xref: /openbsd-src/gnu/usr.bin/perl/pod/perlfunc.pod (revision 3d61058aa5c692477b6d18acfbbdb653a9930ff9)
1=head1 NAME
2X<function>
3
4perlfunc - Perl builtin functions
5
6=head1 DESCRIPTION
7
8The functions in this section can serve as terms in an expression.
9They fall into two major categories: list operators and named unary
10operators.  These differ in their precedence relationship with a
11following comma.  (See the precedence table in L<perlop>.)  List
12operators take more than one argument, while unary operators can never
13take more than one argument.  Thus, a comma terminates the argument of
14a unary operator, but merely separates the arguments of a list
15operator.  A unary operator generally provides scalar context to its
16argument, while a list operator may provide either scalar or list
17contexts for its arguments.  If it does both, scalar arguments
18come first and list argument follow, and there can only ever
19be one such list argument.  For instance,
20L<C<splice>|/splice ARRAY,OFFSET,LENGTH,LIST> has three scalar arguments
21followed by a list, whereas L<C<gethostbyname>|/gethostbyname NAME> has
22four scalar arguments.
23
24In the syntax descriptions that follow, list operators that expect a
25list (and provide list context for elements of the list) are shown
26with LIST as an argument.  Such a list may consist of any combination
27of scalar arguments or list values; the list values will be included
28in the list as if each individual element were interpolated at that
29point in the list, forming a longer single-dimensional list value.
30Commas should separate literal elements of the LIST.
31
32Any function in the list below may be used either with or without
33parentheses around its arguments.  (The syntax descriptions omit the
34parentheses.)  If you use parentheses, the simple but occasionally
35surprising rule is this: It I<looks> like a function, therefore it I<is> a
36function, and precedence doesn't matter.  Otherwise it's a list
37operator or unary operator, and precedence does matter.  Whitespace
38between the function and left parenthesis doesn't count, so sometimes
39you need to be careful:
40
41    print 1+2+4;      # Prints 7.
42    print(1+2) + 4;   # Prints 3.
43    print (1+2)+4;    # Also prints 3!
44    print +(1+2)+4;   # Prints 7.
45    print ((1+2)+4);  # Prints 7.
46
47If you run Perl with the L<C<use warnings>|warnings> pragma, it can warn
48you about this.  For example, the third line above produces:
49
50    print (...) interpreted as function at - line 1.
51    Useless use of integer addition in void context at - line 1.
52
53A few functions take no arguments at all, and therefore work as neither
54unary nor list operators.  These include such functions as
55L<C<time>|/time> and L<C<endpwent>|/endpwent>.  For example,
56C<time+86_400> always means C<time() + 86_400>.
57
58For functions that can be used in either a scalar or list context,
59nonabortive failure is generally indicated in scalar context by
60returning the undefined value, and in list context by returning the
61empty list.
62
63Remember the following important rule: There is B<no rule> that relates
64the behavior of an expression in list context to its behavior in scalar
65context, or vice versa.  It might do two totally different things.
66Each operator and function decides which sort of value would be most
67appropriate to return in scalar context.  Some operators return the
68length of the list that would have been returned in list context.  Some
69operators return the first value in the list.  Some operators return the
70last value in the list.  Some operators return a count of successful
71operations.  In general, they do what you want, unless you want
72consistency.
73X<context>
74
75A named array in scalar context is quite different from what would at
76first glance appear to be a list in scalar context.  You can't get a list
77like C<(1,2,3)> into being in scalar context, because the compiler knows
78the context at compile time.  It would generate the scalar comma operator
79there, not the list concatenation version of the comma.  That means it
80was never a list to start with.
81
82In general, functions in Perl that serve as wrappers for system calls
83("syscalls") of the same name (like L<chown(2)>, L<fork(2)>,
84L<closedir(2)>, etc.) return true when they succeed and
85L<C<undef>|/undef EXPR> otherwise, as is usually mentioned in the
86descriptions below.  This is different from the C interfaces, which
87return C<-1> on failure.  Exceptions to this rule include
88L<C<wait>|/wait>, L<C<waitpid>|/waitpid PID,FLAGS>, and
89L<C<syscall>|/syscall NUMBER, LIST>.  System calls also set the special
90L<C<$!>|perlvar/$!> variable on failure.  Other functions do not, except
91accidentally.
92
93Extension modules can also hook into the Perl parser to define new
94kinds of keyword-headed expression.  These may look like functions, but
95may also look completely different.  The syntax following the keyword
96is defined entirely by the extension.  If you are an implementor, see
97L<perlapi/PL_keyword_plugin> for the mechanism.  If you are using such
98a module, see the module's documentation for details of the syntax that
99it defines.
100
101=head2 Perl Functions by Category
102X<function>
103
104Here are Perl's functions (including things that look like
105functions, like some keywords and named operators)
106arranged by category.  Some functions appear in more
107than one place.  Any warnings, including those produced by
108keywords, are described in L<perldiag> and L<warnings>.
109
110=over 4
111
112=item Functions for SCALARs or strings
113X<scalar> X<string> X<character>
114
115=for Pod::Functions =String
116
117L<C<chomp>|/chomp VARIABLE>, L<C<chop>|/chop VARIABLE>,
118L<C<chr>|/chr NUMBER>, L<C<crypt>|/crypt PLAINTEXT,SALT>,
119L<C<fc>|/fc EXPR>, L<C<hex>|/hex EXPR>,
120L<C<index>|/index STR,SUBSTR,POSITION>, L<C<lc>|/lc EXPR>,
121L<C<lcfirst>|/lcfirst EXPR>, L<C<length>|/length EXPR>,
122L<C<oct>|/oct EXPR>, L<C<ord>|/ord EXPR>,
123L<C<pack>|/pack TEMPLATE,LIST>,
124L<C<qE<sol>E<sol>>|/qE<sol>STRINGE<sol>>,
125L<C<qqE<sol>E<sol>>|/qqE<sol>STRINGE<sol>>, L<C<reverse>|/reverse LIST>,
126L<C<rindex>|/rindex STR,SUBSTR,POSITION>,
127L<C<sprintf>|/sprintf FORMAT, LIST>,
128L<C<substr>|/substr EXPR,OFFSET,LENGTH,REPLACEMENT>,
129L<C<trE<sol>E<sol>E<sol>>|/trE<sol>E<sol>E<sol>>, L<C<uc>|/uc EXPR>,
130L<C<ucfirst>|/ucfirst EXPR>,
131L<C<yE<sol>E<sol>E<sol>>|/yE<sol>E<sol>E<sol>>
132
133L<C<fc>|/fc EXPR> is available only if the
134L<C<"fc"> feature|feature/The 'fc' feature> is enabled or if it is
135prefixed with C<CORE::>.  The
136L<C<"fc"> feature|feature/The 'fc' feature> is enabled automatically
137with a C<use v5.16> (or higher) declaration in the current scope.
138
139=item Regular expressions and pattern matching
140X<regular expression> X<regex> X<regexp>
141
142=for Pod::Functions =Regexp
143
144L<C<mE<sol>E<sol>>|/mE<sol>E<sol>>, L<C<pos>|/pos SCALAR>,
145L<C<qrE<sol>E<sol>>|/qrE<sol>STRINGE<sol>>,
146L<C<quotemeta>|/quotemeta EXPR>,
147L<C<sE<sol>E<sol>E<sol>>|/sE<sol>E<sol>E<sol>>,
148L<C<split>|/split E<sol>PATTERNE<sol>,EXPR,LIMIT>,
149L<C<study>|/study SCALAR>
150
151=item Numeric functions
152X<numeric> X<number> X<trigonometric> X<trigonometry>
153
154=for Pod::Functions =Math
155
156L<C<abs>|/abs VALUE>, L<C<atan2>|/atan2 Y,X>, L<C<cos>|/cos EXPR>,
157L<C<exp>|/exp EXPR>, L<C<hex>|/hex EXPR>, L<C<int>|/int EXPR>,
158L<C<log>|/log EXPR>, L<C<oct>|/oct EXPR>, L<C<rand>|/rand EXPR>,
159L<C<sin>|/sin EXPR>, L<C<sqrt>|/sqrt EXPR>, L<C<srand>|/srand EXPR>
160
161=item Functions for real @ARRAYs
162X<array>
163
164=for Pod::Functions =ARRAY
165
166L<C<each>|/each HASH>, L<C<keys>|/keys HASH>, L<C<pop>|/pop ARRAY>,
167L<C<push>|/push ARRAY,LIST>, L<C<shift>|/shift ARRAY>,
168L<C<splice>|/splice ARRAY,OFFSET,LENGTH,LIST>,
169L<C<unshift>|/unshift ARRAY,LIST>, L<C<values>|/values HASH>
170
171=item Functions for list data
172X<list>
173
174=for Pod::Functions =LIST
175
176L<C<grep>|/grep BLOCK LIST>, L<C<join>|/join EXPR,LIST>,
177L<C<map>|/map BLOCK LIST>, L<C<qwE<sol>E<sol>>|/qwE<sol>STRINGE<sol>>,
178L<C<reverse>|/reverse LIST>, L<C<sort>|/sort SUBNAME LIST>,
179L<C<unpack>|/unpack TEMPLATE,EXPR>
180
181=item Functions for real %HASHes
182X<hash>
183
184=for Pod::Functions =HASH
185
186L<C<delete>|/delete EXPR>, L<C<each>|/each HASH>,
187L<C<exists>|/exists EXPR>, L<C<keys>|/keys HASH>,
188L<C<values>|/values HASH>
189
190=item Input and output functions
191X<I/O> X<input> X<output> X<dbm>
192
193=for Pod::Functions =I/O
194
195L<C<binmode>|/binmode FILEHANDLE, LAYER>, L<C<close>|/close FILEHANDLE>,
196L<C<closedir>|/closedir DIRHANDLE>, L<C<dbmclose>|/dbmclose HASH>,
197L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK>, L<C<die>|/die LIST>,
198L<C<eof>|/eof FILEHANDLE>, L<C<fileno>|/fileno FILEHANDLE>,
199L<C<flock>|/flock FILEHANDLE,OPERATION>, L<C<format>|/format>,
200L<C<getc>|/getc FILEHANDLE>, L<C<print>|/print FILEHANDLE LIST>,
201L<C<printf>|/printf FILEHANDLE FORMAT, LIST>,
202L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET>,
203L<C<readdir>|/readdir DIRHANDLE>, L<C<readline>|/readline EXPR>,
204L<C<rewinddir>|/rewinddir DIRHANDLE>, L<C<say>|/say FILEHANDLE LIST>,
205L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
206L<C<seekdir>|/seekdir DIRHANDLE,POS>,
207L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT>,
208L<C<syscall>|/syscall NUMBER, LIST>,
209L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET>,
210L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>,
211L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET>,
212L<C<tell>|/tell FILEHANDLE>, L<C<telldir>|/telldir DIRHANDLE>,
213L<C<truncate>|/truncate FILEHANDLE,LENGTH>, L<C<warn>|/warn LIST>,
214L<C<write>|/write FILEHANDLE>
215
216L<C<say>|/say FILEHANDLE LIST> is available only if the
217L<C<"say"> feature|feature/The 'say' feature> is enabled or if it is
218prefixed with C<CORE::>.  The
219L<C<"say"> feature|feature/The 'say' feature> is enabled automatically
220with a C<use v5.10> (or higher) declaration in the current scope.
221
222=item Functions for fixed-length data or records
223
224=for Pod::Functions =Binary
225
226L<C<pack>|/pack TEMPLATE,LIST>,
227L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET>,
228L<C<syscall>|/syscall NUMBER, LIST>,
229L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET>,
230L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>,
231L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET>,
232L<C<unpack>|/unpack TEMPLATE,EXPR>, L<C<vec>|/vec EXPR,OFFSET,BITS>
233
234=item Functions for filehandles, files, or directories
235X<file> X<filehandle> X<directory> X<pipe> X<link> X<symlink>
236
237=for Pod::Functions =File
238
239L<C<-I<X>>|/-X FILEHANDLE>, L<C<chdir>|/chdir EXPR>,
240L<C<chmod>|/chmod LIST>, L<C<chown>|/chown LIST>,
241L<C<chroot>|/chroot FILENAME>,
242L<C<fcntl>|/fcntl FILEHANDLE,FUNCTION,SCALAR>, L<C<glob>|/glob EXPR>,
243L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR>,
244L<C<link>|/link OLDFILE,NEWFILE>, L<C<lstat>|/lstat FILEHANDLE>,
245L<C<mkdir>|/mkdir FILENAME,MODE>, L<C<open>|/open FILEHANDLE,MODE,EXPR>,
246L<C<opendir>|/opendir DIRHANDLE,EXPR>, L<C<readlink>|/readlink EXPR>,
247L<C<rename>|/rename OLDNAME,NEWNAME>, L<C<rmdir>|/rmdir FILENAME>,
248L<C<select>|/select FILEHANDLE>, L<C<stat>|/stat FILEHANDLE>,
249L<C<symlink>|/symlink OLDFILE,NEWFILE>,
250L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE>,
251L<C<umask>|/umask EXPR>, L<C<unlink>|/unlink LIST>,
252L<C<utime>|/utime LIST>
253
254=item Keywords related to the control flow of your Perl program
255X<control flow>
256
257=for Pod::Functions =Flow
258
259L<C<break>|/break>, L<C<caller>|/caller EXPR>,
260L<C<continue>|/continue BLOCK>, L<C<die>|/die LIST>, L<C<do>|/do BLOCK>,
261L<C<dump>|/dump LABEL>, L<C<eval>|/eval EXPR>,
262L<C<evalbytes>|/evalbytes EXPR>, L<C<exit>|/exit EXPR>,
263L<C<__FILE__>|/__FILE__>, L<C<goto>|/goto LABEL>,
264L<C<last>|/last LABEL>, L<C<__LINE__>|/__LINE__>,
265L<C<method>|/method NAME BLOCK>,
266L<C<next>|/next LABEL>, L<C<__PACKAGE__>|/__PACKAGE__>,
267L<C<redo>|/redo LABEL>, L<C<return>|/return EXPR>,
268L<C<sub>|/sub NAME BLOCK>, L<C<__SUB__>|/__SUB__>,
269L<C<wantarray>|/wantarray>
270
271L<C<break>|/break> is available only if you enable the experimental
272L<C<"switch"> feature|feature/The 'switch' feature> or use the C<CORE::>
273prefix.  The L<C<"switch"> feature|feature/The 'switch' feature> also
274enables the C<default>, C<given> and C<when> statements, which are
275documented in L<perlsyn/"Switch Statements">.
276The L<C<"switch"> feature|feature/The 'switch' feature> is enabled
277automatically with a C<use v5.10> (or higher) declaration in the current
278scope.  In Perl v5.14 and earlier, L<C<continue>|/continue BLOCK>
279required the L<C<"switch"> feature|feature/The 'switch' feature>, like
280the other keywords.
281
282L<C<evalbytes>|/evalbytes EXPR> is only available with the
283L<C<"evalbytes"> feature|feature/The 'unicode_eval' and 'evalbytes' features>
284(see L<feature>) or if prefixed with C<CORE::>.  L<C<__SUB__>|/__SUB__>
285is only available with the
286L<C<"current_sub"> feature|feature/The 'current_sub' feature> or if
287prefixed with C<CORE::>.  Both the
288L<C<"evalbytes">|feature/The 'unicode_eval' and 'evalbytes' features>
289and L<C<"current_sub">|feature/The 'current_sub' feature> features are
290enabled automatically with a C<use v5.16> (or higher) declaration in the
291current scope.
292
293=item Keywords related to scoping
294
295=for Pod::Functions =Namespace
296
297L<C<caller>|/caller EXPR>,
298L<C<class>|/class NAMESPACE>,
299L<C<field>|/field VARNAME>,
300L<C<import>|/import LIST>,
301L<C<local>|/local EXPR>,
302L<C<my>|/my VARLIST>,
303L<C<our>|/our VARLIST>,
304L<C<package>|/package NAMESPACE>,
305L<C<state>|/state VARLIST>,
306L<C<use>|/use Module VERSION LIST>
307
308L<C<state>|/state VARLIST> is available only if the
309L<C<"state"> feature|feature/The 'state' feature> is enabled or if it is
310prefixed with C<CORE::>.  The
311L<C<"state"> feature|feature/The 'state' feature> is enabled
312automatically with a C<use v5.10> (or higher) declaration in the current
313scope.
314
315=item Miscellaneous functions
316
317=for Pod::Functions =Misc
318
319L<C<defined>|/defined EXPR>, L<C<formline>|/formline PICTURE,LIST>,
320L<C<lock>|/lock THING>, L<C<prototype>|/prototype FUNCTION>,
321L<C<reset>|/reset EXPR>, L<C<scalar>|/scalar EXPR>,
322L<C<undef>|/undef EXPR>
323
324=item Functions for processes and process groups
325X<process> X<pid> X<process id>
326
327=for Pod::Functions =Process
328
329L<C<alarm>|/alarm SECONDS>, L<C<exec>|/exec LIST>, L<C<fork>|/fork>,
330L<C<getpgrp>|/getpgrp PID>, L<C<getppid>|/getppid>,
331L<C<getpriority>|/getpriority WHICH,WHO>, L<C<kill>|/kill SIGNAL, LIST>,
332L<C<pipe>|/pipe READHANDLE,WRITEHANDLE>,
333L<C<qxE<sol>E<sol>>|/qxE<sol>STRINGE<sol>>,
334L<C<readpipe>|/readpipe EXPR>, L<C<setpgrp>|/setpgrp PID,PGRP>,
335L<C<setpriority>|/setpriority WHICH,WHO,PRIORITY>,
336L<C<sleep>|/sleep EXPR>, L<C<system>|/system LIST>, L<C<times>|/times>,
337L<C<wait>|/wait>, L<C<waitpid>|/waitpid PID,FLAGS>
338
339=item Keywords related to Perl modules
340X<module>
341
342=for Pod::Functions =Modules
343
344L<C<do>|/do EXPR>, L<C<import>|/import LIST>,
345L<C<no>|/no MODULE VERSION LIST>, L<C<package>|/package NAMESPACE>,
346L<C<require>|/require VERSION>, L<C<use>|/use Module VERSION LIST>
347
348=item Keywords related to classes and object-orientation
349X<object> X<class> X<package>
350
351=for Pod::Functions =Objects
352
353L<C<bless>|/bless REF,CLASSNAME>,
354L<C<class>|/class NAMESPACE>,
355L<C<__CLASS__>|/__CLASS__>,
356L<C<dbmclose>|/dbmclose HASH>,
357L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK>,
358L<C<field>|/field VARNAME>,
359L<C<method>|/method NAME BLOCK>,
360L<C<package>|/package NAMESPACE>,
361L<C<ref>|/ref EXPR>,
362L<C<tie>|/tie VARIABLE,CLASSNAME,LIST>,
363L<C<tied>|/tied VARIABLE>,
364L<C<untie>|/untie VARIABLE>,
365L<C<use>|/use Module VERSION LIST>
366
367=item Low-level socket functions
368X<socket> X<sock>
369
370=for Pod::Functions =Socket
371
372L<C<accept>|/accept NEWSOCKET,GENERICSOCKET>,
373L<C<bind>|/bind SOCKET,NAME>, L<C<connect>|/connect SOCKET,NAME>,
374L<C<getpeername>|/getpeername SOCKET>,
375L<C<getsockname>|/getsockname SOCKET>,
376L<C<getsockopt>|/getsockopt SOCKET,LEVEL,OPTNAME>,
377L<C<listen>|/listen SOCKET,QUEUESIZE>,
378L<C<recv>|/recv SOCKET,SCALAR,LENGTH,FLAGS>,
379L<C<send>|/send SOCKET,MSG,FLAGS,TO>,
380L<C<setsockopt>|/setsockopt SOCKET,LEVEL,OPTNAME,OPTVAL>,
381L<C<shutdown>|/shutdown SOCKET,HOW>,
382L<C<socket>|/socket SOCKET,DOMAIN,TYPE,PROTOCOL>,
383L<C<socketpair>|/socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL>
384
385=item System V interprocess communication functions
386X<IPC> X<System V> X<semaphore> X<shared memory> X<memory> X<message>
387
388=for Pod::Functions =SysV
389
390L<C<msgctl>|/msgctl ID,CMD,ARG>, L<C<msgget>|/msgget KEY,FLAGS>,
391L<C<msgrcv>|/msgrcv ID,VAR,SIZE,TYPE,FLAGS>,
392L<C<msgsnd>|/msgsnd ID,MSG,FLAGS>,
393L<C<semctl>|/semctl ID,SEMNUM,CMD,ARG>,
394L<C<semget>|/semget KEY,NSEMS,FLAGS>, L<C<semop>|/semop KEY,OPSTRING>,
395L<C<shmctl>|/shmctl ID,CMD,ARG>, L<C<shmget>|/shmget KEY,SIZE,FLAGS>,
396L<C<shmread>|/shmread ID,VAR,POS,SIZE>,
397L<C<shmwrite>|/shmwrite ID,STRING,POS,SIZE>
398
399=item Fetching user and group info
400X<user> X<group> X<password> X<uid> X<gid>  X<passwd> X</etc/passwd>
401
402=for Pod::Functions =User
403
404L<C<endgrent>|/endgrent>, L<C<endhostent>|/endhostent>,
405L<C<endnetent>|/endnetent>, L<C<endpwent>|/endpwent>,
406L<C<getgrent>|/getgrent>, L<C<getgrgid>|/getgrgid GID>,
407L<C<getgrnam>|/getgrnam NAME>, L<C<getlogin>|/getlogin>,
408L<C<getpwent>|/getpwent>, L<C<getpwnam>|/getpwnam NAME>,
409L<C<getpwuid>|/getpwuid UID>, L<C<setgrent>|/setgrent>,
410L<C<setpwent>|/setpwent>
411
412=item Fetching network info
413X<network> X<protocol> X<host> X<hostname> X<IP> X<address> X<service>
414
415=for Pod::Functions =Network
416
417L<C<endprotoent>|/endprotoent>, L<C<endservent>|/endservent>,
418L<C<gethostbyaddr>|/gethostbyaddr ADDR,ADDRTYPE>,
419L<C<gethostbyname>|/gethostbyname NAME>, L<C<gethostent>|/gethostent>,
420L<C<getnetbyaddr>|/getnetbyaddr ADDR,ADDRTYPE>,
421L<C<getnetbyname>|/getnetbyname NAME>, L<C<getnetent>|/getnetent>,
422L<C<getprotobyname>|/getprotobyname NAME>,
423L<C<getprotobynumber>|/getprotobynumber NUMBER>,
424L<C<getprotoent>|/getprotoent>,
425L<C<getservbyname>|/getservbyname NAME,PROTO>,
426L<C<getservbyport>|/getservbyport PORT,PROTO>,
427L<C<getservent>|/getservent>, L<C<sethostent>|/sethostent STAYOPEN>,
428L<C<setnetent>|/setnetent STAYOPEN>,
429L<C<setprotoent>|/setprotoent STAYOPEN>,
430L<C<setservent>|/setservent STAYOPEN>
431
432=item Time-related functions
433X<time> X<date>
434
435=for Pod::Functions =Time
436
437L<C<gmtime>|/gmtime EXPR>, L<C<localtime>|/localtime EXPR>,
438L<C<time>|/time>, L<C<times>|/times>
439
440=item Non-function keywords
441
442=for Pod::Functions =!Non-functions
443
444C<ADJUST>,
445C<and>,
446C<AUTOLOAD>,
447C<BEGIN>,
448C<catch>,
449C<CHECK>,
450C<cmp>,
451C<CORE>,
452C<__DATA__>,
453C<default>,
454C<defer>,
455C<DESTROY>,
456C<else>,
457C<elseif>,
458C<elsif>,
459C<END>,
460C<__END__>,
461C<eq>,
462C<finally>,
463C<for>,
464C<foreach>,
465C<ge>,
466C<given>,
467C<gt>,
468C<if>,
469C<INIT>,
470C<isa>,
471C<le>,
472C<lt>,
473C<ne>,
474C<not>,
475C<or>,
476C<try>,
477C<UNITCHECK>,
478C<unless>,
479C<until>,
480C<when>,
481C<while>,
482C<x>,
483C<xor>
484
485=back
486
487=head2 Portability
488X<portability> X<Unix> X<portable>
489
490Perl was born in Unix and can therefore access all common Unix
491system calls.  In non-Unix environments, the functionality of some
492Unix system calls may not be available or details of the available
493functionality may differ slightly.  The Perl functions affected
494by this are:
495
496L<C<-I<X>>|/-X FILEHANDLE>, L<C<binmode>|/binmode FILEHANDLE, LAYER>,
497L<C<chmod>|/chmod LIST>, L<C<chown>|/chown LIST>,
498L<C<chroot>|/chroot FILENAME>, L<C<crypt>|/crypt PLAINTEXT,SALT>,
499L<C<dbmclose>|/dbmclose HASH>, L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK>,
500L<C<dump>|/dump LABEL>, L<C<endgrent>|/endgrent>,
501L<C<endhostent>|/endhostent>, L<C<endnetent>|/endnetent>,
502L<C<endprotoent>|/endprotoent>, L<C<endpwent>|/endpwent>,
503L<C<endservent>|/endservent>, L<C<exec>|/exec LIST>,
504L<C<fcntl>|/fcntl FILEHANDLE,FUNCTION,SCALAR>,
505L<C<flock>|/flock FILEHANDLE,OPERATION>, L<C<fork>|/fork>,
506L<C<getgrent>|/getgrent>, L<C<getgrgid>|/getgrgid GID>,
507L<C<gethostbyname>|/gethostbyname NAME>, L<C<gethostent>|/gethostent>,
508L<C<getlogin>|/getlogin>,
509L<C<getnetbyaddr>|/getnetbyaddr ADDR,ADDRTYPE>,
510L<C<getnetbyname>|/getnetbyname NAME>, L<C<getnetent>|/getnetent>,
511L<C<getppid>|/getppid>, L<C<getpgrp>|/getpgrp PID>,
512L<C<getpriority>|/getpriority WHICH,WHO>,
513L<C<getprotobynumber>|/getprotobynumber NUMBER>,
514L<C<getprotoent>|/getprotoent>, L<C<getpwent>|/getpwent>,
515L<C<getpwnam>|/getpwnam NAME>, L<C<getpwuid>|/getpwuid UID>,
516L<C<getservbyport>|/getservbyport PORT,PROTO>,
517L<C<getservent>|/getservent>,
518L<C<getsockopt>|/getsockopt SOCKET,LEVEL,OPTNAME>,
519L<C<glob>|/glob EXPR>, L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR>,
520L<C<kill>|/kill SIGNAL, LIST>, L<C<link>|/link OLDFILE,NEWFILE>,
521L<C<lstat>|/lstat FILEHANDLE>, L<C<msgctl>|/msgctl ID,CMD,ARG>,
522L<C<msgget>|/msgget KEY,FLAGS>,
523L<C<msgrcv>|/msgrcv ID,VAR,SIZE,TYPE,FLAGS>,
524L<C<msgsnd>|/msgsnd ID,MSG,FLAGS>, L<C<open>|/open FILEHANDLE,MODE,EXPR>,
525L<C<pipe>|/pipe READHANDLE,WRITEHANDLE>, L<C<readlink>|/readlink EXPR>,
526L<C<rename>|/rename OLDNAME,NEWNAME>,
527L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT>,
528L<C<semctl>|/semctl ID,SEMNUM,CMD,ARG>,
529L<C<semget>|/semget KEY,NSEMS,FLAGS>, L<C<semop>|/semop KEY,OPSTRING>,
530L<C<setgrent>|/setgrent>, L<C<sethostent>|/sethostent STAYOPEN>,
531L<C<setnetent>|/setnetent STAYOPEN>, L<C<setpgrp>|/setpgrp PID,PGRP>,
532L<C<setpriority>|/setpriority WHICH,WHO,PRIORITY>,
533L<C<setprotoent>|/setprotoent STAYOPEN>, L<C<setpwent>|/setpwent>,
534L<C<setservent>|/setservent STAYOPEN>,
535L<C<setsockopt>|/setsockopt SOCKET,LEVEL,OPTNAME,OPTVAL>,
536L<C<shmctl>|/shmctl ID,CMD,ARG>, L<C<shmget>|/shmget KEY,SIZE,FLAGS>,
537L<C<shmread>|/shmread ID,VAR,POS,SIZE>,
538L<C<shmwrite>|/shmwrite ID,STRING,POS,SIZE>,
539L<C<socket>|/socket SOCKET,DOMAIN,TYPE,PROTOCOL>,
540L<C<socketpair>|/socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL>,
541L<C<stat>|/stat FILEHANDLE>, L<C<symlink>|/symlink OLDFILE,NEWFILE>,
542L<C<syscall>|/syscall NUMBER, LIST>,
543L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE>,
544L<C<system>|/system LIST>, L<C<times>|/times>,
545L<C<truncate>|/truncate FILEHANDLE,LENGTH>, L<C<umask>|/umask EXPR>,
546L<C<unlink>|/unlink LIST>, L<C<utime>|/utime LIST>, L<C<wait>|/wait>,
547L<C<waitpid>|/waitpid PID,FLAGS>
548
549For more information about the portability of these functions, see
550L<perlport> and other available platform-specific documentation.
551
552=head2 Alphabetical Listing of Perl Functions
553
554=over
555
556=item -X FILEHANDLE
557X<-r>X<-w>X<-x>X<-o>X<-R>X<-W>X<-X>X<-O>X<-e>X<-z>X<-s>X<-f>X<-d>X<-l>X<-p>
558X<-S>X<-b>X<-c>X<-t>X<-u>X<-g>X<-k>X<-T>X<-B>X<-M>X<-A>X<-C>
559
560=item -X EXPR
561
562=item -X DIRHANDLE
563
564=item -X
565
566=for Pod::Functions a file test (-r, -x, etc)
567
568A file test, where X is one of the letters listed below.  This unary
569operator takes one argument, either a filename, a filehandle, or a dirhandle,
570and tests the associated file to see if something is true about it.  If the
571argument is omitted, tests L<C<$_>|perlvar/$_>, except for C<-t>, which
572tests STDIN.  Unless otherwise documented, it returns C<1> for true and
573C<''> for false.  If the file doesn't exist or can't be examined, it
574returns L<C<undef>|/undef EXPR> and sets L<C<$!>|perlvar/$!> (errno).
575With the exception of the C<-l> test they all follow symbolic links
576because they use C<stat()> and not C<lstat()> (so dangling symlinks can't
577be examined and will therefore report failure).
578
579Despite the funny names, precedence is the same as any other named unary
580operator.  The operator may be any of:
581
582    -r  File is readable by effective uid/gid.
583    -w  File is writable by effective uid/gid.
584    -x  File is executable by effective uid/gid.
585    -o  File is owned by effective uid.
586
587    -R  File is readable by real uid/gid.
588    -W  File is writable by real uid/gid.
589    -X  File is executable by real uid/gid.
590    -O  File is owned by real uid.
591
592    -e  File exists.
593    -z  File has zero size (is empty).
594    -s  File has nonzero size (returns size in bytes).
595
596    -f  File is a plain file.
597    -d  File is a directory.
598    -l  File is a symbolic link (false if symlinks aren't
599        supported by the file system).
600    -p  File is a named pipe (FIFO), or Filehandle is a pipe.
601    -S  File is a socket.
602    -b  File is a block special file.
603    -c  File is a character special file.
604    -t  Filehandle is opened to a tty.
605
606    -u  File has setuid bit set.
607    -g  File has setgid bit set.
608    -k  File has sticky bit set.
609
610    -T  File is an ASCII or UTF-8 text file (heuristic guess).
611    -B  File is a "binary" file (opposite of -T).
612
613    -M  Script start time minus file modification time, in days.
614    -A  Same for access time.
615    -C  Same for inode change time (Unix, may differ for other
616	platforms)
617
618Example:
619
620    while (<>) {
621        chomp;
622        next unless -f $_;  # ignore specials
623        #...
624    }
625
626Note that C<-s/a/b/> does not do a negated substitution.  Saying
627C<-exp($foo)> still works as expected, however: only single letters
628following a minus are interpreted as file tests.
629
630These operators are exempt from the "looks like a function rule" described
631above.  That is, an opening parenthesis after the operator does not affect
632how much of the following code constitutes the argument.  Put the opening
633parentheses before the operator to separate it from code that follows (this
634applies only to operators with higher precedence than unary operators, of
635course):
636
637    -s($file) + 1024   # probably wrong; same as -s($file + 1024)
638    (-s $file) + 1024  # correct
639
640The interpretation of the file permission operators C<-r>, C<-R>,
641C<-w>, C<-W>, C<-x>, and C<-X> is by default based solely on the mode
642of the file and the uids and gids of the user.  There may be other
643reasons you can't actually read, write, or execute the file: for
644example network filesystem access controls, ACLs (access control lists),
645read-only filesystems, and unrecognized executable formats.  Note
646that the use of these six specific operators to verify if some operation
647is possible is usually a mistake, because it may be open to race
648conditions.
649
650Also note that, for the superuser on the local filesystems, the C<-r>,
651C<-R>, C<-w>, and C<-W> tests always return 1, and C<-x> and C<-X> return 1
652if any execute bit is set in the mode.  Scripts run by the superuser
653may thus need to do a L<C<stat>|/stat FILEHANDLE> to determine the
654actual mode of the file, or temporarily set their effective uid to
655something else.
656
657If you are using ACLs, there is a pragma called L<C<filetest>|filetest>
658that may produce more accurate results than the bare
659L<C<stat>|/stat FILEHANDLE> mode bits.
660When under C<use filetest 'access'>, the above-mentioned filetests
661test whether the permission can(not) be granted using the L<access(2)>
662family of system calls.  Also note that the C<-x> and C<-X> tests may
663under this pragma return true even if there are no execute permission
664bits set (nor any extra execute permission ACLs).  This strangeness is
665due to the underlying system calls' definitions.  Note also that, due to
666the implementation of C<use filetest 'access'>, the C<_> special
667filehandle won't cache the results of the file tests when this pragma is
668in effect.  Read the documentation for the L<C<filetest>|filetest>
669pragma for more information.
670
671The C<-T> and C<-B> tests work as follows.  The first block or so of
672the file is examined to see if it is valid UTF-8 that includes non-ASCII
673characters.  If so, it's a C<-T> file.  Otherwise, that same portion of
674the file is examined for odd characters such as strange control codes or
675characters with the high bit set.  If more than a third of the
676characters are strange, it's a C<-B> file; otherwise it's a C<-T> file.
677Also, any file containing a zero byte in the examined portion is
678considered a binary file.  (If executed within the scope of a L<S<use
679locale>|perllocale> which includes C<LC_CTYPE>, odd characters are
680anything that isn't a printable nor space in the current locale.)  If
681C<-T> or C<-B> is used on a filehandle, the current IO buffer is
682examined
683rather than the first block.  Both C<-T> and C<-B> return true on an empty
684file, or a file at EOF when testing a filehandle.  Because you have to
685read a file to do the C<-T> test, on most occasions you want to use a C<-f>
686against the file first, as in C<next unless -f $file && -T $file>.
687
688If any of the file tests (or either the L<C<stat>|/stat FILEHANDLE> or
689L<C<lstat>|/lstat FILEHANDLE> operator) is given the special filehandle
690consisting of a solitary underline, then the stat structure of the
691previous file test (or L<C<stat>|/stat FILEHANDLE> operator) is used,
692saving a system call.  (This doesn't work with C<-t>, and you need to
693remember that L<C<lstat>|/lstat FILEHANDLE> and C<-l> leave values in
694the stat structure for the symbolic link, not the real file.)  (Also, if
695the stat buffer was filled by an L<C<lstat>|/lstat FILEHANDLE> call,
696C<-T> and C<-B> will reset it with the results of C<stat _>).
697Example:
698
699    print "Can do.\n" if -r $x || -w _ || -x _;
700
701    stat($filename);
702    print "Readable\n" if -r _;
703    print "Writable\n" if -w _;
704    print "Executable\n" if -x _;
705    print "Setuid\n" if -u _;
706    print "Setgid\n" if -g _;
707    print "Sticky\n" if -k _;
708    print "Text\n" if -T _;
709    print "Binary\n" if -B _;
710
711As of Perl 5.10.0, as a form of purely syntactic sugar, you can stack file
712test operators, in a way that C<-f -w -x $file> is equivalent to
713C<-x $file && -w _ && -f _>.  (This is only fancy syntax: if you use
714the return value of C<-f $file> as an argument to another filetest
715operator, no special magic will happen.)
716
717Portability issues: L<perlport/-X>.
718
719To avoid confusing would-be users of your code with mysterious
720syntax errors, put something like this at the top of your script:
721
722    use v5.10;  # so filetest ops can stack
723
724=item abs VALUE
725X<abs> X<absolute>
726
727=item abs
728
729=for Pod::Functions absolute value function
730
731Returns the absolute value of its argument.
732If VALUE is omitted, uses L<C<$_>|perlvar/$_>.
733
734=item accept NEWSOCKET,GENERICSOCKET
735X<accept>
736
737=for Pod::Functions accept an incoming socket connect
738
739Accepts an incoming socket connect, just as L<accept(2)>
740does.  Returns the packed address if it succeeded, false otherwise.
741See the example in L<perlipc/"Sockets: Client/Server Communication">.
742
743On systems that support a close-on-exec flag on files, the flag will
744be set for the newly opened file descriptor, as determined by the
745value of L<C<$^F>|perlvar/$^F>.  See L<perlvar/$^F>.
746
747=item alarm SECONDS
748X<alarm>
749X<SIGALRM>
750X<timer>
751
752=item alarm
753
754=for Pod::Functions schedule a SIGALRM
755
756Arranges to have a SIGALRM delivered to this process after the
757specified number of wallclock seconds has elapsed.  If SECONDS is not
758specified, the value stored in L<C<$_>|perlvar/$_> is used.  (On some
759machines, unfortunately, the elapsed time may be up to one second less
760or more than you specified because of how seconds are counted, and
761process scheduling may delay the delivery of the signal even further.)
762
763Only one timer may be counting at once.  Each call disables the
764previous timer, and an argument of C<0> may be supplied to cancel the
765previous timer without starting a new one.  The returned value is the
766amount of time remaining on the previous timer.
767
768For delays of finer granularity than one second, the L<Time::HiRes> module
769(from CPAN, and starting from Perl 5.8 part of the standard
770distribution) provides
771L<C<ualarm>|Time::HiRes/ualarm ( $useconds [, $interval_useconds ] )>.
772You may also use Perl's four-argument version of
773L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT> leaving the first three
774arguments undefined, or you might be able to use the
775L<C<syscall>|/syscall NUMBER, LIST> interface to access L<setitimer(2)>
776if your system supports it.  See L<perlfaq8> for details.
777
778It is usually a mistake to intermix L<C<alarm>|/alarm SECONDS> and
779L<C<sleep>|/sleep EXPR> calls, because L<C<sleep>|/sleep EXPR> may be
780internally implemented on your system with L<C<alarm>|/alarm SECONDS>.
781
782If you want to use L<C<alarm>|/alarm SECONDS> to time out a system call
783you need to use an L<C<eval>|/eval EXPR>/L<C<die>|/die LIST> pair.  You
784can't rely on the alarm causing the system call to fail with
785L<C<$!>|perlvar/$!> set to C<EINTR> because Perl sets up signal handlers
786to restart system calls on some systems.  Using
787L<C<eval>|/eval EXPR>/L<C<die>|/die LIST> always works, modulo the
788caveats given in L<perlipc/"Signals">.
789
790    eval {
791        local $SIG{ALRM} = sub { die "alarm\n" }; # NB: \n required
792        alarm $timeout;
793        my $nread = sysread $socket, $buffer, $size;
794        alarm 0;
795    };
796    if ($@) {
797        die unless $@ eq "alarm\n";   # propagate unexpected errors
798        # timed out
799    }
800    else {
801        # didn't
802    }
803
804For more information see L<perlipc>.
805
806Portability issues: L<perlport/alarm>.
807
808=item atan2 Y,X
809X<atan2> X<arctangent> X<tan> X<tangent>
810
811=for Pod::Functions arctangent of Y/X in the range -PI to PI
812
813Returns the arctangent of Y/X in the range -PI to PI.
814
815For the tangent operation, you may use the
816L<C<Math::Trig::tan>|Math::Trig/B<tan>> function, or use the familiar
817relation:
818
819    sub tan { sin($_[0]) / cos($_[0])  }
820
821The return value for C<atan2(0,0)> is implementation-defined; consult
822your L<atan2(3)> manpage for more information.
823
824Portability issues: L<perlport/atan2>.
825
826=item bind SOCKET,NAME
827X<bind>
828
829=for Pod::Functions binds an address to a socket
830
831Binds a network address to a socket, just as L<bind(2)>
832does.  Returns true if it succeeded, false otherwise.  NAME should be a
833packed address of the appropriate type for the socket.  See the examples in
834L<perlipc/"Sockets: Client/Server Communication">.
835
836=item binmode FILEHANDLE, LAYER
837X<binmode> X<binary> X<text> X<DOS> X<Windows>
838
839=item binmode FILEHANDLE
840
841=for Pod::Functions prepare binary files for I/O
842
843Arranges for FILEHANDLE to be read or written in "binary" or "text"
844mode on systems where the run-time libraries distinguish between
845binary and text files.  If FILEHANDLE is an expression, the value is
846taken as the name of the filehandle.  Returns true on success,
847otherwise it returns L<C<undef>|/undef EXPR> and sets
848L<C<$!>|perlvar/$!> (errno).
849
850On some systems (in general, DOS- and Windows-based systems)
851L<C<binmode>|/binmode FILEHANDLE, LAYER> is necessary when you're not
852working with a text file.  For the sake of portability it is a good idea
853always to use it when appropriate, and never to use it when it isn't
854appropriate.  Also, people can set their I/O to be by default
855UTF8-encoded Unicode, not bytes.
856
857In other words: regardless of platform, use
858L<C<binmode>|/binmode FILEHANDLE, LAYER> on binary data, like images,
859for example.
860
861If LAYER is present it is a single string, but may contain multiple
862directives.  The directives alter the behaviour of the filehandle.
863When LAYER is present, using binmode on a text file makes sense.
864
865If LAYER is omitted or specified as C<:raw> the filehandle is made
866suitable for passing binary data.  This includes turning off possible CRLF
867translation and marking it as bytes (as opposed to Unicode characters).
868Note that, despite what may be implied in I<"Programming Perl"> (the
869Camel, 3rd edition) or elsewhere, C<:raw> is I<not> simply the inverse of C<:crlf>.
870Other layers that would affect the binary nature of the stream are
871I<also> disabled.  See L<PerlIO>, and the discussion about the PERLIO
872environment variable in L<perlrun|perlrun/PERLIO>.
873
874The C<:bytes>, C<:crlf>, C<:utf8>, and any other directives of the
875form C<:...>, are called I/O I<layers>.  The L<open> pragma can be used to
876establish default I/O layers.
877
878I<The LAYER parameter of the L<C<binmode>|/binmode FILEHANDLE, LAYER>
879function is described as "DISCIPLINE" in "Programming Perl, 3rd
880Edition".  However, since the publishing of this book, by many known as
881"Camel III", the consensus of the naming of this functionality has moved
882from "discipline" to "layer".  All documentation of this version of Perl
883therefore refers to "layers" rather than to "disciplines".  Now back to
884the regularly scheduled documentation...>
885
886To mark FILEHANDLE as UTF-8, use C<:utf8> or C<:encoding(UTF-8)>.
887C<:utf8> just marks the data as UTF-8 without further checking,
888while C<:encoding(UTF-8)> checks the data for actually being valid
889UTF-8.  More details can be found in L<PerlIO::encoding>.
890
891In general, L<C<binmode>|/binmode FILEHANDLE, LAYER> should be called
892after L<C<open>|/open FILEHANDLE,MODE,EXPR> but before any I/O is done on the
893filehandle.  Calling L<C<binmode>|/binmode FILEHANDLE, LAYER> normally
894flushes any pending buffered output data (and perhaps pending input
895data) on the handle.  An exception to this is the C<:encoding> layer
896that changes the default character encoding of the handle.
897The C<:encoding> layer sometimes needs to be called in
898mid-stream, and it doesn't flush the stream.  C<:encoding>
899also implicitly pushes on top of itself the C<:utf8> layer because
900internally Perl operates on UTF8-encoded Unicode characters.
901
902The operating system, device drivers, C libraries, and Perl run-time
903system all conspire to let the programmer treat a single
904character (C<\n>) as the line terminator, irrespective of external
905representation.  On many operating systems, the native text file
906representation matches the internal representation, but on some
907platforms the external representation of C<\n> is made up of more than
908one character.
909
910All variants of Unix, Mac OS (old and new), and Stream_LF files on VMS use
911a single character to end each line in the external representation of text
912(even though that single character is CARRIAGE RETURN on old, pre-Darwin
913flavors of Mac OS, and is LINE FEED on Unix and most VMS files).  In other
914systems like OS/2, DOS, and the various flavors of MS-Windows, your program
915sees a C<\n> as a simple C<\cJ>, but what's stored in text files are the
916two characters C<\cM\cJ>.  That means that if you don't use
917L<C<binmode>|/binmode FILEHANDLE, LAYER> on these systems, C<\cM\cJ>
918sequences on disk will be converted to C<\n> on input, and any C<\n> in
919your program will be converted back to C<\cM\cJ> on output.  This is
920what you want for text files, but it can be disastrous for binary files.
921
922Another consequence of using L<C<binmode>|/binmode FILEHANDLE, LAYER>
923(on some systems) is that special end-of-file markers will be seen as
924part of the data stream.  For systems from the Microsoft family this
925means that, if your binary data contain C<\cZ>, the I/O subsystem will
926regard it as the end of the file, unless you use
927L<C<binmode>|/binmode FILEHANDLE, LAYER>.
928
929L<C<binmode>|/binmode FILEHANDLE, LAYER> is important not only for
930L<C<readline>|/readline EXPR> and L<C<print>|/print FILEHANDLE LIST>
931operations, but also when using
932L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET>,
933L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
934L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET>,
935L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET> and
936L<C<tell>|/tell FILEHANDLE> (see L<perlport> for more details).  See the
937L<C<$E<sol>>|perlvar/$E<sol>> and L<C<$\>|perlvar/$\> variables in
938L<perlvar> for how to manually set your input and output
939line-termination sequences.
940
941Portability issues: L<perlport/binmode>.
942
943=item bless REF,CLASSNAME
944X<bless>
945
946=item bless REF
947
948=for Pod::Functions create an object
949
950C<bless> tells Perl to mark the item referred to by C<REF> as an
951object in a package.  The two-argument version of C<bless> is
952always preferable unless there is a specific reason to I<not>
953use it.
954
955=over
956
957=item * Bless the referred-to item into a specific package
958(recommended form):
959
960    bless $ref, $package;
961
962The two-argument form adds the object to the package specified
963as the second argument.
964
965=item * Bless the referred-to item into package C<main>:
966
967    bless $ref, "";
968
969If the second argument is an empty string, C<bless> adds the
970object to package C<main>.
971
972=item * Bless the referred-to item into the current package (not
973inheritable):
974
975    bless $ref;
976
977If C<bless> is used without its second argument, the object is
978created in the current package. The second argument should
979always be supplied if a derived class might inherit a method
980executing C<bless>. Because it is a potential source of bugs,
981one-argument C<bless> is discouraged.
982
983=back
984
985See L<perlobj> for more about the blessing (and blessings) of
986objects.
987
988L<C<bless>|/bless REF,CLASSNAME> returns its first argument, the
989supplied reference, as the value of the function; since C<bless>
990is commonly the last thing executed in constructors, this means
991that the reference to the object is returned as the
992constructor's value and allows the caller to immediately use
993this returned object in method calls.
994
995C<CLASSNAME> should always be a mixed-case name, as
996all-uppercase and all-lowercase names are meant to be used only
997for Perl builtin types and pragmas, respectively. Avoid creating
998all-uppercase or all-lowercase package names to prevent
999confusion.
1000
1001Also avoid C<bless>ing things into the class name C<0>; this
1002will cause code which (erroneously) checks the result of
1003L<C<ref>|/ref EXPR> to see if a reference is C<bless>ed to fail,
1004as "0", a falsy value, is returned.
1005
1006See L<perlmod/"Perl Modules"> for more details.
1007
1008=item break
1009
1010=for Pod::Functions +switch break out of a C<given> block
1011
1012Break out of a C<given> block.
1013
1014L<C<break>|/break> is available only if the
1015L<C<"switch"> feature|feature/The 'switch' feature> is enabled or if it
1016is prefixed with C<CORE::>. The
1017L<C<"switch"> feature|feature/The 'switch' feature> is enabled
1018automatically with a C<use v5.10> (or higher) declaration in the current
1019scope.
1020
1021=item caller EXPR
1022X<caller> X<call stack> X<stack> X<stack trace>
1023
1024=item caller
1025
1026=for Pod::Functions get context of the current subroutine call
1027
1028Returns the context of the current pure perl subroutine call.  In scalar
1029context, returns the caller's package name if there I<is> a caller (that is, if
1030we're in a subroutine or L<C<eval>|/eval EXPR> or
1031L<C<require>|/require VERSION>) and the undefined value otherwise.
1032C<caller> never returns XS subs and they are skipped.  The next pure perl
1033sub will appear instead of the XS sub in caller's return values.  In
1034list context, caller returns
1035
1036       # 0         1          2
1037    my ($package, $filename, $line) = caller;
1038
1039Like L<C<__FILE__>|/__FILE__> and L<C<__LINE__>|/__LINE__>, the filename and
1040line number returned here may be altered by the mechanism described at
1041L<perlsyn/"Plain Old Comments (Not!)">.
1042
1043With EXPR, it returns some extra information that the debugger uses to
1044print a stack trace.  The value of EXPR indicates how many call frames
1045to go back before the current one.
1046
1047    #  0         1          2      3            4
1048 my ($package, $filename, $line, $subroutine, $hasargs,
1049
1050    #  5          6          7            8       9         10
1051    $wantarray, $evaltext, $is_require, $hints, $bitmask, $hinthash)
1052  = caller($i);
1053
1054Here, $subroutine is the function that the caller called (rather than the
1055function containing the caller).  Note that $subroutine may be C<(eval)> if
1056the frame is not a subroutine call, but an L<C<eval>|/eval EXPR>.  In
1057such a case additional elements $evaltext and C<$is_require> are set:
1058C<$is_require> is true if the frame is created by a
1059L<C<require>|/require VERSION> or L<C<use>|/use Module VERSION LIST>
1060statement, $evaltext contains the text of the C<eval EXPR> statement.
1061In particular, for an C<eval BLOCK> statement, $subroutine is C<(eval)>,
1062but $evaltext is undefined.  (Note also that each
1063L<C<use>|/use Module VERSION LIST> statement creates a
1064L<C<require>|/require VERSION> frame inside an C<eval EXPR> frame.)
1065$subroutine may also be C<(unknown)> if this particular subroutine
1066happens to have been deleted from the symbol table.  C<$hasargs> is true
1067if a new instance of L<C<@_>|perlvar/@_> was set up for the frame.
1068C<$hints> and C<$bitmask> contain pragmatic hints that the caller was
1069compiled with.  C<$hints> corresponds to L<C<$^H>|perlvar/$^H>, and
1070C<$bitmask> corresponds to
1071L<C<${^WARNING_BITS}>|perlvar/${^WARNING_BITS}>.  The C<$hints> and
1072C<$bitmask> values are subject to change between versions of Perl, and
1073are not meant for external use.
1074
1075C<$hinthash> is a reference to a hash containing the value of
1076L<C<%^H>|perlvar/%^H> when the caller was compiled, or
1077L<C<undef>|/undef EXPR> if L<C<%^H>|perlvar/%^H> was empty.  Do not
1078modify the values of this hash, as they are the actual values stored in
1079the optree.
1080
1081Note that the only types of call frames that are visible are subroutine
1082calls and C<eval>. Other forms of context, such as C<while> or C<foreach>
1083loops or C<try> blocks are not considered interesting to C<caller>, as they
1084do not alter the behaviour of the C<return> expression.
1085
1086Furthermore, when called from within the DB package in
1087list context, and with an argument, caller returns more
1088detailed information: it sets the list variable C<@DB::args> to be the
1089arguments with which the subroutine was invoked.
1090
1091Be aware that the optimizer might have optimized call frames away before
1092L<C<caller>|/caller EXPR> had a chance to get the information.  That
1093means that C<caller(N)> might not return information about the call
1094frame you expect it to, for C<< N > 1 >>.  In particular, C<@DB::args>
1095might have information from the previous time L<C<caller>|/caller EXPR>
1096was called.
1097
1098Be aware that setting C<@DB::args> is I<best effort>, intended for
1099debugging or generating backtraces, and should not be relied upon.  In
1100particular, as L<C<@_>|perlvar/@_> contains aliases to the caller's
1101arguments, Perl does not take a copy of L<C<@_>|perlvar/@_>, so
1102C<@DB::args> will contain modifications the subroutine makes to
1103L<C<@_>|perlvar/@_> or its contents, not the original values at call
1104time.  C<@DB::args>, like L<C<@_>|perlvar/@_>, does not hold explicit
1105references to its elements, so under certain cases its elements may have
1106become freed and reallocated for other variables or temporary values.
1107Finally, a side effect of the current implementation is that the effects
1108of C<shift @_> can I<normally> be undone (but not C<pop @_> or other
1109splicing, I<and> not if a reference to L<C<@_>|perlvar/@_> has been
1110taken, I<and> subject to the caveat about reallocated elements), so
1111C<@DB::args> is actually a hybrid of the current state and initial state
1112of L<C<@_>|perlvar/@_>.  Buyer beware.
1113
1114=item chdir EXPR
1115X<chdir>
1116X<cd>
1117X<directory, change>
1118
1119=item chdir FILEHANDLE
1120
1121=item chdir DIRHANDLE
1122
1123=item chdir
1124
1125=for Pod::Functions change your current working directory
1126
1127Changes the working directory to EXPR, if possible.  If EXPR is omitted,
1128changes to the directory specified by C<$ENV{HOME}>, if set; if not,
1129changes to the directory specified by C<$ENV{LOGDIR}>.  (Under VMS, the
1130variable C<$ENV{'SYS$LOGIN'}> is also checked, and used if it is set.)  If
1131neither is set, L<C<chdir>|/chdir EXPR> does nothing and fails.  It
1132returns true on success, false otherwise.  See the example under
1133L<C<die>|/die LIST>.
1134
1135On systems that support L<fchdir(2)>, you may pass a filehandle or
1136directory handle as the argument.  On systems that don't support L<fchdir(2)>,
1137passing handles raises an exception.
1138
1139=item chmod LIST
1140X<chmod> X<permission> X<mode>
1141
1142=for Pod::Functions changes the permissions on a list of files
1143
1144Changes the permissions of a list of files.  The first element of the
1145list must be the numeric mode, which should probably be an octal
1146number, and which definitely should I<not> be a string of octal digits:
1147C<0644> is okay, but C<"0644"> is not.  Returns the number of files
1148successfully changed.  See also L<C<oct>|/oct EXPR> if all you have is a
1149string.
1150
1151    my $cnt = chmod 0755, "foo", "bar";
1152    chmod 0755, @executables;
1153    my $mode = "0644"; chmod $mode, "foo";      # !!! sets mode to
1154                                                # --w----r-T
1155    my $mode = "0644"; chmod oct($mode), "foo"; # this is better
1156    my $mode = 0644;   chmod $mode, "foo";      # this is best
1157
1158On systems that support L<fchmod(2)>, you may pass filehandles among the
1159files.  On systems that don't support L<fchmod(2)>, passing filehandles raises
1160an exception.  Filehandles must be passed as globs or glob references to be
1161recognized; barewords are considered filenames.
1162
1163    open(my $fh, "<", "foo");
1164    my $perm = (stat $fh)[2] & 07777;
1165    chmod($perm | 0600, $fh);
1166
1167You can also import the symbolic C<S_I*> constants from the
1168L<C<Fcntl>|Fcntl> module:
1169
1170    use Fcntl qw( :mode );
1171    chmod S_IRWXU|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH, @executables;
1172    # Identical to the chmod 0755 of the example above.
1173
1174Portability issues: L<perlport/chmod>.
1175
1176=item chomp VARIABLE
1177X<chomp> X<INPUT_RECORD_SEPARATOR> X<$/> X<newline> X<eol>
1178
1179=item chomp( LIST )
1180
1181=item chomp
1182
1183=for Pod::Functions remove a trailing record separator from a string
1184
1185This safer version of L<C<chop>|/chop VARIABLE> removes any trailing
1186string that corresponds to the current value of
1187L<C<$E<sol>>|perlvar/$E<sol>> (also known as C<$INPUT_RECORD_SEPARATOR>
1188in the L<C<English>|English> module).  It returns the total
1189number of characters removed from all its arguments.  It's often used to
1190remove the newline from the end of an input record when you're worried
1191that the final record may be missing its newline.  When in paragraph
1192mode (C<$/ = ''>), it removes all trailing newlines from the string.
1193When in slurp mode (C<$/ = undef>) or fixed-length record mode
1194(L<C<$E<sol>>|perlvar/$E<sol>> is a reference to an integer or the like;
1195see L<perlvar>), L<C<chomp>|/chomp VARIABLE> won't remove anything.
1196If VARIABLE is omitted, it chomps L<C<$_>|perlvar/$_>.  Example:
1197
1198    while (<>) {
1199        chomp;  # avoid \n on last field
1200        my @array = split(/:/);
1201        # ...
1202    }
1203
1204If VARIABLE is a hash, it chomps the hash's values, but not its keys,
1205resetting the L<C<each>|/each HASH> iterator in the process.
1206
1207You can actually chomp anything that's an lvalue, including an assignment:
1208
1209    chomp(my $cwd = `pwd`);
1210    chomp(my $answer = <STDIN>);
1211
1212If you chomp a list, each element is chomped, and the total number of
1213characters removed is returned.
1214
1215Note that parentheses are necessary when you're chomping anything
1216that is not a simple variable.  This is because C<chomp $cwd = `pwd`;>
1217is interpreted as C<(chomp $cwd) = `pwd`;>, rather than as
1218C<chomp( $cwd = `pwd` )> which you might expect.  Similarly,
1219C<chomp $x, $y> is interpreted as C<chomp($x), $y> rather than
1220as C<chomp($x, $y)>.
1221
1222=item chop VARIABLE
1223X<chop>
1224
1225=item chop( LIST )
1226
1227=item chop
1228
1229=for Pod::Functions remove the last character from a string
1230
1231Chops off the last character of a string and returns the character
1232chopped.  It is much more efficient than C<s/.$//s> because it neither
1233scans nor copies the string.  If VARIABLE is omitted, chops
1234L<C<$_>|perlvar/$_>.
1235If VARIABLE is a hash, it chops the hash's values, but not its keys,
1236resetting the L<C<each>|/each HASH> iterator in the process.
1237
1238You can actually chop anything that's an lvalue, including an assignment.
1239
1240If you chop a list, each element is chopped.  Only the value of the
1241last L<C<chop>|/chop VARIABLE> is returned.
1242
1243Note that L<C<chop>|/chop VARIABLE> returns the last character.  To
1244return all but the last character, use C<substr($string, 0, -1)>.
1245
1246See also L<C<chomp>|/chomp VARIABLE>.
1247
1248=item chown LIST
1249X<chown> X<owner> X<user> X<group>
1250
1251=for Pod::Functions change the ownership on a list of files
1252
1253Changes the owner (and group) of a list of files.  The first two
1254elements of the list must be the I<numeric> uid and gid, in that
1255order.  A value of -1 in either position is interpreted by most
1256systems to leave that value unchanged.  Returns the number of files
1257successfully changed.
1258
1259    my $cnt = chown $uid, $gid, 'foo', 'bar';
1260    chown $uid, $gid, @filenames;
1261
1262On systems that support L<fchown(2)>, you may pass filehandles among the
1263files.  On systems that don't support L<fchown(2)>, passing filehandles raises
1264an exception.  Filehandles must be passed as globs or glob references to be
1265recognized; barewords are considered filenames.
1266
1267Here's an example that looks up nonnumeric uids in the passwd file:
1268
1269    print "User: ";
1270    chomp(my $user = <STDIN>);
1271    print "Files: ";
1272    chomp(my $pattern = <STDIN>);
1273
1274    my ($login,$pass,$uid,$gid) = getpwnam($user)
1275        or die "$user not in passwd file";
1276
1277    my @ary = glob($pattern);  # expand filenames
1278    chown $uid, $gid, @ary;
1279
1280On most systems, you are not allowed to change the ownership of the
1281file unless you're the superuser, although you should be able to change
1282the group to any of your secondary groups.  On insecure systems, these
1283restrictions may be relaxed, but this is not a portable assumption.
1284On POSIX systems, you can detect this condition this way:
1285
1286    use POSIX qw(pathconf _PC_CHOWN_RESTRICTED);
1287    my $can_chown_giveaway =
1288        ! pathconf($path_of_interest, _PC_CHOWN_RESTRICTED);
1289
1290Portability issues: L<perlport/chown>.
1291
1292=item chr NUMBER
1293X<chr> X<character> X<ASCII> X<Unicode>
1294
1295=item chr
1296
1297=for Pod::Functions get character this number represents
1298
1299Returns the character represented by that NUMBER in the character set.
1300For example, C<chr(65)> is C<"A"> in either ASCII or Unicode, and
1301C<chr(0x263a)> is a Unicode smiley face.
1302
1303Negative values give the Unicode replacement character (C<chr(0xfffd)>),
1304except under the L<bytes> pragma, where the low eight bits of the value
1305(truncated to an integer) are used.
1306
1307If NUMBER is omitted, uses L<C<$_>|perlvar/$_>.
1308
1309For the reverse, use L<C<ord>|/ord EXPR>.
1310
1311Note that characters from 128 to 255 (inclusive) are by default
1312internally not encoded as UTF-8 for backward compatibility reasons.
1313
1314See L<perlunicode> for more about Unicode.
1315
1316=item chroot FILENAME
1317X<chroot> X<root>
1318
1319=item chroot
1320
1321=for Pod::Functions make directory new root for path lookups
1322
1323This function works like the system call by the same name: it makes the
1324named directory the new root directory for all further pathnames that
1325begin with a C</> by your process and all its children.  (It doesn't
1326change your current working directory, which is unaffected.)  For security
1327reasons, this call is restricted to the superuser.  If FILENAME is
1328omitted, does a L<C<chroot>|/chroot FILENAME> to L<C<$_>|perlvar/$_>.
1329
1330B<NOTE:>  It is mandatory for security to C<chdir("/")>
1331(L<C<chdir>|/chdir EXPR> to the root directory) immediately after a
1332L<C<chroot>|/chroot FILENAME>, otherwise the current working directory
1333may be outside of the new root.
1334
1335Portability issues: L<perlport/chroot>.
1336
1337=item class NAMESPACE
1338
1339=item class NAMESPACE VERSION
1340
1341=item class NAMESPACE BLOCK
1342
1343=item class NAMESPACE VERSION BLOCK
1344
1345=for Pod::Functions declare a separate global namespace that is an object class
1346
1347Declares the BLOCK or the rest of the compilation unit as being in the given
1348namespace, which implements an object class.  This behaves similarly to
1349L<C<package>|/package NAMESPACE>, except that the newly-created package behaves
1350as a class.
1351
1352=item close FILEHANDLE
1353X<close>
1354
1355=item close
1356
1357=for Pod::Functions close file (or pipe or socket) handle
1358
1359Closes the file or pipe associated with the filehandle, flushes the IO
1360buffers, and closes the system file descriptor.  Returns true if those
1361operations succeed, and if no error was reported by any PerlIO layer,
1362and there was no existing error on the filehandle.
1363
1364If there was an existing error on the filehandle, close will return
1365false and L<C<$!>|perlvar/$!> will be set to the error from the
1366failing operation, so you can safely use its value when reporting the
1367error.
1368
1369Closes the currently selected filehandle if the argument is
1370omitted.
1371
1372You don't have to close FILEHANDLE if you are immediately going to do
1373another L<C<open>|/open FILEHANDLE,MODE,EXPR> on it, because
1374L<C<open>|/open FILEHANDLE,MODE,EXPR> closes it for you.  (See
1375L<C<open>|/open FILEHANDLE,MODE,EXPR>.) However, an explicit
1376L<C<close>|/close FILEHANDLE> on an input file resets the line counter
1377(L<C<$.>|perlvar/$.>), while the implicit close done by
1378L<C<open>|/open FILEHANDLE,MODE,EXPR> does not.
1379
1380If the filehandle came from a piped open, L<C<close>|/close FILEHANDLE>
1381returns false if one of the other syscalls involved fails or if its
1382program exits with non-zero status.  If the only problem was that the
1383program exited non-zero, L<C<$!>|perlvar/$!> will be set to C<0>.
1384Closing a pipe also waits for the process executing on the pipe to
1385exit--in case you wish to look at the output of the pipe afterwards--and
1386implicitly puts the exit status value of that command into
1387L<C<$?>|perlvar/$?> and
1388L<C<${^CHILD_ERROR_NATIVE}>|perlvar/${^CHILD_ERROR_NATIVE}>.
1389
1390If there are multiple threads running, L<C<close>|/close FILEHANDLE> on
1391a filehandle from a piped open returns true without waiting for the
1392child process to terminate, if the filehandle is still open in another
1393thread.
1394
1395Closing the read end of a pipe before the process writing to it at the
1396other end is done writing results in the writer receiving a SIGPIPE.  If
1397the other end can't handle that, be sure to read all the data before
1398closing the pipe.
1399
1400Example:
1401
1402    open(OUTPUT, '|sort >foo')  # pipe to sort
1403        or die "Can't start sort: $!";
1404    #...                        # print stuff to output
1405    close OUTPUT                # wait for sort to finish
1406        or warn $! ? "Error closing sort pipe: $!"
1407                   : "Exit status $? from sort";
1408    open(INPUT, 'foo')          # get sort's results
1409        or die "Can't open 'foo' for input: $!";
1410
1411FILEHANDLE may be an expression whose value can be used as an indirect
1412filehandle, usually the real filehandle name or an autovivified handle.
1413
1414If an error occurs when perl implicitly closes a handle, perl will
1415produce a L<warning|perldiag/"Warning: unable to close filehandle %s
1416properly: %s">.  Explicitly calling close on the handle prevents that
1417warning.
1418
1419=item closedir DIRHANDLE
1420X<closedir>
1421
1422=for Pod::Functions close directory handle
1423
1424Closes a directory opened by L<C<opendir>|/opendir DIRHANDLE,EXPR> and
1425returns the success of that system call.
1426
1427=item connect SOCKET,NAME
1428X<connect>
1429
1430=for Pod::Functions connect to a remote socket
1431
1432Attempts to connect to a remote socket, just like L<connect(2)>.
1433Returns true if it succeeded, false otherwise.  NAME should be a
1434packed address of the appropriate type for the socket.  See the examples in
1435L<perlipc/"Sockets: Client/Server Communication">.
1436
1437=item continue BLOCK
1438X<continue>
1439
1440=item continue
1441
1442=for Pod::Functions optional trailing block in a while or foreach
1443
1444When followed by a BLOCK, L<C<continue>|/continue BLOCK> is actually a
1445flow control statement rather than a function.  If there is a
1446L<C<continue>|/continue BLOCK> BLOCK attached to a BLOCK (typically in a
1447C<while> or C<foreach>), it is always executed just before the
1448conditional is about to be evaluated again, just like the third part of
1449a C<for> loop in C.  Thus it can be used to increment a loop variable,
1450even when the loop has been continued via the L<C<next>|/next LABEL>
1451statement (which is similar to the C L<C<continue>|/continue BLOCK>
1452statement).
1453
1454L<C<last>|/last LABEL>, L<C<next>|/next LABEL>, or
1455L<C<redo>|/redo LABEL> may appear within a
1456L<C<continue>|/continue BLOCK> block; L<C<last>|/last LABEL> and
1457L<C<redo>|/redo LABEL> behave as if they had been executed within the
1458main block.  So will L<C<next>|/next LABEL>, but since it will execute a
1459L<C<continue>|/continue BLOCK> block, it may be more entertaining.
1460
1461    while (EXPR) {
1462        ### redo always comes here
1463        do_something;
1464    } continue {
1465        ### next always comes here
1466        do_something_else;
1467        # then back to the top to re-check EXPR
1468    }
1469    ### last always comes here
1470
1471Omitting the L<C<continue>|/continue BLOCK> section is equivalent to
1472using an empty one, logically enough, so L<C<next>|/next LABEL> goes
1473directly back to check the condition at the top of the loop.
1474
1475When there is no BLOCK, L<C<continue>|/continue BLOCK> is a function
1476that falls through the current C<when> or C<default> block instead of
1477iterating a dynamically enclosing C<foreach> or exiting a lexically
1478enclosing C<given>.  In Perl 5.14 and earlier, this form of
1479L<C<continue>|/continue BLOCK> was only available when the
1480L<C<"switch"> feature|feature/The 'switch' feature> was enabled.  See
1481L<feature> and L<perlsyn/"Switch Statements"> for more information.
1482
1483=item cos EXPR
1484X<cos> X<cosine> X<acos> X<arccosine>
1485
1486=item cos
1487
1488=for Pod::Functions cosine function
1489
1490Returns the cosine of EXPR (expressed in radians).  If EXPR is omitted,
1491takes the cosine of L<C<$_>|perlvar/$_>.
1492
1493For the inverse cosine operation, you may use the
1494L<C<Math::Trig::acos>|Math::Trig> function, or use this relation:
1495
1496    sub acos { atan2( sqrt(1 - $_[0] * $_[0]), $_[0] ) }
1497
1498=item crypt PLAINTEXT,SALT
1499X<crypt> X<digest> X<hash> X<salt> X<plaintext> X<password>
1500X<decrypt> X<cryptography> X<passwd> X<encrypt>
1501
1502=for Pod::Functions one-way passwd-style encryption
1503
1504Creates a digest string exactly like the L<crypt(3)> function in the C
1505library (assuming that you actually have a version there that has not
1506been extirpated as a potential munition).
1507
1508L<C<crypt>|/crypt PLAINTEXT,SALT> is a one-way hash function.  The
1509PLAINTEXT and SALT are turned
1510into a short string, called a digest, which is returned.  The same
1511PLAINTEXT and SALT will always return the same string, but there is no
1512(known) way to get the original PLAINTEXT from the hash.  Small
1513changes in the PLAINTEXT or SALT will result in large changes in the
1514digest.
1515
1516There is no decrypt function.  This function isn't all that useful for
1517cryptography (for that, look for F<Crypt> modules on your nearby CPAN
1518mirror) and the name "crypt" is a bit of a misnomer.  Instead it is
1519primarily used to check if two pieces of text are the same without
1520having to transmit or store the text itself.  An example is checking
1521if a correct password is given.  The digest of the password is stored,
1522not the password itself.  The user types in a password that is
1523L<C<crypt>|/crypt PLAINTEXT,SALT>'d with the same salt as the stored
1524digest.  If the two digests match, the password is correct.
1525
1526When verifying an existing digest string you should use the digest as
1527the salt (like C<crypt($plain, $digest) eq $digest>).  The SALT used
1528to create the digest is visible as part of the digest.  This ensures
1529L<C<crypt>|/crypt PLAINTEXT,SALT> will hash the new string with the same
1530salt as the digest.  This allows your code to work with the standard
1531L<C<crypt>|/crypt PLAINTEXT,SALT> and with more exotic implementations.
1532In other words, assume nothing about the returned string itself nor
1533about how many bytes of SALT may matter.
1534
1535Traditionally the result is a string of 13 bytes: two first bytes of
1536the salt, followed by 11 bytes from the set C<[./0-9A-Za-z]>, and only
1537the first eight bytes of PLAINTEXT mattered.  But alternative
1538hashing schemes (like MD5), higher level security schemes (like C2),
1539and implementations on non-Unix platforms may produce different
1540strings.
1541
1542When choosing a new salt create a random two character string whose
1543characters come from the set C<[./0-9A-Za-z]> (like C<join '', ('.',
1544'/', 0..9, 'A'..'Z', 'a'..'z')[rand 64, rand 64]>).  This set of
1545characters is just a recommendation; the characters allowed in
1546the salt depend solely on your system's crypt library, and Perl can't
1547restrict what salts L<C<crypt>|/crypt PLAINTEXT,SALT> accepts.
1548
1549Here's an example that makes sure that whoever runs this program knows
1550their password:
1551
1552    my $pwd = (getpwuid($<))[1];
1553
1554    system "stty -echo";
1555    print "Password: ";
1556    chomp(my $word = <STDIN>);
1557    print "\n";
1558    system "stty echo";
1559
1560    if (crypt($word, $pwd) ne $pwd) {
1561        die "Sorry...\n";
1562    } else {
1563        print "ok\n";
1564    }
1565
1566Of course, typing in your own password to whoever asks you
1567for it is unwise.
1568
1569The L<C<crypt>|/crypt PLAINTEXT,SALT> function is unsuitable for hashing
1570large quantities of data, not least of all because you can't get the
1571information back.  Look at the L<Digest> module for more robust
1572algorithms.
1573
1574If using L<C<crypt>|/crypt PLAINTEXT,SALT> on a Unicode string (which
1575I<potentially> has characters with codepoints above 255), Perl tries to
1576make sense of the situation by trying to downgrade (a copy of) the
1577string back to an eight-bit byte string before calling
1578L<C<crypt>|/crypt PLAINTEXT,SALT> (on that copy).  If that works, good.
1579If not, L<C<crypt>|/crypt PLAINTEXT,SALT> dies with
1580L<C<Wide character in crypt>|perldiag/Wide character in %s>.
1581
1582Portability issues: L<perlport/crypt>.
1583
1584=item dbmclose HASH
1585X<dbmclose>
1586
1587=for Pod::Functions breaks binding on a tied dbm file
1588
1589[This function has been largely superseded by the
1590L<C<untie>|/untie VARIABLE> function.]
1591
1592Breaks the binding between a DBM file and a hash.
1593
1594Portability issues: L<perlport/dbmclose>.
1595
1596=item dbmopen HASH,DBNAME,MASK
1597X<dbmopen> X<dbm> X<ndbm> X<sdbm> X<gdbm>
1598
1599=for Pod::Functions create binding on a tied dbm file
1600
1601[This function has been largely superseded by the
1602L<C<tie>|/tie VARIABLE,CLASSNAME,LIST> function.]
1603
1604This binds a L<dbm(3)>, L<ndbm(3)>, L<sdbm(3)>, L<gdbm(3)>, or Berkeley
1605DB file to a hash.  HASH is the name of the hash.  (Unlike normal
1606L<C<open>|/open FILEHANDLE,MODE,EXPR>, the first argument is I<not> a
1607filehandle, even though it looks like one).  DBNAME is the name of the
1608database (without the F<.dir> or F<.pag> extension if any).  If the
1609database does not exist, it is created with protection specified by MASK
1610(as modified by the L<C<umask>|/umask EXPR>).  To prevent creation of
1611the database if it doesn't exist, you may specify a MASK of 0, and the
1612function will return a false value if it can't find an existing
1613database.  If your system supports only the older DBM functions, you may
1614make only one L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK> call in your
1615program.  In older versions of Perl, if your system had neither DBM nor
1616ndbm, calling L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK> produced a fatal
1617error; it now falls back to L<sdbm(3)>.
1618
1619If you don't have write access to the DBM file, you can only read hash
1620variables, not set them.  If you want to test whether you can write,
1621either use file tests or try setting a dummy hash entry inside an
1622L<C<eval>|/eval EXPR> to trap the error.
1623
1624Note that functions such as L<C<keys>|/keys HASH> and
1625L<C<values>|/values HASH> may return huge lists when used on large DBM
1626files.  You may prefer to use the L<C<each>|/each HASH> function to
1627iterate over large DBM files.  Example:
1628
1629    # print out history file offsets
1630    dbmopen(%HIST,'/usr/lib/news/history',0666);
1631    while (($key,$val) = each %HIST) {
1632        print $key, ' = ', unpack('L',$val), "\n";
1633    }
1634    dbmclose(%HIST);
1635
1636See also L<AnyDBM_File> for a more general description of the pros and
1637cons of the various dbm approaches, as well as L<DB_File> for a particularly
1638rich implementation.
1639
1640You can control which DBM library you use by loading that library
1641before you call L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK>:
1642
1643    use DB_File;
1644    dbmopen(%NS_Hist, "$ENV{HOME}/.netscape/history.db")
1645        or die "Can't open netscape history file: $!";
1646
1647Portability issues: L<perlport/dbmopen>.
1648
1649=item defined EXPR
1650X<defined> X<undef> X<undefined>
1651
1652=item defined
1653
1654=for Pod::Functions test whether a value, variable, or function is defined
1655
1656Returns a Boolean value telling whether EXPR has a value other than the
1657undefined value L<C<undef>|/undef EXPR>.  If EXPR is not present,
1658L<C<$_>|perlvar/$_> is checked.
1659
1660Many operations return L<C<undef>|/undef EXPR> to indicate failure, end
1661of file, system error, uninitialized variable, and other exceptional
1662conditions.  This function allows you to distinguish
1663L<C<undef>|/undef EXPR> from other values.  (A simple Boolean test will
1664not distinguish among L<C<undef>|/undef EXPR>, zero, the empty string,
1665and C<"0">, which are all equally false.)  Note that since
1666L<C<undef>|/undef EXPR> is a valid scalar, its presence doesn't
1667I<necessarily> indicate an exceptional condition: L<C<pop>|/pop ARRAY>
1668returns L<C<undef>|/undef EXPR> when its argument is an empty array,
1669I<or> when the element to return happens to be L<C<undef>|/undef EXPR>.
1670
1671You may also use C<defined(&func)> to check whether subroutine C<func>
1672has ever been defined.  The return value is unaffected by any forward
1673declarations of C<func>.  A subroutine that is not defined
1674may still be callable: its package may have an C<AUTOLOAD> method that
1675makes it spring into existence the first time that it is called; see
1676L<perlsub>.
1677
1678Use of L<C<defined>|/defined EXPR> on aggregates (hashes and arrays) is
1679no longer supported. It used to report whether memory for that
1680aggregate had ever been allocated.  You should instead use a simple
1681test for size:
1682
1683    if (@an_array) { print "has array elements\n" }
1684    if (%a_hash)   { print "has hash members\n"   }
1685
1686When used on a hash element, it tells you whether the value is defined,
1687not whether the key exists in the hash.  Use L<C<exists>|/exists EXPR>
1688for the latter purpose.
1689
1690Examples:
1691
1692    print if defined $switch{D};
1693    print "$val\n" while defined($val = pop(@ary));
1694    die "Can't readlink $sym: $!"
1695        unless defined($value = readlink $sym);
1696    sub foo { defined &$bar ? $bar->(@_) : die "No bar"; }
1697    $debugging = 0 unless defined $debugging;
1698
1699Note:  Many folks tend to overuse L<C<defined>|/defined EXPR> and are
1700then surprised to discover that the number C<0> and C<""> (the
1701zero-length string) are, in fact, defined values.  For example, if you
1702say
1703
1704    "ab" =~ /a(.*)b/;
1705
1706The pattern match succeeds and C<$1> is defined, although it
1707matched "nothing".  It didn't really fail to match anything.  Rather, it
1708matched something that happened to be zero characters long.  This is all
1709very above-board and honest.  When a function returns an undefined value,
1710it's an admission that it couldn't give you an honest answer.  So you
1711should use L<C<defined>|/defined EXPR> only when questioning the
1712integrity of what you're trying to do.  At other times, a simple
1713comparison to C<0> or C<""> is what you want.
1714
1715See also L<C<undef>|/undef EXPR>, L<C<exists>|/exists EXPR>,
1716L<C<ref>|/ref EXPR>.
1717
1718=item delete EXPR
1719X<delete>
1720
1721=for Pod::Functions deletes a value from a hash
1722
1723Given an expression that specifies an element or slice of a hash,
1724L<C<delete>|/delete EXPR> deletes the specified elements from that hash
1725so that L<C<exists>|/exists EXPR> on that element no longer returns
1726true.  Setting a hash element to the undefined value does not remove its
1727key, but deleting it does; see L<C<exists>|/exists EXPR>.
1728
1729In list context, usually returns the value or values deleted, or the last such
1730element in scalar context.  The return list's length corresponds to that of
1731the argument list: deleting non-existent elements returns the undefined value
1732in their corresponding positions. Since Perl 5.28, a
1733L<keyE<sol>value hash slice|perldata/KeyE<sol>Value Hash Slices> can be passed
1734to C<delete>, and the return value is a list of key/value pairs (two elements
1735for each item deleted from the hash).
1736
1737L<C<delete>|/delete EXPR> may also be used on arrays and array slices,
1738but its behavior is less straightforward.  Although
1739L<C<exists>|/exists EXPR> will return false for deleted entries,
1740deleting array elements never changes indices of existing values; use
1741L<C<shift>|/shift ARRAY> or L<C<splice>|/splice
1742ARRAY,OFFSET,LENGTH,LIST> for that.  However, if any deleted elements
1743fall at the end of an array, the array's size shrinks to the position of
1744the highest element that still tests true for L<C<exists>|/exists EXPR>,
1745or to 0 if none do.  In other words, an array won't have trailing
1746nonexistent elements after a delete.
1747
1748B<WARNING:> Calling L<C<delete>|/delete EXPR> on array values is
1749strongly discouraged.  The
1750notion of deleting or checking the existence of Perl array elements is not
1751conceptually coherent, and can lead to surprising behavior.
1752
1753Deleting from L<C<%ENV>|perlvar/%ENV> modifies the environment.
1754Deleting from a hash tied to a DBM file deletes the entry from the DBM
1755file.  Deleting from a L<C<tied>|/tied VARIABLE> hash or array may not
1756necessarily return anything; it depends on the implementation of the
1757L<C<tied>|/tied VARIABLE> package's DELETE method, which may do whatever
1758it pleases.
1759
1760The C<delete local EXPR> construct localizes the deletion to the current
1761block at run time.  Until the block exits, elements locally deleted
1762temporarily no longer exist.  See L<perlsub/"Localized deletion of elements
1763of composite types">.
1764
1765    my %hash = (foo => 11, bar => 22, baz => 33);
1766    my $scalar = delete $hash{foo};         # $scalar is 11
1767    $scalar = delete @hash{qw(foo bar)}; # $scalar is 22
1768    my @array  = delete @hash{qw(foo baz)}; # @array  is (undef,33)
1769
1770The following (inefficiently) deletes all the values of %HASH and @ARRAY:
1771
1772    foreach my $key (keys %HASH) {
1773        delete $HASH{$key};
1774    }
1775
1776    foreach my $index (0 .. $#ARRAY) {
1777        delete $ARRAY[$index];
1778    }
1779
1780And so do these:
1781
1782    delete @HASH{keys %HASH};
1783
1784    delete @ARRAY[0 .. $#ARRAY];
1785
1786But both are slower than assigning the empty list
1787or undefining %HASH or @ARRAY, which is the customary
1788way to empty out an aggregate:
1789
1790    %HASH = ();     # completely empty %HASH
1791    undef %HASH;    # forget %HASH ever existed
1792
1793    @ARRAY = ();    # completely empty @ARRAY
1794    undef @ARRAY;   # forget @ARRAY ever existed
1795
1796The EXPR can be arbitrarily complicated provided its
1797final operation is an element or slice of an aggregate:
1798
1799    delete $ref->[$x][$y]{$key};
1800    delete $ref->[$x][$y]->@{$key1, $key2, @morekeys};
1801
1802    delete $ref->[$x][$y][$index];
1803    delete $ref->[$x][$y]->@[$index1, $index2, @moreindices];
1804
1805=item die LIST
1806X<die> X<throw> X<exception> X<raise> X<$@> X<abort>
1807
1808=for Pod::Functions raise an exception or bail out
1809
1810L<C<die>|/die LIST> raises an exception.  Inside an L<C<eval>|/eval EXPR>
1811the exception is stuffed into L<C<$@>|perlvar/$@> and the L<C<eval>|/eval
1812EXPR> is terminated with the undefined value.  If the exception is
1813outside of all enclosing L<C<eval>|/eval EXPR>s, then the uncaught
1814exception is printed to C<STDERR> and perl exits with an exit code
1815indicating failure.  If you need to exit the process with a specific
1816exit code, see L<C<exit>|/exit EXPR>.
1817
1818Equivalent examples:
1819
1820    die "Can't cd to spool: $!\n" unless chdir '/usr/spool/news';
1821    chdir '/usr/spool/news' or die "Can't cd to spool: $!\n"
1822
1823Most of the time, C<die> is called with a string to use as the exception.
1824You may either give a single non-reference operand to serve as the
1825exception, or a list of two or more items, which will be stringified
1826and concatenated to make the exception.
1827
1828If the string exception does not end in a newline, the current
1829script line number and input line number (if any) and a newline
1830are appended to it.  Note that the "input line number" (also
1831known as "chunk") is subject to whatever notion of "line" happens to
1832be currently in effect, and is also available as the special variable
1833L<C<$.>|perlvar/$.>.  See L<perlvar/"$/"> and L<perlvar/"$.">.
1834
1835Hint: sometimes appending C<", stopped"> to your message will cause it
1836to make better sense when the string C<"at foo line 123"> is appended.
1837Suppose you are running script "canasta".
1838
1839    die "/etc/games is no good";
1840    die "/etc/games is no good, stopped";
1841
1842produce, respectively
1843
1844    /etc/games is no good at canasta line 123.
1845    /etc/games is no good, stopped at canasta line 123.
1846
1847If LIST was empty or made an empty string, and L<C<$@>|perlvar/$@>
1848already contains an exception value (typically from a previous
1849L<C<eval>|/eval EXPR>), then that value is reused after
1850appending C<"\t...propagated">.  This is useful for propagating exceptions:
1851
1852    eval { ... };
1853    die unless $@ =~ /Expected exception/;
1854
1855If LIST was empty or made an empty string,
1856and L<C<$@>|perlvar/$@> contains an object
1857reference that has a C<PROPAGATE> method, that method will be called
1858with additional file and line number parameters.  The return value
1859replaces the value in L<C<$@>|perlvar/$@>;  i.e., as if
1860C<< $@ = eval { $@->PROPAGATE(__FILE__, __LINE__) }; >> were called.
1861
1862If LIST was empty or made an empty string, and L<C<$@>|perlvar/$@>
1863is also empty, then the string C<"Died"> is used.
1864
1865You can also call L<C<die>|/die LIST> with a reference argument, and if
1866this is trapped within an L<C<eval>|/eval EXPR>, L<C<$@>|perlvar/$@>
1867contains that reference.  This permits more elaborate exception handling
1868using objects that maintain arbitrary state about the exception.  Such a
1869scheme is sometimes preferable to matching particular string values of
1870L<C<$@>|perlvar/$@> with regular expressions.
1871
1872Because Perl stringifies uncaught exception messages before display,
1873you'll probably want to overload stringification operations on
1874exception objects.  See L<overload> for details about that.
1875The stringified message should be non-empty, and should end in a newline,
1876in order to fit in with the treatment of string exceptions.
1877Also, because an exception object reference cannot be stringified
1878without destroying it, Perl doesn't attempt to append location or other
1879information to a reference exception.  If you want location information
1880with a complex exception object, you'll have to arrange to put the
1881location information into the object yourself.
1882
1883Because L<C<$@>|perlvar/$@> is a global variable, be careful that
1884analyzing an exception caught by C<eval> doesn't replace the reference
1885in the global variable.  It's
1886easiest to make a local copy of the reference before any manipulations.
1887Here's an example:
1888
1889    use Scalar::Util "blessed";
1890
1891    eval { ... ; die Some::Module::Exception->new( FOO => "bar" ) };
1892    if (my $ev_err = $@) {
1893        if (blessed($ev_err)
1894            && $ev_err->isa("Some::Module::Exception")) {
1895            # handle Some::Module::Exception
1896        }
1897        else {
1898            # handle all other possible exceptions
1899        }
1900    }
1901
1902If an uncaught exception results in interpreter exit, the exit code is
1903determined from the values of L<C<$!>|perlvar/$!> and
1904L<C<$?>|perlvar/$?> with this pseudocode:
1905
1906    exit $! if $!;              # errno
1907    exit $? >> 8 if $? >> 8;    # child exit status
1908    exit 255;                   # last resort
1909
1910As with L<C<exit>|/exit EXPR>, L<C<$?>|perlvar/$?> is set prior to
1911unwinding the call stack; any C<DESTROY> or C<END> handlers can then
1912alter this value, and thus Perl's exit code.
1913
1914The intent is to squeeze as much possible information about the likely cause
1915into the limited space of the system exit code.  However, as
1916L<C<$!>|perlvar/$!> is the value of C's C<errno>, which can be set by
1917any system call, this means that the value of the exit code used by
1918L<C<die>|/die LIST> can be non-predictable, so should not be relied
1919upon, other than to be non-zero.
1920
1921You can arrange for a callback to be run just before the
1922L<C<die>|/die LIST> does its deed, by setting the
1923L<C<$SIG{__DIE__}>|perlvar/%SIG> hook.  The associated handler is called
1924with the exception as an argument, and can change the exception,
1925if it sees fit, by
1926calling L<C<die>|/die LIST> again.  See L<perlvar/%SIG> for details on
1927setting L<C<%SIG>|perlvar/%SIG> entries, and L<C<eval>|/eval EXPR> for some
1928examples.  Although this feature was to be run only right before your
1929program was to exit, this is not currently so: the
1930L<C<$SIG{__DIE__}>|perlvar/%SIG> hook is currently called even inside
1931L<C<eval>|/eval EXPR>ed blocks/strings!  If one wants the hook to do
1932nothing in such situations, put
1933
1934    die @_ if $^S;
1935
1936as the first line of the handler (see L<perlvar/$^S>).  Because
1937this promotes strange action at a distance, this counterintuitive
1938behavior may be fixed in a future release.
1939
1940See also L<C<exit>|/exit EXPR>, L<C<warn>|/warn LIST>, and the L<Carp>
1941module.
1942
1943=item do BLOCK
1944X<do> X<block>
1945
1946=for Pod::Functions turn a BLOCK into a TERM
1947
1948Not really a function.  Returns the value of the last command in the
1949sequence of commands indicated by BLOCK.  When modified by the C<while> or
1950C<until> loop modifier, executes the BLOCK once before testing the loop
1951condition.  (On other statements the loop modifiers test the conditional
1952first.)
1953
1954C<do BLOCK> does I<not> count as a loop, so the loop control statements
1955L<C<next>|/next LABEL>, L<C<last>|/last LABEL>, or
1956L<C<redo>|/redo LABEL> cannot be used to leave or restart the block.
1957See L<perlsyn> for alternative strategies.
1958
1959=item do EXPR
1960X<do>
1961
1962Uses the value of EXPR as a filename and executes the contents of the
1963file as a Perl script:
1964
1965    # load the exact specified file (./ and ../ special-cased)
1966    do '/foo/stat.pl';
1967    do './stat.pl';
1968    do '../foo/stat.pl';
1969
1970    # search for the named file within @INC
1971    do 'stat.pl';
1972    do 'foo/stat.pl';
1973
1974C<do './stat.pl'> is largely like
1975
1976    eval `cat stat.pl`;
1977
1978except that it's more concise, runs no external processes, and keeps
1979track of the current filename for error messages. It also differs in that
1980code evaluated with C<do FILE> cannot see lexicals in the enclosing
1981scope; C<eval STRING> does.  It's the same, however, in that it does
1982reparse the file every time you call it, so you probably don't want
1983to do this inside a loop.
1984
1985Using C<do> with a relative path (except for F<./> and F<../>), like
1986
1987    do 'foo/stat.pl';
1988
1989will search the L<C<@INC>|perlvar/@INC> directories, and update
1990L<C<%INC>|perlvar/%INC> if the file is found.  See L<perlvar/@INC>
1991and L<perlvar/%INC> for these variables. In particular, note that
1992whilst historically L<C<@INC>|perlvar/@INC> contained '.' (the
1993current directory) making these two cases equivalent, that is no
1994longer necessarily the case, as '.' is not included in C<@INC> by default
1995in perl versions 5.26.0 onwards. Instead, perl will now warn:
1996
1997    do "stat.pl" failed, '.' is no longer in @INC;
1998    did you mean do "./stat.pl"?
1999
2000If L<C<do>|/do EXPR> can read the file but cannot compile it, it
2001returns L<C<undef>|/undef EXPR> and sets an error message in
2002L<C<$@>|perlvar/$@>.  If L<C<do>|/do EXPR> cannot read the file, it
2003returns undef and sets L<C<$!>|perlvar/$!> to the error.  Always check
2004L<C<$@>|perlvar/$@> first, as compilation could fail in a way that also
2005sets L<C<$!>|perlvar/$!>.  If the file is successfully compiled,
2006L<C<do>|/do EXPR> returns the value of the last expression evaluated.
2007
2008Inclusion of library modules is better done with the
2009L<C<use>|/use Module VERSION LIST> and L<C<require>|/require VERSION>
2010operators, which also do automatic error checking and raise an exception
2011if there's a problem.
2012
2013You might like to use L<C<do>|/do EXPR> to read in a program
2014configuration file.  Manual error checking can be done this way:
2015
2016    # Read in config files: system first, then user.
2017    # Beware of using relative pathnames here.
2018    for $file ("/share/prog/defaults.rc",
2019               "$ENV{HOME}/.someprogrc")
2020    {
2021        unless ($return = do $file) {
2022            warn "couldn't parse $file: $@" if $@;
2023            warn "couldn't do $file: $!"    unless defined $return;
2024            warn "couldn't run $file"       unless $return;
2025        }
2026    }
2027
2028=item dump LABEL
2029X<dump> X<core> X<undump>
2030
2031=item dump EXPR
2032
2033=item dump
2034
2035=for Pod::Functions create an immediate core dump
2036
2037This function causes an immediate core dump.  See also the B<-u>
2038command-line switch in L<perlrun|perlrun/-u>, which does the same thing.
2039Primarily this is so that you can use the B<undump> program (not
2040supplied) to turn your core dump into an executable binary after
2041having initialized all your variables at the beginning of the
2042program.  When the new binary is executed it will begin by executing
2043a C<goto LABEL> (with all the restrictions that L<C<goto>|/goto LABEL>
2044suffers).
2045Think of it as a goto with an intervening core dump and reincarnation.
2046If C<LABEL> is omitted, restarts the program from the top.  The
2047C<dump EXPR> form, available starting in Perl 5.18.0, allows a name to be
2048computed at run time, being otherwise identical to C<dump LABEL>.
2049
2050B<WARNING>: Any files opened at the time of the dump will I<not>
2051be open any more when the program is reincarnated, with possible
2052resulting confusion by Perl.
2053
2054This function is now largely obsolete, mostly because it's very hard to
2055convert a core file into an executable.  As of Perl 5.30, it must be invoked
2056as C<CORE::dump()>.
2057
2058Unlike most named operators, this has the same precedence as assignment.
2059It is also exempt from the looks-like-a-function rule, so
2060C<dump ("foo")."bar"> will cause "bar" to be part of the argument to
2061L<C<dump>|/dump LABEL>.
2062
2063Portability issues: L<perlport/dump>.
2064
2065=item each HASH
2066X<each> X<hash, iterator>
2067
2068=item each ARRAY
2069X<array, iterator>
2070
2071=for Pod::Functions retrieve the next key/value pair from a hash
2072
2073When called on a hash in list context, returns a 2-element list
2074consisting of the key and value for the next element of a hash.  In Perl
20755.12 and later only, it will also return the index and value for the next
2076element of an array so that you can iterate over it; older Perls consider
2077this a syntax error.  When called in scalar context, returns only the key
2078(not the value) in a hash, or the index in an array.
2079
2080Hash entries are returned in an apparently random order.  The actual random
2081order is specific to a given hash; the exact same series of operations
2082on two hashes may result in a different order for each hash.  Any insertion
2083into the hash may change the order, as will any deletion, with the exception
2084that the most recent key returned by L<C<each>|/each HASH> or
2085L<C<keys>|/keys HASH> may be deleted without changing the order.  So
2086long as a given hash is unmodified you may rely on
2087L<C<keys>|/keys HASH>, L<C<values>|/values HASH> and
2088L<C<each>|/each HASH> to repeatedly return the same order
2089as each other.  See L<perlsec/"Algorithmic Complexity Attacks"> for
2090details on why hash order is randomized.  Aside from the guarantees
2091provided here the exact details of Perl's hash algorithm and the hash
2092traversal order are subject to change in any release of Perl.
2093
2094After L<C<each>|/each HASH> has returned all entries from the hash or
2095array, the next call to L<C<each>|/each HASH> returns the empty list in
2096list context and L<C<undef>|/undef EXPR> in scalar context; the next
2097call following I<that> one restarts iteration.  Each hash or array has
2098its own internal iterator, accessed by L<C<each>|/each HASH>,
2099L<C<keys>|/keys HASH>, and L<C<values>|/values HASH>.  The iterator is
2100implicitly reset when L<C<each>|/each HASH> has reached the end as just
2101described; it can be explicitly reset by calling L<C<keys>|/keys HASH>
2102or L<C<values>|/values HASH> on the hash or array, or by referencing
2103the hash (but not array) in list context.  If you add or delete
2104a hash's elements while iterating over it, the effect on the iterator is
2105unspecified; for example, entries may be skipped or duplicated--so don't
2106do that.  Exception: It is always safe to delete the item most recently
2107returned by L<C<each>|/each HASH>, so the following code works properly:
2108
2109    while (my ($key, $value) = each %hash) {
2110        print $key, "\n";
2111        delete $hash{$key};   # This is safe
2112    }
2113
2114Tied hashes may have a different ordering behaviour to perl's hash
2115implementation.
2116
2117The iterator used by C<each> is attached to the hash or array, and is
2118shared between all iteration operations applied to the same hash or array.
2119Thus all uses of C<each> on a single hash or array advance the same
2120iterator location.  All uses of C<each> are also subject to having the
2121iterator reset by any use of C<keys> or C<values> on the same hash or
2122array, or by the hash (but not array) being referenced in list context.
2123This makes C<each>-based loops quite fragile: it is easy to arrive at
2124such a loop with the iterator already part way through the object, or to
2125accidentally clobber the iterator state during execution of the loop body.
2126It's easy enough to explicitly reset the iterator before starting a loop,
2127but there is no way to insulate the iterator state used by a loop from
2128the iterator state used by anything else that might execute during the
2129loop body.  To avoid these problems, use a C<foreach> loop rather than
2130C<while>-C<each>.
2131
2132This extends to using C<each> on the result of an anonymous hash or
2133array constructor.  A new underlying array or hash is created each
2134time so each will always start iterating from scratch, eg:
2135
2136  # loops forever
2137  while (my ($key, $value) = each @{ +{ a => 1 } }) {
2138      print "$key=$value\n";
2139  }
2140
2141This prints out your environment like the L<printenv(1)> program,
2142but in a different order:
2143
2144    while (my ($key,$value) = each %ENV) {
2145        print "$key=$value\n";
2146    }
2147
2148Starting with Perl 5.14, an experimental feature allowed
2149L<C<each>|/each HASH> to take a scalar expression. This experiment has
2150been deemed unsuccessful, and was removed as of Perl 5.24.
2151
2152As of Perl 5.18 you can use a bare L<C<each>|/each HASH> in a C<while>
2153loop, which will set L<C<$_>|perlvar/$_> on every iteration.
2154If either an C<each> expression or an explicit assignment of an C<each>
2155expression to a scalar is used as a C<while>/C<for> condition, then
2156the condition actually tests for definedness of the expression's value,
2157not for its regular truth value.
2158
2159    while (each %ENV) {
2160	print "$_=$ENV{$_}\n";
2161    }
2162
2163To avoid confusing would-be users of your code who are running earlier
2164versions of Perl with mysterious syntax errors, put this sort of thing at
2165the top of your file to signal that your code will work I<only> on Perls of
2166a recent vintage:
2167
2168    use v5.12;	# so keys/values/each work on arrays
2169    use v5.18;	# so each assigns to $_ in a lone while test
2170
2171See also L<C<keys>|/keys HASH>, L<C<values>|/values HASH>, and
2172L<C<sort>|/sort SUBNAME LIST>.
2173
2174=item eof FILEHANDLE
2175X<eof>
2176X<end of file>
2177X<end-of-file>
2178
2179=item eof ()
2180
2181=item eof
2182
2183=for Pod::Functions test a filehandle for its end
2184
2185Returns 1 if the next read on FILEHANDLE will return end of file I<or> if
2186FILEHANDLE is not open.  FILEHANDLE may be an expression whose value
2187gives the real filehandle.  (Note that this function actually
2188reads a character and then C<ungetc>s it, so isn't useful in an
2189interactive context.)  Do not read from a terminal file (or call
2190C<eof(FILEHANDLE)> on it) after end-of-file is reached.  File types such
2191as terminals may lose the end-of-file condition if you do.
2192
2193An L<C<eof>|/eof FILEHANDLE> without an argument uses the last file
2194read.  Using L<C<eof()>|/eof FILEHANDLE> with empty parentheses is
2195different.  It refers to the pseudo file formed from the files listed on
2196the command line and accessed via the C<< <> >> operator.  Since
2197C<< <> >> isn't explicitly opened, as a normal filehandle is, an
2198L<C<eof()>|/eof FILEHANDLE> before C<< <> >> has been used will cause
2199L<C<@ARGV>|perlvar/@ARGV> to be examined to determine if input is
2200available.   Similarly, an L<C<eof()>|/eof FILEHANDLE> after C<< <> >>
2201has returned end-of-file will assume you are processing another
2202L<C<@ARGV>|perlvar/@ARGV> list, and if you haven't set
2203L<C<@ARGV>|perlvar/@ARGV>, will read input from C<STDIN>; see
2204L<perlop/"I/O Operators">.
2205
2206In a C<< while (<>) >> loop, L<C<eof>|/eof FILEHANDLE> or C<eof(ARGV)>
2207can be used to detect the end of each file, whereas
2208L<C<eof()>|/eof FILEHANDLE> will detect the end of the very last file
2209only.  Examples:
2210
2211    # reset line numbering on each input file
2212    while (<>) {
2213        next if /^\s*#/;  # skip comments
2214        print "$.\t$_";
2215    } continue {
2216        close ARGV if eof;  # Not eof()!
2217    }
2218
2219    # insert dashes just before last line of last file
2220    while (<>) {
2221        if (eof()) {  # check for end of last file
2222            print "--------------\n";
2223        }
2224        print;
2225        last if eof();     # needed if we're reading from a terminal
2226    }
2227
2228Practical hint: you almost never need to use L<C<eof>|/eof FILEHANDLE>
2229in Perl, because the input operators typically return L<C<undef>|/undef
2230EXPR> when they run out of data or encounter an error.
2231
2232=item eval EXPR
2233X<eval> X<try> X<catch> X<evaluate> X<parse> X<execute>
2234X<error, handling> X<exception, handling>
2235
2236=item eval BLOCK
2237
2238=item eval
2239
2240=for Pod::Functions catch exceptions or compile and run code
2241
2242C<eval> in all its forms is used to execute a little Perl program,
2243trapping any errors encountered so they don't crash the calling program.
2244
2245Plain C<eval> with no argument is just C<eval EXPR>, where the
2246expression is understood to be contained in L<C<$_>|perlvar/$_>.  Thus
2247there are only two real C<eval> forms; the one with an EXPR is often
2248called "string eval".  In a string eval, the value of the expression
2249(which is itself determined within scalar context) is first parsed, and
2250if there were no errors, executed as a block within the lexical context
2251of the current Perl program.  This form is typically used to delay
2252parsing and subsequent execution of the text of EXPR until run time.
2253Note that the value is parsed every time the C<eval> executes.
2254
2255The other form is called "block eval".  It is less general than string
2256eval, but the code within the BLOCK is parsed only once (at the same
2257time the code surrounding the C<eval> itself was parsed) and executed
2258within the context of the current Perl program.  This form is typically
2259used to trap exceptions more efficiently than the first, while also
2260providing the benefit of checking the code within BLOCK at compile time.
2261BLOCK is parsed and compiled just once.  Since errors are trapped, it
2262often is used to check if a given feature is available.
2263
2264In both forms, the value returned is the value of the last expression
2265evaluated inside the mini-program; a return statement may also be used, just
2266as with subroutines.  The expression providing the return value is evaluated
2267in void, scalar, or list context, depending on the context of the
2268C<eval> itself.  See L<C<wantarray>|/wantarray> for more
2269on how the evaluation context can be determined.
2270
2271If there is a syntax error or runtime error, or a L<C<die>|/die LIST>
2272statement is executed, C<eval> returns
2273L<C<undef>|/undef EXPR> in scalar context, or an empty list in list
2274context, and L<C<$@>|perlvar/$@> is set to the error message.  (Prior to
22755.16, a bug caused L<C<undef>|/undef EXPR> to be returned in list
2276context for syntax errors, but not for runtime errors.) If there was no
2277error, L<C<$@>|perlvar/$@> is set to the empty string.  A control flow
2278operator like L<C<last>|/last LABEL> or L<C<goto>|/goto LABEL> can
2279bypass the setting of L<C<$@>|perlvar/$@>.  Beware that using
2280C<eval> neither silences Perl from printing warnings to
2281STDERR, nor does it stuff the text of warning messages into
2282L<C<$@>|perlvar/$@>.  To do either of those, you have to use the
2283L<C<$SIG{__WARN__}>|perlvar/%SIG> facility, or turn off warnings inside
2284the BLOCK or EXPR using S<C<no warnings 'all'>>.  See
2285L<C<warn>|/warn LIST>, L<perlvar>, and L<warnings>.
2286
2287Note that, because C<eval> traps otherwise-fatal errors,
2288it is useful for determining whether a particular feature (such as
2289L<C<socket>|/socket SOCKET,DOMAIN,TYPE,PROTOCOL> or
2290L<C<symlink>|/symlink OLDFILE,NEWFILE>) is implemented.  It is also
2291Perl's exception-trapping mechanism, where the L<C<die>|/die LIST>
2292operator is used to raise exceptions.
2293
2294Before Perl 5.14, the assignment to L<C<$@>|perlvar/$@> occurred before
2295restoration
2296of localized variables, which means that for your code to run on older
2297versions, a temporary is required if you want to mask some, but not all
2298errors:
2299
2300 # alter $@ on nefarious repugnancy only
2301 {
2302    my $e;
2303    {
2304      local $@; # protect existing $@
2305      eval { test_repugnancy() };
2306      # $@ =~ /nefarious/ and die $@; # Perl 5.14 and higher only
2307      $@ =~ /nefarious/ and $e = $@;
2308    }
2309    die $e if defined $e
2310 }
2311
2312There are some different considerations for each form:
2313
2314=over 4
2315
2316=item String eval
2317
2318Since the return value of EXPR is executed as a block within the lexical
2319context of the current Perl program, any outer lexical variables are
2320visible to it, and any package variable settings or subroutine and
2321format definitions remain afterwards.
2322
2323Note that when C<BEGIN {}> blocks are embedded inside of an eval block
2324the contents of the block will be executed immediately and before the rest
2325of the eval code is executed. You can disable this entirely by
2326
2327   local ${^MAX_NESTED_EVAL_BEGIN_BLOCKS} = 0;
2328   eval $string;
2329
2330which will cause any embedded C<BEGIN> blocks in C<$string> to throw an
2331exception.
2332
2333=over 4
2334
2335=item Under the L<C<"unicode_eval"> feature|feature/The 'unicode_eval' and 'evalbytes' features>
2336
2337If this feature is enabled (which is the default under a C<use 5.16> or
2338higher declaration), Perl assumes that EXPR is a character string.
2339Any S<C<use utf8>> or S<C<no utf8>> declarations within
2340the string thus have no effect. Source filters are forbidden as well.
2341(C<unicode_strings>, however, can appear within the string.)
2342
2343See also the L<C<evalbytes>|/evalbytes EXPR> operator, which works properly
2344with source filters.
2345
2346=item Outside the C<"unicode_eval"> feature
2347
2348In this case, the behavior is problematic and is not so easily
2349described.  Here are two bugs that cannot easily be fixed without
2350breaking existing programs:
2351
2352=over 4
2353
2354=item *
2355
2356Perl's internal storage of EXPR affects the behavior of the executed code.
2357For example:
2358
2359    my $v = eval "use utf8; '$expr'";
2360
2361If $expr is C<"\xc4\x80"> (U+0100 in UTF-8), then the value stored in C<$v>
2362will depend on whether Perl stores $expr "upgraded" (cf. L<utf8>) or
2363not:
2364
2365=over
2366
2367=item * If upgraded, C<$v> will be C<"\xc4\x80"> (i.e., the
2368C<use utf8> has no effect.)
2369
2370=item * If non-upgraded, C<$v> will be C<"\x{100}">.
2371
2372=back
2373
2374This is undesirable since being
2375upgraded or not should not affect a string's behavior.
2376
2377=item *
2378
2379Source filters activated within C<eval> leak out into whichever file
2380scope is currently being compiled.  To give an example with the CPAN module
2381L<Semi::Semicolons>:
2382
2383 BEGIN { eval "use Semi::Semicolons; # not filtered" }
2384 # filtered here!
2385
2386L<C<evalbytes>|/evalbytes EXPR> fixes that to work the way one would
2387expect:
2388
2389 use feature "evalbytes";
2390 BEGIN { evalbytes "use Semi::Semicolons; # filtered" }
2391 # not filtered
2392
2393=back
2394
2395=back
2396
2397Problems can arise if the string expands a scalar containing a floating
2398point number.  That scalar can expand to letters, such as C<"NaN"> or
2399C<"Infinity">; or, within the scope of a L<C<use locale>|locale>, the
2400decimal point character may be something other than a dot (such as a
2401comma).  None of these are likely to parse as you are likely expecting.
2402
2403You should be especially careful to remember what's being looked at
2404when:
2405
2406    eval $x;        # CASE 1
2407    eval "$x";      # CASE 2
2408
2409    eval '$x';      # CASE 3
2410    eval { $x };    # CASE 4
2411
2412    eval "\$$x++";  # CASE 5
2413    $$x++;          # CASE 6
2414
2415Cases 1 and 2 above behave identically: they run the code contained in
2416the variable $x.  (Although case 2 has misleading double quotes making
2417the reader wonder what else might be happening (nothing is).)  Cases 3
2418and 4 likewise behave in the same way: they run the code C<'$x'>, which
2419does nothing but return the value of $x.  (Case 4 is preferred for
2420purely visual reasons, but it also has the advantage of compiling at
2421compile-time instead of at run-time.)  Case 5 is a place where
2422normally you I<would> like to use double quotes, except that in this
2423particular situation, you can just use symbolic references instead, as
2424in case 6.
2425
2426An C<eval ''> executed within a subroutine defined
2427in the C<DB> package doesn't see the usual
2428surrounding lexical scope, but rather the scope of the first non-DB piece
2429of code that called it.  You don't normally need to worry about this unless
2430you are writing a Perl debugger.
2431
2432The final semicolon, if any, may be omitted from the value of EXPR.
2433
2434=item Block eval
2435
2436If the code to be executed doesn't vary, you may use the eval-BLOCK
2437form to trap run-time errors without incurring the penalty of
2438recompiling each time.  The error, if any, is still returned in
2439L<C<$@>|perlvar/$@>.
2440Examples:
2441
2442    # make divide-by-zero nonfatal
2443    eval { $answer = $x / $y; }; warn $@ if $@;
2444
2445    # same thing, but less efficient
2446    eval '$answer = $x / $y'; warn $@ if $@;
2447
2448    # a compile-time error
2449    eval { $answer = }; # WRONG
2450
2451    # a run-time error
2452    eval '$answer =';   # sets $@
2453
2454If you want to trap errors when loading an XS module, some problems with
2455the binary interface (such as Perl version skew) may be fatal even with
2456C<eval> unless C<$ENV{PERL_DL_NONLAZY}> is set.  See
2457L<perlrun|perlrun/PERL_DL_NONLAZY>.
2458
2459Using the C<eval {}> form as an exception trap in libraries does have some
2460issues.  Due to the current arguably broken state of C<__DIE__> hooks, you
2461may wish not to trigger any C<__DIE__> hooks that user code may have installed.
2462You can use the C<local $SIG{__DIE__}> construct for this purpose,
2463as this example shows:
2464
2465    # a private exception trap for divide-by-zero
2466    eval { local $SIG{'__DIE__'}; $answer = $x / $y; };
2467    warn $@ if $@;
2468
2469This is especially significant, given that C<__DIE__> hooks can call
2470L<C<die>|/die LIST> again, which has the effect of changing their error
2471messages:
2472
2473    # __DIE__ hooks may modify error messages
2474    {
2475       local $SIG{'__DIE__'} =
2476              sub { (my $x = $_[0]) =~ s/foo/bar/g; die $x };
2477       eval { die "foo lives here" };
2478       print $@ if $@;                # prints "bar lives here"
2479    }
2480
2481Because this promotes action at a distance, this counterintuitive behavior
2482may be fixed in a future release.
2483
2484C<eval BLOCK> does I<not> count as a loop, so the loop control statements
2485L<C<next>|/next LABEL>, L<C<last>|/last LABEL>, or
2486L<C<redo>|/redo LABEL> cannot be used to leave or restart the block.
2487
2488The final semicolon, if any, may be omitted from within the BLOCK.
2489
2490=back
2491
2492=item evalbytes EXPR
2493X<evalbytes>
2494
2495=item evalbytes
2496
2497=for Pod::Functions +evalbytes similar to string eval, but intend to parse a bytestream
2498
2499This function is similar to a L<string eval|/eval EXPR>, except it
2500always parses its argument (or L<C<$_>|perlvar/$_> if EXPR is omitted)
2501as a byte string. If the string contains any code points above 255, then
2502it cannot be a byte string, and the C<evalbytes> will fail with the error
2503stored in C<$@>.
2504
2505C<use utf8> and C<no utf8> within the string have their usual effect.
2506
2507Source filters activated within the evaluated code apply to the code
2508itself.
2509
2510L<C<evalbytes>|/evalbytes EXPR> is available starting in Perl v5.16.  To
2511access it, you must say C<CORE::evalbytes>, but you can omit the
2512C<CORE::> if the
2513L<C<"evalbytes"> feature|feature/The 'unicode_eval' and 'evalbytes' features>
2514is enabled.  This is enabled automatically with a C<use v5.16> (or
2515higher) declaration in the current scope.
2516
2517=item exec LIST
2518X<exec> X<execute>
2519
2520=item exec PROGRAM LIST
2521
2522=for Pod::Functions abandon this program to run another
2523
2524The L<C<exec>|/exec LIST> function executes a system command I<and never
2525returns>; use L<C<system>|/system LIST> instead of L<C<exec>|/exec LIST>
2526if you want it to return.  It fails and
2527returns false only if the command does not exist I<and> it is executed
2528directly instead of via your system's command shell (see below).
2529
2530Since it's a common mistake to use L<C<exec>|/exec LIST> instead of
2531L<C<system>|/system LIST>, Perl warns you if L<C<exec>|/exec LIST> is
2532called in void context and if there is a following statement that isn't
2533L<C<die>|/die LIST>, L<C<warn>|/warn LIST>, or L<C<exit>|/exit EXPR> (if
2534L<warnings> are enabled--but you always do that, right?).  If you
2535I<really> want to follow an L<C<exec>|/exec LIST> with some other
2536statement, you can use one of these styles to avoid the warning:
2537
2538    exec ('foo')   or print STDERR "couldn't exec foo: $!";
2539    { exec ('foo') }; print STDERR "couldn't exec foo: $!";
2540
2541If there is more than one argument in LIST, this calls L<execvp(3)> with the
2542arguments in LIST.  If there is only one element in LIST, the argument is
2543checked for shell metacharacters, and if there are any, the entire
2544argument is passed to the system's command shell for parsing (this is
2545C</bin/sh -c> on Unix platforms, but varies on other platforms).  If
2546there are no shell metacharacters in the argument, it is split into words
2547and passed directly to C<execvp>, which is more efficient.  Examples:
2548
2549    exec '/bin/echo', 'Your arguments are: ', @ARGV;
2550    exec "sort $outfile | uniq";
2551
2552If you don't really want to execute the first argument, but want to lie
2553to the program you are executing about its own name, you can specify
2554the program you actually want to run as an "indirect object" (without a
2555comma) in front of the LIST, as in C<exec PROGRAM LIST>.  (This always
2556forces interpretation of the LIST as a multivalued list, even if there
2557is only a single scalar in the list.)  Example:
2558
2559    my $shell = '/bin/csh';
2560    exec $shell '-sh';    # pretend it's a login shell
2561
2562or, more directly,
2563
2564    exec {'/bin/csh'} '-sh';  # pretend it's a login shell
2565
2566When the arguments get executed via the system shell, results are
2567subject to its quirks and capabilities.  See L<perlop/"`STRING`">
2568for details.
2569
2570Using an indirect object with L<C<exec>|/exec LIST> or
2571L<C<system>|/system LIST> is also more secure.  This usage (which also
2572works fine with L<C<system>|/system LIST>) forces
2573interpretation of the arguments as a multivalued list, even if the
2574list had just one argument.  That way you're safe from the shell
2575expanding wildcards or splitting up words with whitespace in them.
2576
2577    my @args = ( "echo surprise" );
2578
2579    exec @args;               # subject to shell escapes
2580                                # if @args == 1
2581    exec { $args[0] } @args;  # safe even with one-arg list
2582
2583The first version, the one without the indirect object, ran the I<echo>
2584program, passing it C<"surprise"> an argument.  The second version didn't;
2585it tried to run a program named I<"echo surprise">, didn't find it, and set
2586L<C<$?>|perlvar/$?> to a non-zero value indicating failure.
2587
2588On Windows, only the C<exec PROGRAM LIST> indirect object syntax will
2589reliably avoid using the shell; C<exec LIST>, even with more than one
2590element, will fall back to the shell if the first spawn fails.
2591
2592Perl attempts to flush all files opened for output before the exec,
2593but this may not be supported on some platforms (see L<perlport>).
2594To be safe, you may need to set L<C<$E<verbar>>|perlvar/$E<verbar>>
2595(C<$AUTOFLUSH> in L<English>) or call the C<autoflush> method of
2596L<C<IO::Handle>|IO::Handle/METHODS> on any open handles to avoid lost
2597output.
2598
2599Note that L<C<exec>|/exec LIST> will not call your C<END> blocks, nor
2600will it invoke C<DESTROY> methods on your objects.
2601
2602Portability issues: L<perlport/exec>.
2603
2604=item exists EXPR
2605X<exists> X<autovivification>
2606
2607=for Pod::Functions test whether a hash key is present
2608
2609Given an expression that specifies an element of a hash, returns true if the
2610specified element in the hash has ever been initialized, even if the
2611corresponding value is undefined.
2612
2613    print "Exists\n"    if exists $hash{$key};
2614    print "Defined\n"   if defined $hash{$key};
2615    print "True\n"      if $hash{$key};
2616
2617exists may also be called on array elements, but its behavior is much less
2618obvious and is strongly tied to the use of L<C<delete>|/delete EXPR> on
2619arrays.
2620
2621B<WARNING:> Calling L<C<exists>|/exists EXPR> on array values is
2622strongly discouraged.  The
2623notion of deleting or checking the existence of Perl array elements is not
2624conceptually coherent, and can lead to surprising behavior.
2625
2626    print "Exists\n"    if exists $array[$index];
2627    print "Defined\n"   if defined $array[$index];
2628    print "True\n"      if $array[$index];
2629
2630A hash or array element can be true only if it's defined and defined only if
2631it exists, but the reverse doesn't necessarily hold true.
2632
2633Given an expression that specifies the name of a subroutine,
2634returns true if the specified subroutine has ever been declared, even
2635if it is undefined.  Mentioning a subroutine name for exists or defined
2636does not count as declaring it.  Note that a subroutine that does not
2637exist may still be callable: its package may have an C<AUTOLOAD>
2638method that makes it spring into existence the first time that it is
2639called; see L<perlsub>.
2640
2641    print "Exists\n"  if exists &subroutine;
2642    print "Defined\n" if defined &subroutine;
2643
2644Note that the EXPR can be arbitrarily complicated as long as the final
2645operation is a hash or array key lookup or subroutine name:
2646
2647    if (exists $ref->{A}->{B}->{$key})  { }
2648    if (exists $hash{A}{B}{$key})       { }
2649
2650    if (exists $ref->{A}->{B}->[$ix])   { }
2651    if (exists $hash{A}{B}[$ix])        { }
2652
2653    if (exists &{$ref->{A}{B}{$key}})   { }
2654
2655Although the most deeply nested array or hash element will not spring into
2656existence just because its existence was tested, any intervening ones will.
2657Thus C<< $ref->{"A"} >> and C<< $ref->{"A"}->{"B"} >> will spring
2658into existence due to the existence test for the C<$key> element above.
2659This happens anywhere the arrow operator is used, including even here:
2660
2661    undef $ref;
2662    if (exists $ref->{"Some key"})    { }
2663    print $ref;  # prints HASH(0x80d3d5c)
2664
2665Use of a subroutine call, rather than a subroutine name, as an argument
2666to L<C<exists>|/exists EXPR> is an error.
2667
2668    exists &sub;    # OK
2669    exists &sub();  # Error
2670
2671=item exit EXPR
2672X<exit> X<terminate> X<abort>
2673
2674=item exit
2675
2676=for Pod::Functions terminate this program
2677
2678Evaluates EXPR and exits immediately with that value.    Example:
2679
2680    my $ans = <STDIN>;
2681    exit 0 if $ans =~ /^[Xx]/;
2682
2683See also L<C<die>|/die LIST>.  If EXPR is omitted, exits with C<0>
2684status.  The only
2685universally recognized values for EXPR are C<0> for success and C<1>
2686for error; other values are subject to interpretation depending on the
2687environment in which the Perl program is running.  For example, exiting
268869 (EX_UNAVAILABLE) from a I<sendmail> incoming-mail filter will cause
2689the mailer to return the item undelivered, but that's not true everywhere.
2690
2691Don't use L<C<exit>|/exit EXPR> to abort a subroutine if there's any
2692chance that someone might want to trap whatever error happened.  Use
2693L<C<die>|/die LIST> instead, which can be trapped by an
2694L<C<eval>|/eval EXPR>.
2695
2696The L<C<exit>|/exit EXPR> function does not always exit immediately.  It
2697calls any defined C<END> routines first, but these C<END> routines may
2698not themselves abort the exit.  Likewise any object destructors that
2699need to be called are called before the real exit.  C<END> routines and
2700destructors can change the exit status by modifying L<C<$?>|perlvar/$?>.
2701If this is a problem, you can call
2702L<C<POSIX::_exit($status)>|POSIX/C<_exit>> to avoid C<END> and destructor
2703processing.  See L<perlmod> for details.
2704
2705Portability issues: L<perlport/exit>.
2706
2707=item exp EXPR
2708X<exp> X<exponential> X<antilog> X<antilogarithm> X<e>
2709
2710=item exp
2711
2712=for Pod::Functions raise I<e> to a power
2713
2714Returns I<e> (the natural logarithm base) to the power of EXPR.
2715If EXPR is omitted, gives C<exp($_)>.
2716
2717=item fc EXPR
2718X<fc> X<foldcase> X<casefold> X<fold-case> X<case-fold>
2719
2720=item fc
2721
2722=for Pod::Functions +fc return casefolded version of a string
2723
2724Returns the casefolded version of EXPR.  This is the internal function
2725implementing the C<\F> escape in double-quoted strings.
2726
2727Casefolding is the process of mapping strings to a form where case
2728differences are erased; comparing two strings in their casefolded
2729form is effectively a way of asking if two strings are equal,
2730regardless of case.
2731
2732Roughly, if you ever found yourself writing this
2733
2734    lc($this) eq lc($that)    # Wrong!
2735        # or
2736    uc($this) eq uc($that)    # Also wrong!
2737        # or
2738    $this =~ /^\Q$that\E\z/i  # Right!
2739
2740Now you can write
2741
2742    fc($this) eq fc($that)
2743
2744And get the correct results.
2745
2746Perl only implements the full form of casefolding, but you can access
2747the simple folds using L<Unicode::UCD/B<casefold()>> and
2748L<Unicode::UCD/B<prop_invmap()>>.
2749For further information on casefolding, refer to
2750the Unicode Standard, specifically sections 3.13 C<Default Case Operations>,
27514.2 C<Case-Normative>, and 5.18 C<Case Mappings>,
2752available at L<https://www.unicode.org/versions/latest/>, as well as the
2753Case Charts available at L<https://www.unicode.org/charts/case/>.
2754
2755If EXPR is omitted, uses L<C<$_>|perlvar/$_>.
2756
2757This function behaves the same way under various pragmas, such as within
2758L<S<C<"use feature 'unicode_strings">>|feature/The 'unicode_strings' feature>,
2759as L<C<lc>|/lc EXPR> does, with the single exception of
2760L<C<fc>|/fc EXPR> of I<LATIN CAPITAL LETTER SHARP S> (U+1E9E) within the
2761scope of L<S<C<use locale>>|locale>.  The foldcase of this character
2762would normally be C<"ss">, but as explained in the L<C<lc>|/lc EXPR>
2763section, case
2764changes that cross the 255/256 boundary are problematic under locales,
2765and are hence prohibited.  Therefore, this function under locale returns
2766instead the string C<"\x{17F}\x{17F}">, which is the I<LATIN SMALL LETTER
2767LONG S>.  Since that character itself folds to C<"s">, the string of two
2768of them together should be equivalent to a single U+1E9E when foldcased.
2769
2770While the Unicode Standard defines two additional forms of casefolding,
2771one for Turkic languages and one that never maps one character into multiple
2772characters, these are not provided by the Perl core.  However, the CPAN module
2773L<C<Unicode::Casing>|Unicode::Casing> may be used to provide an implementation.
2774
2775L<C<fc>|/fc EXPR> is available only if the
2776L<C<"fc"> feature|feature/The 'fc' feature> is enabled or if it is
2777prefixed with C<CORE::>.  The
2778L<C<"fc"> feature|feature/The 'fc' feature> is enabled automatically
2779with a C<use v5.16> (or higher) declaration in the current scope.
2780
2781=item fcntl FILEHANDLE,FUNCTION,SCALAR
2782X<fcntl>
2783
2784=for Pod::Functions file control system call
2785
2786Implements the L<fcntl(2)> function.  You'll probably have to say
2787
2788    use Fcntl;
2789
2790first to get the correct constant definitions.  Argument processing and
2791value returned work just like L<C<ioctl>|/ioctl
2792FILEHANDLE,FUNCTION,SCALAR> below.  For example:
2793
2794    use Fcntl;
2795    my $flags = fcntl($filehandle, F_GETFL, 0)
2796        or die "Can't fcntl F_GETFL: $!";
2797
2798You don't have to check for L<C<defined>|/defined EXPR> on the return
2799from L<C<fcntl>|/fcntl FILEHANDLE,FUNCTION,SCALAR>.  Like
2800L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR>, it maps a C<0> return
2801from the system call into C<"0 but true"> in Perl.  This string is true
2802in boolean context and C<0> in numeric context.  It is also exempt from
2803the normal
2804L<C<Argument "..." isn't numeric>|perldiag/Argument "%s" isn't numeric%s>
2805L<warnings> on improper numeric conversions.
2806
2807Note that L<C<fcntl>|/fcntl FILEHANDLE,FUNCTION,SCALAR> raises an
2808exception if used on a machine that doesn't implement L<fcntl(2)>.  See
2809the L<Fcntl> module or your L<fcntl(2)> manpage to learn what functions
2810are available on your system.
2811
2812Here's an example of setting a filehandle named C<$REMOTE> to be
2813non-blocking at the system level.  You'll have to negotiate
2814L<C<$E<verbar>>|perlvar/$E<verbar>> on your own, though.
2815
2816    use Fcntl qw(F_GETFL F_SETFL O_NONBLOCK);
2817
2818    my $flags = fcntl($REMOTE, F_GETFL, 0)
2819        or die "Can't get flags for the socket: $!\n";
2820
2821    fcntl($REMOTE, F_SETFL, $flags | O_NONBLOCK)
2822        or die "Can't set flags for the socket: $!\n";
2823
2824Portability issues: L<perlport/fcntl>.
2825
2826=item __FILE__
2827X<__FILE__>
2828
2829=for Pod::Functions the name of the current source file
2830
2831A special token that returns the name of the file in which it occurs.
2832It can be altered by the mechanism described at
2833L<perlsyn/"Plain Old Comments (Not!)">.
2834
2835=item field VARNAME
2836X<field>
2837
2838=for Pod::Functions declare a field variable of the current class
2839
2840Declares a new field variable within the current class.  Methods and
2841C<ADJUST> blocks of the class will have access to this variable as if it
2842was a lexical in scope at that point.
2843
2844=item fileno FILEHANDLE
2845X<fileno>
2846
2847=item fileno DIRHANDLE
2848
2849=for Pod::Functions return file descriptor from filehandle
2850
2851Returns the file descriptor for a filehandle or directory handle,
2852or undefined if the
2853filehandle is not open.  If there is no real file descriptor at the OS
2854level, as can happen with filehandles connected to memory objects via
2855L<C<open>|/open FILEHANDLE,MODE,EXPR> with a reference for the third
2856argument, -1 is returned.
2857
2858This is mainly useful for constructing bitmaps for
2859L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT> and low-level POSIX
2860tty-handling operations.
2861If FILEHANDLE is an expression, the value is taken as an indirect
2862filehandle, generally its name.
2863
2864You can use this to find out whether two handles refer to the
2865same underlying descriptor:
2866
2867    if (fileno($this) != -1 && fileno($this) == fileno($that)) {
2868        print "\$this and \$that are dups\n";
2869    } elsif (fileno($this) != -1 && fileno($that) != -1) {
2870        print "\$this and \$that have different " .
2871            "underlying file descriptors\n";
2872    } else {
2873        print "At least one of \$this and \$that does " .
2874            "not have a real file descriptor\n";
2875    }
2876
2877The behavior of L<C<fileno>|/fileno FILEHANDLE> on a directory handle
2878depends on the operating system.  On a system with L<dirfd(3)> or
2879similar, L<C<fileno>|/fileno FILEHANDLE> on a directory
2880handle returns the underlying file descriptor associated with the
2881handle; on systems with no such support, it returns the undefined value,
2882and sets L<C<$!>|perlvar/$!> (errno).
2883
2884=item flock FILEHANDLE,OPERATION
2885X<flock> X<lock> X<locking>
2886
2887=for Pod::Functions lock an entire file with an advisory lock
2888
2889Calls L<flock(2)>, or an emulation of it, on FILEHANDLE.  Returns true
2890for success, false on failure.  Produces a fatal error if used on a
2891machine that doesn't implement L<flock(2)>, L<fcntl(2)> locking, or
2892L<lockf(3)>.  L<C<flock>|/flock FILEHANDLE,OPERATION> is Perl's portable
2893file-locking interface, although it locks entire files only, not
2894records.
2895
2896Two potentially non-obvious but traditional L<C<flock>|/flock
2897FILEHANDLE,OPERATION> semantics are
2898that it waits indefinitely until the lock is granted, and that its locks
2899are B<merely advisory>.  Such discretionary locks are more flexible, but
2900offer fewer guarantees.  This means that programs that do not also use
2901L<C<flock>|/flock FILEHANDLE,OPERATION> may modify files locked with
2902L<C<flock>|/flock FILEHANDLE,OPERATION>.  See L<perlport>,
2903your port's specific documentation, and your system-specific local manpages
2904for details.  It's best to assume traditional behavior if you're writing
2905portable programs.  (But if you're not, you should as always feel perfectly
2906free to write for your own system's idiosyncrasies (sometimes called
2907"features").  Slavish adherence to portability concerns shouldn't get
2908in the way of your getting your job done.)
2909
2910OPERATION is one of LOCK_SH, LOCK_EX, or LOCK_UN, possibly combined with
2911LOCK_NB.  These constants are traditionally valued 1, 2, 8 and 4, but
2912you can use the symbolic names if you import them from the L<Fcntl> module,
2913either individually, or as a group using the C<:flock> tag.  LOCK_SH
2914requests a shared lock, LOCK_EX requests an exclusive lock, and LOCK_UN
2915releases a previously requested lock.  If LOCK_NB is bitwise-or'ed with
2916LOCK_SH or LOCK_EX, then L<C<flock>|/flock FILEHANDLE,OPERATION> returns
2917immediately rather than blocking waiting for the lock; check the return
2918status to see if you got it.
2919
2920To avoid the possibility of miscoordination, Perl now flushes FILEHANDLE
2921before locking or unlocking it.
2922
2923Note that the emulation built with L<lockf(3)> doesn't provide shared
2924locks, and it requires that FILEHANDLE be open with write intent.  These
2925are the semantics that L<lockf(3)> implements.  Most if not all systems
2926implement L<lockf(3)> in terms of L<fcntl(2)> locking, though, so the
2927differing semantics shouldn't bite too many people.
2928
2929Note that the L<fcntl(2)> emulation of L<flock(3)> requires that FILEHANDLE
2930be open with read intent to use LOCK_SH and requires that it be open
2931with write intent to use LOCK_EX.
2932
2933Note also that some versions of L<C<flock>|/flock FILEHANDLE,OPERATION>
2934cannot lock things over the network; you would need to use the more
2935system-specific L<C<fcntl>|/fcntl FILEHANDLE,FUNCTION,SCALAR> for
2936that.  If you like you can force Perl to ignore your system's L<flock(2)>
2937function, and so provide its own L<fcntl(2)>-based emulation, by passing
2938the switch C<-Ud_flock> to the F<Configure> program when you configure
2939and build a new Perl.
2940
2941Here's a mailbox appender for BSD systems.
2942
2943    # import LOCK_* and SEEK_END constants
2944    use Fcntl qw(:flock SEEK_END);
2945
2946    sub lock {
2947        my ($fh) = @_;
2948        flock($fh, LOCK_EX) or die "Cannot lock mailbox - $!\n";
2949        # and, in case we're running on a very old UNIX
2950        # variant without the modern O_APPEND semantics...
2951        seek($fh, 0, SEEK_END) or die "Cannot seek - $!\n";
2952    }
2953
2954    sub unlock {
2955        my ($fh) = @_;
2956        flock($fh, LOCK_UN) or die "Cannot unlock mailbox - $!\n";
2957    }
2958
2959    open(my $mbox, ">>", "/usr/spool/mail/$ENV{'USER'}")
2960        or die "Can't open mailbox: $!";
2961
2962    lock($mbox);
2963    print $mbox $msg,"\n\n";
2964    unlock($mbox);
2965
2966On systems that support a real L<flock(2)>, locks are inherited across
2967L<C<fork>|/fork> calls, whereas those that must resort to the more
2968capricious L<fcntl(2)> function lose their locks, making it seriously
2969harder to write servers.
2970
2971See also L<DB_File> for other L<C<flock>|/flock FILEHANDLE,OPERATION>
2972examples.
2973
2974Portability issues: L<perlport/flock>.
2975
2976=item fork
2977X<fork> X<child> X<parent>
2978
2979=for Pod::Functions create a new process just like this one
2980
2981Does a L<fork(2)> system call to create a new process running the
2982same program at the same point.  It returns the child pid to the
2983parent process, C<0> to the child process, or L<C<undef>|/undef EXPR> if
2984the fork is
2985unsuccessful.  File descriptors (and sometimes locks on those descriptors)
2986are shared, while everything else is copied.  On most systems supporting
2987L<fork(2)>, great care has gone into making it extremely efficient (for
2988example, using copy-on-write technology on data pages), making it the
2989dominant paradigm for multitasking over the last few decades.
2990
2991Perl attempts to flush all files opened for output before forking the
2992child process, but this may not be supported on some platforms (see
2993L<perlport>).  To be safe, you may need to set
2994L<C<$E<verbar>>|perlvar/$E<verbar>> (C<$AUTOFLUSH> in L<English>) or
2995call the C<autoflush> method of L<C<IO::Handle>|IO::Handle/METHODS> on
2996any open handles to avoid duplicate output.
2997
2998If you L<C<fork>|/fork> without ever waiting on your children, you will
2999accumulate zombies.  On some systems, you can avoid this by setting
3000L<C<$SIG{CHLD}>|perlvar/%SIG> to C<"IGNORE">.  See also L<perlipc> for
3001more examples of forking and reaping moribund children.
3002
3003Note that if your forked child inherits system file descriptors like
3004STDIN and STDOUT that are actually connected by a pipe or socket, even
3005if you exit, then the remote server (such as, say, a CGI script or a
3006backgrounded job launched from a remote shell) won't think you're done.
3007You should reopen those to F</dev/null> if it's any issue.
3008
3009On some platforms such as Windows, where the L<fork(2)> system call is
3010not available, Perl can be built to emulate L<C<fork>|/fork> in the Perl
3011interpreter.  The emulation is designed, at the level of the Perl
3012program, to be as compatible as possible with the "Unix" L<fork(2)>.
3013However it has limitations that have to be considered in code intended
3014to be portable.  See L<perlfork> for more details.
3015
3016Portability issues: L<perlport/fork>.
3017
3018=item format
3019X<format>
3020
3021=for Pod::Functions declare a picture format with use by the write() function
3022
3023Declare a picture format for use by the L<C<write>|/write FILEHANDLE>
3024function.  For example:
3025
3026    format Something =
3027        Test: @<<<<<<<< @||||| @>>>>>
3028              $str,     $%,    '$' . int($num)
3029    .
3030
3031    $str = "widget";
3032    $num = $cost/$quantity;
3033    $~ = 'Something';
3034    write;
3035
3036See L<perlform> for many details and examples.
3037
3038=item formline PICTURE,LIST
3039X<formline>
3040
3041=for Pod::Functions internal function used for formats
3042
3043This is an internal function used by L<C<format>|/format>s, though you
3044may call it, too.  It formats (see L<perlform>) a list of values
3045according to the contents of PICTURE, placing the output into the format
3046output accumulator, L<C<$^A>|perlvar/$^A> (or C<$ACCUMULATOR> in
3047L<English>).  Eventually, when a L<C<write>|/write FILEHANDLE> is done,
3048the contents of L<C<$^A>|perlvar/$^A> are written to some filehandle.
3049You could also read L<C<$^A>|perlvar/$^A> and then set
3050L<C<$^A>|perlvar/$^A> back to C<"">.  Note that a format typically does
3051one L<C<formline>|/formline PICTURE,LIST> per line of form, but the
3052L<C<formline>|/formline PICTURE,LIST> function itself doesn't care how
3053many newlines are embedded in the PICTURE.  This means that the C<~> and
3054C<~~> tokens treat the entire PICTURE as a single line.  You may
3055therefore need to use multiple formlines to implement a single record
3056format, just like the L<C<format>|/format> compiler.
3057
3058Be careful if you put double quotes around the picture, because an C<@>
3059character may be taken to mean the beginning of an array name.
3060L<C<formline>|/formline PICTURE,LIST> always returns true.  See
3061L<perlform> for other examples.
3062
3063If you are trying to use this instead of L<C<write>|/write FILEHANDLE>
3064to capture the output, you may find it easier to open a filehandle to a
3065scalar (C<< open my $fh, ">", \$output >>) and write to that instead.
3066
3067=item getc FILEHANDLE
3068X<getc> X<getchar> X<character> X<file, read>
3069
3070=item getc
3071
3072=for Pod::Functions get the next character from the filehandle
3073
3074Returns the next character from the input file attached to FILEHANDLE,
3075or the undefined value at end of file or if there was an error (in
3076the latter case L<C<$!>|perlvar/$!> is set).  If FILEHANDLE is omitted,
3077reads from
3078STDIN.  This is not particularly efficient.  However, it cannot be
3079used by itself to fetch single characters without waiting for the user
3080to hit enter.  For that, try something more like:
3081
3082    if ($BSD_STYLE) {
3083        system "stty cbreak </dev/tty >/dev/tty 2>&1";
3084    }
3085    else {
3086        system "stty", '-icanon', 'eol', "\001";
3087    }
3088
3089    my $key = getc(STDIN);
3090
3091    if ($BSD_STYLE) {
3092        system "stty -cbreak </dev/tty >/dev/tty 2>&1";
3093    }
3094    else {
3095        system 'stty', 'icanon', 'eol', '^@'; # ASCII NUL
3096    }
3097    print "\n";
3098
3099Determination of whether C<$BSD_STYLE> should be set is left as an
3100exercise to the reader.
3101
3102The L<C<POSIX::getattr>|POSIX/C<getattr>> function can do this more
3103portably on systems purporting POSIX compliance.  See also the
3104L<C<Term::ReadKey>|Term::ReadKey> module on CPAN.
3105
3106=item getlogin
3107X<getlogin> X<login>
3108
3109=for Pod::Functions return who logged in at this tty
3110
3111This implements the C library function of the same name, which on most
3112systems returns the current login from F</etc/utmp>, if any.  If it
3113returns the empty string, use L<C<getpwuid>|/getpwuid UID>.
3114
3115    my $login = getlogin || getpwuid($<) || "Kilroy";
3116
3117Do not consider L<C<getlogin>|/getlogin> for authentication: it is not
3118as secure as L<C<getpwuid>|/getpwuid UID>.
3119
3120Portability issues: L<perlport/getlogin>.
3121
3122=item getpeername SOCKET
3123X<getpeername> X<peer>
3124
3125=for Pod::Functions find the other end of a socket connection
3126
3127Returns the packed sockaddr address of the other end of the SOCKET
3128connection.
3129
3130    use Socket;
3131    my $hersockaddr    = getpeername($sock);
3132    my ($port, $iaddr) = sockaddr_in($hersockaddr);
3133    my $herhostname    = gethostbyaddr($iaddr, AF_INET);
3134    my $herstraddr     = inet_ntoa($iaddr);
3135
3136=item getpgrp PID
3137X<getpgrp> X<group>
3138
3139=for Pod::Functions get process group
3140
3141Returns the current process group for the specified PID.  Use
3142a PID of C<0> to get the current process group for the
3143current process.  Will raise an exception if used on a machine that
3144doesn't implement L<getpgrp(2)>.  If PID is omitted, returns the process
3145group of the current process.
3146
3147Some very old machines may not support C<PID != 0> and will throw an
3148exception if C<PID != 0>.
3149
3150Portability issues: L<perlport/getpgrp>.
3151
3152=item getppid
3153X<getppid> X<parent> X<pid>
3154
3155=for Pod::Functions get parent process ID
3156
3157Returns the process id of the parent process.
3158
3159Note for Linux users: Between v5.8.1 and v5.16.0 Perl would work
3160around non-POSIX thread semantics the minority of Linux systems (and
3161Debian GNU/kFreeBSD systems) that used LinuxThreads, this emulation
3162has since been removed.  See the documentation for L<$$|perlvar/$$> for
3163details.
3164
3165Portability issues: L<perlport/getppid>.
3166
3167=item getpriority WHICH,WHO
3168X<getpriority> X<priority> X<nice>
3169
3170=for Pod::Functions get current nice value
3171
3172Returns the current priority for a process, a process group, or a user.
3173(See L<getpriority(2)>.)  Will raise a fatal exception if used on a
3174machine that doesn't implement L<getpriority(2)>.
3175
3176C<WHICH> can be any of C<PRIO_PROCESS>, C<PRIO_PGRP> or C<PRIO_USER>
3177imported from L<POSIX/RESOURCE CONSTANTS>.
3178
3179Portability issues: L<perlport/getpriority>.
3180
3181=item getpwnam NAME
3182X<getpwnam> X<getgrnam> X<gethostbyname> X<getnetbyname> X<getprotobyname>
3183X<getpwuid> X<getgrgid> X<getservbyname> X<gethostbyaddr> X<getnetbyaddr>
3184X<getprotobynumber> X<getservbyport> X<getpwent> X<getgrent> X<gethostent>
3185X<getnetent> X<getprotoent> X<getservent> X<setpwent> X<setgrent> X<sethostent>
3186X<setnetent> X<setprotoent> X<setservent> X<endpwent> X<endgrent> X<endhostent>
3187X<endnetent> X<endprotoent> X<endservent>
3188
3189=for Pod::Functions get passwd record given user login name
3190
3191=item getgrnam NAME
3192
3193=for Pod::Functions get group record given group name
3194
3195=item gethostbyname NAME
3196
3197=for Pod::Functions get host record given name
3198
3199=item getnetbyname NAME
3200
3201=for Pod::Functions get networks record given name
3202
3203=item getprotobyname NAME
3204
3205=for Pod::Functions get protocol record given name
3206
3207=item getpwuid UID
3208
3209=for Pod::Functions get passwd record given user ID
3210
3211=item getgrgid GID
3212
3213=for Pod::Functions get group record given group user ID
3214
3215=item getservbyname NAME,PROTO
3216
3217=for Pod::Functions get services record given its name
3218
3219=item gethostbyaddr ADDR,ADDRTYPE
3220
3221=for Pod::Functions get host record given its address
3222
3223=item getnetbyaddr ADDR,ADDRTYPE
3224
3225=for Pod::Functions get network record given its address
3226
3227=item getprotobynumber NUMBER
3228
3229=for Pod::Functions get protocol record numeric protocol
3230
3231=item getservbyport PORT,PROTO
3232
3233=for Pod::Functions get services record given numeric port
3234
3235=item getpwent
3236
3237=for Pod::Functions get next passwd record
3238
3239=item getgrent
3240
3241=for Pod::Functions get next group record
3242
3243=item gethostent
3244
3245=for Pod::Functions get next hosts record
3246
3247=item getnetent
3248
3249=for Pod::Functions get next networks record
3250
3251=item getprotoent
3252
3253=for Pod::Functions get next protocols record
3254
3255=item getservent
3256
3257=for Pod::Functions get next services record
3258
3259=item setpwent
3260
3261=for Pod::Functions prepare passwd file for use
3262
3263=item setgrent
3264
3265=for Pod::Functions prepare group file for use
3266
3267=item sethostent STAYOPEN
3268
3269=for Pod::Functions prepare hosts file for use
3270
3271=item setnetent STAYOPEN
3272
3273=for Pod::Functions prepare networks file for use
3274
3275=item setprotoent STAYOPEN
3276
3277=for Pod::Functions prepare protocols file for use
3278
3279=item setservent STAYOPEN
3280
3281=for Pod::Functions prepare services file for use
3282
3283=item endpwent
3284
3285=for Pod::Functions be done using passwd file
3286
3287=item endgrent
3288
3289=for Pod::Functions be done using group file
3290
3291=item endhostent
3292
3293=for Pod::Functions be done using hosts file
3294
3295=item endnetent
3296
3297=for Pod::Functions be done using networks file
3298
3299=item endprotoent
3300
3301=for Pod::Functions be done using protocols file
3302
3303=item endservent
3304
3305=for Pod::Functions be done using services file
3306
3307These routines are the same as their counterparts in the
3308system C library.  In list context, the return values from the
3309various get routines are as follows:
3310
3311 #    0        1          2           3         4
3312 my ( $name,   $passwd,   $gid,       $members  ) = getgr*
3313 my ( $name,   $aliases,  $addrtype,  $net      ) = getnet*
3314 my ( $name,   $aliases,  $port,      $proto    ) = getserv*
3315 my ( $name,   $aliases,  $proto                ) = getproto*
3316 my ( $name,   $aliases,  $addrtype,  $length,  @addrs ) = gethost*
3317 my ( $name,   $passwd,   $uid,       $gid,     $quota,
3318    $comment,  $gcos,     $dir,       $shell,   $expire ) = getpw*
3319 #    5        6          7           8         9
3320
3321(If the entry doesn't exist, the return value is a single meaningless true
3322value.)
3323
3324The exact meaning of the $gcos field varies but it usually contains
3325the real name of the user (as opposed to the login name) and other
3326information pertaining to the user.  Beware, however, that in many
3327system users are able to change this information and therefore it
3328cannot be trusted and therefore the $gcos is tainted (see
3329L<perlsec>).  The $passwd and $shell, user's encrypted password and
3330login shell, are also tainted, for the same reason.
3331
3332In scalar context, you get the name, unless the function was a
3333lookup by name, in which case you get the other thing, whatever it is.
3334(If the entry doesn't exist you get the undefined value.)  For example:
3335
3336    my $uid   = getpwnam($name);
3337    my $name  = getpwuid($num);
3338    my $name  = getpwent();
3339    my $gid   = getgrnam($name);
3340    my $name  = getgrgid($num);
3341    my $name  = getgrent();
3342    # etc.
3343
3344In I<getpw*()> the fields $quota, $comment, and $expire are special
3345in that they are unsupported on many systems.  If the
3346$quota is unsupported, it is an empty scalar.  If it is supported, it
3347usually encodes the disk quota.  If the $comment field is unsupported,
3348it is an empty scalar.  If it is supported it usually encodes some
3349administrative comment about the user.  In some systems the $quota
3350field may be $change or $age, fields that have to do with password
3351aging.  In some systems the $comment field may be $class.  The $expire
3352field, if present, encodes the expiration period of the account or the
3353password.  For the availability and the exact meaning of these fields
3354in your system, please consult L<getpwnam(3)> and your system's
3355F<pwd.h> file.  You can also find out from within Perl what your
3356$quota and $comment fields mean and whether you have the $expire field
3357by using the L<C<Config>|Config> module and the values C<d_pwquota>, C<d_pwage>,
3358C<d_pwchange>, C<d_pwcomment>, and C<d_pwexpire>.  Shadow password
3359files are supported only if your vendor has implemented them in the
3360intuitive fashion that calling the regular C library routines gets the
3361shadow versions if you're running under privilege or if there exists
3362the L<shadow(3)> functions as found in System V (this includes Solaris
3363and Linux).  Those systems that implement a proprietary shadow password
3364facility are unlikely to be supported.
3365
3366The $members value returned by I<getgr*()> is a space-separated list of
3367the login names of the members of the group.
3368
3369For the I<gethost*()> functions, if the C<h_errno> variable is supported in
3370C, it will be returned to you via L<C<$?>|perlvar/$?> if the function
3371call fails.  The
3372C<@addrs> value returned by a successful call is a list of raw
3373addresses returned by the corresponding library call.  In the
3374Internet domain, each address is four bytes long; you can unpack it
3375by saying something like:
3376
3377    my ($w,$x,$y,$z) = unpack('W4',$addr[0]);
3378
3379The Socket library makes this slightly easier:
3380
3381    use Socket;
3382    my $iaddr = inet_aton("127.1"); # or whatever address
3383    my $name  = gethostbyaddr($iaddr, AF_INET);
3384
3385    # or going the other way
3386    my $straddr = inet_ntoa($iaddr);
3387
3388In the opposite way, to resolve a hostname to the IP address
3389you can write this:
3390
3391    use Socket;
3392    my $packed_ip = gethostbyname("www.perl.org");
3393    my $ip_address;
3394    if (defined $packed_ip) {
3395        $ip_address = inet_ntoa($packed_ip);
3396    }
3397
3398Make sure L<C<gethostbyname>|/gethostbyname NAME> is called in SCALAR
3399context and that its return value is checked for definedness.
3400
3401The L<C<getprotobynumber>|/getprotobynumber NUMBER> function, even
3402though it only takes one argument, has the precedence of a list
3403operator, so beware:
3404
3405    getprotobynumber $number eq 'icmp'   # WRONG
3406    getprotobynumber($number eq 'icmp')  # actually means this
3407    getprotobynumber($number) eq 'icmp'  # better this way
3408
3409If you get tired of remembering which element of the return list
3410contains which return value, by-name interfaces are provided in standard
3411modules: L<C<File::stat>|File::stat>, L<C<Net::hostent>|Net::hostent>,
3412L<C<Net::netent>|Net::netent>, L<C<Net::protoent>|Net::protoent>,
3413L<C<Net::servent>|Net::servent>, L<C<Time::gmtime>|Time::gmtime>,
3414L<C<Time::localtime>|Time::localtime>, and
3415L<C<User::grent>|User::grent>.  These override the normal built-ins,
3416supplying versions that return objects with the appropriate names for
3417each field.  For example:
3418
3419   use File::stat;
3420   use User::pwent;
3421   my $is_theirs = (stat($filename)->uid == getpwnam($whoever)->uid);
3422
3423Even though it looks as though they're the same method calls (uid),
3424they aren't, because a C<File::stat> object is different from
3425a C<User::pwent> object.
3426
3427Many of these functions are not safe in a multi-threaded environment
3428where more than one thread can be using them.  In particular, functions
3429like C<getpwent()> iterate per-process and not per-thread, so if two
3430threads are simultaneously iterating, neither will get all the records.
3431
3432Some systems have thread-safe versions of some of the functions, such as
3433C<getpwnam_r()> instead of C<getpwnam()>.  There, Perl automatically and
3434invisibly substitutes the thread-safe version, without notice.  This
3435means that code that safely runs on some systems can fail on others that
3436lack the thread-safe versions.
3437
3438Portability issues: L<perlport/getpwnam> to L<perlport/endservent>.
3439
3440=item getsockname SOCKET
3441X<getsockname>
3442
3443=for Pod::Functions retrieve the sockaddr for a given socket
3444
3445Returns the packed sockaddr address of this end of the SOCKET connection,
3446in case you don't know the address because you have several different
3447IPs that the connection might have come in on.
3448
3449    use Socket;
3450    my $mysockaddr = getsockname($sock);
3451    my ($port, $myaddr) = sockaddr_in($mysockaddr);
3452    printf "Connect to %s [%s]\n",
3453       scalar gethostbyaddr($myaddr, AF_INET),
3454       inet_ntoa($myaddr);
3455
3456=item getsockopt SOCKET,LEVEL,OPTNAME
3457X<getsockopt>
3458
3459=for Pod::Functions get socket options on a given socket
3460
3461Queries the option named OPTNAME associated with SOCKET at a given LEVEL.
3462Options may exist at multiple protocol levels depending on the socket
3463type, but at least the uppermost socket level SOL_SOCKET (defined in the
3464L<C<Socket>|Socket> module) will exist.  To query options at another
3465level the protocol number of the appropriate protocol controlling the
3466option should be supplied.  For example, to indicate that an option is
3467to be interpreted by the TCP protocol, LEVEL should be set to the
3468protocol number of TCP, which you can get using
3469L<C<getprotobyname>|/getprotobyname NAME>.
3470
3471The function returns a packed string representing the requested socket
3472option, or L<C<undef>|/undef EXPR> on error, with the reason for the
3473error placed in L<C<$!>|perlvar/$!>.  Just what is in the packed string
3474depends on LEVEL and OPTNAME; consult L<getsockopt(2)> for details.  A
3475common case is that the option is an integer, in which case the result
3476is a packed integer, which you can decode using
3477L<C<unpack>|/unpack TEMPLATE,EXPR> with the C<i> (or C<I>) format.
3478
3479Here's an example to test whether Nagle's algorithm is enabled on a socket:
3480
3481    use Socket qw(:all);
3482
3483    defined(my $tcp = getprotobyname("tcp"))
3484        or die "Could not determine the protocol number for tcp";
3485    # my $tcp = IPPROTO_TCP; # Alternative
3486    my $packed = getsockopt($socket, $tcp, TCP_NODELAY)
3487        or die "getsockopt TCP_NODELAY: $!";
3488    my $nodelay = unpack("I", $packed);
3489    print "Nagle's algorithm is turned ",
3490           $nodelay ? "off\n" : "on\n";
3491
3492Portability issues: L<perlport/getsockopt>.
3493
3494=item glob EXPR
3495X<glob> X<wildcard> X<filename, expansion> X<expand>
3496
3497=item glob
3498
3499=for Pod::Functions expand filenames using wildcards
3500
3501In list context, returns a (possibly empty) list of filename expansions on
3502the value of EXPR such as the Unix shell Bash would do. In
3503scalar context, glob iterates through such filename expansions, returning
3504L<C<undef>|/undef EXPR> when the list is exhausted. If EXPR is omitted,
3505L<C<$_>|perlvar/$_> is used.
3506
3507    # List context
3508    my @txt_files  = glob("*.txt");
3509    my @perl_files = glob("*.pl *.pm");
3510
3511    # Scalar context
3512    while (my $file = glob("*.mp3")) {
3513        # Do stuff
3514    }
3515
3516Glob also supports an alternate syntax using C<< < >> C<< > >> as
3517delimiters. While this syntax is supported, it is recommended that you
3518use C<glob> instead as it is more readable and searchable.
3519
3520    my @txt_files  = <"*.txt">;
3521
3522If you need case insensitive file globbing that can be achieved using the
3523C<:nocase> parameter of the L<C<bsd_glob>|File::Glob/C<bsd_glob>> module.
3524
3525    use File::Glob qw(:globally :nocase);
3526
3527	my @txt = glob("readme*"); # README readme.txt Readme.md
3528
3529Note that L<C<glob>|/glob EXPR> splits its arguments on whitespace and
3530treats
3531each segment as separate pattern.  As such, C<glob("*.c *.h")>
3532matches all files with a F<.c> or F<.h> extension.  The expression
3533C<glob(".* *")> matches all files in the current working directory.
3534If you want to glob filenames that might contain whitespace, you'll
3535have to use extra quotes around the spacey filename to protect it.
3536For example, to glob filenames that have an C<e> followed by a space
3537followed by an C<f>, use one of:
3538
3539    my @spacies = <"*e f*">;
3540    my @spacies = glob('"*e f*"');
3541    my @spacies = glob(q("*e f*"));
3542
3543If you had to get a variable through, you could do this:
3544
3545    my @spacies = glob("'*${var}e f*'");
3546    my @spacies = glob(qq("*${var}e f*"));
3547
3548If non-empty braces are the only wildcard characters used in the
3549L<C<glob>|/glob EXPR>, no filenames are matched, but potentially many
3550strings are returned.  For example, this produces nine strings, one for
3551each pairing of fruits and colors:
3552
3553    my @many = glob("{apple,tomato,cherry}={green,yellow,red}");
3554
3555This operator is implemented using the standard C<File::Glob> extension.
3556See L<C<bsd_glob>|File::Glob/C<bsd_glob>> for details, including
3557L<C<bsd_glob>|File::Glob/C<bsd_glob>>, which does not treat whitespace
3558as a pattern separator.
3559
3560If a C<glob> expression is used as the condition of a C<while> or C<for>
3561loop, then it will be implicitly assigned to C<$_>.  If either a C<glob>
3562expression or an explicit assignment of a C<glob> expression to a scalar
3563is used as a C<while>/C<for> condition, then the condition actually
3564tests for definedness of the expression's value, not for its regular
3565truth value.
3566
3567Internal implemenation details:
3568
3569This is the internal function implementing the C<< <*.c> >> operator,
3570but you can use it directly. The C<< <*.c> >> operator is discussed in
3571more detail in L<perlop/"I/O Operators">.
3572
3573Portability issues: L<perlport/glob>.
3574
3575=item gmtime EXPR
3576X<gmtime> X<UTC> X<Greenwich>
3577
3578=item gmtime
3579
3580=for Pod::Functions convert UNIX time into record or string using Greenwich time
3581
3582Works just like L<C<localtime>|/localtime EXPR>, but the returned values
3583are localized for the standard Greenwich time zone.
3584
3585Note: When called in list context, $isdst, the last value
3586returned by gmtime, is always C<0>.  There is no
3587Daylight Saving Time in GMT.
3588
3589Portability issues: L<perlport/gmtime>.
3590
3591=item goto LABEL
3592X<goto> X<jump> X<jmp>
3593
3594=item goto EXPR
3595
3596=item goto &NAME
3597
3598=for Pod::Functions create spaghetti code
3599
3600The C<goto LABEL> form finds the statement labeled with LABEL and
3601resumes execution there.  It can't be used to get out of a block or
3602subroutine given to L<C<sort>|/sort SUBNAME LIST>.  It can be used to go
3603almost anywhere else within the dynamic scope, including out of
3604subroutines, but it's usually better to use some other construct such as
3605L<C<last>|/last LABEL> or L<C<die>|/die LIST>.  The author of Perl has
3606never felt the need to use this form of L<C<goto>|/goto LABEL> (in Perl,
3607that is; C is another matter).  (The difference is that C does not offer
3608named loops combined with loop control.  Perl does, and this replaces
3609most structured uses of L<C<goto>|/goto LABEL> in other languages.)
3610
3611The C<goto EXPR> form expects to evaluate C<EXPR> to a code reference or
3612a label name.  If it evaluates to a code reference, it will be handled
3613like C<goto &NAME>, below.  This is especially useful for implementing
3614tail recursion via C<goto __SUB__>.
3615
3616If the expression evaluates to a label name, its scope will be resolved
3617dynamically.  This allows for computed L<C<goto>|/goto LABEL>s per
3618FORTRAN, but isn't necessarily recommended if you're optimizing for
3619maintainability:
3620
3621    goto ("FOO", "BAR", "GLARCH")[$i];
3622
3623As shown in this example, C<goto EXPR> is exempt from the "looks like a
3624function" rule.  A pair of parentheses following it does not (necessarily)
3625delimit its argument.  C<goto("NE")."XT"> is equivalent to C<goto NEXT>.
3626Also, unlike most named operators, this has the same precedence as
3627assignment.
3628
3629Use of C<goto LABEL> or C<goto EXPR> to jump into a construct is
3630deprecated and will issue a warning; it will become a fatal error in
3631Perl 5.42. While still available, it may not be used to
3632go into any construct that requires initialization, such as a
3633subroutine, a C<foreach> loop, or a C<given>
3634block.  In general, it may not be used to jump into the parameter
3635of a binary or list operator, but it may be used to jump into the
3636I<first> parameter of a binary operator.  (The C<=>
3637assignment operator's "first" operand is its right-hand
3638operand.)  It also can't be used to go into a
3639construct that is optimized away.
3640
3641The C<goto &NAME> form is quite different from the other forms of
3642L<C<goto>|/goto LABEL>.  In fact, it isn't a goto in the normal sense at
3643all, and doesn't have the stigma associated with other gotos.  Instead,
3644it exits the current subroutine (losing any changes set by
3645L<C<local>|/local EXPR>) and immediately calls in its place the named
3646subroutine using the current value of L<C<@_>|perlvar/@_>.  This is used
3647by C<AUTOLOAD> subroutines that wish to load another subroutine and then
3648pretend that the other subroutine had been called in the first place
3649(except that any modifications to L<C<@_>|perlvar/@_> in the current
3650subroutine are propagated to the other subroutine.) After the
3651L<C<goto>|/goto LABEL>, not even L<C<caller>|/caller EXPR> will be able
3652to tell that this routine was called first.
3653
3654NAME needn't be the name of a subroutine; it can be a scalar variable
3655containing a code reference or a block that evaluates to a code
3656reference.
3657
3658=item grep BLOCK LIST
3659X<grep>
3660
3661=item grep EXPR,LIST
3662
3663=for Pod::Functions locate elements in a list test true against a given criterion
3664
3665This is similar in spirit to, but not the same as, L<grep(1)> and its
3666relatives.  In particular, it is not limited to using regular expressions.
3667
3668Evaluates the BLOCK or EXPR for each element of LIST (locally setting
3669L<C<$_>|perlvar/$_> to each element) and returns the list value
3670consisting of those
3671elements for which the expression evaluated to true.  In scalar
3672context, returns the number of times the expression was true.
3673
3674    my @foo = grep(!/^#/, @bar);    # weed out comments
3675
3676or equivalently,
3677
3678    my @foo = grep {!/^#/} @bar;    # weed out comments
3679
3680Note that L<C<$_>|perlvar/$_> is an alias to the list value, so it can
3681be used to
3682modify the elements of the LIST.  While this is useful and supported,
3683it can cause bizarre results if the elements of LIST are not variables.
3684Similarly, grep returns aliases into the original list, much as a for
3685loop's index variable aliases the list elements.  That is, modifying an
3686element of a list returned by grep (for example, in a C<foreach>,
3687L<C<map>|/map BLOCK LIST> or another L<C<grep>|/grep BLOCK LIST>)
3688actually modifies the element in the original list.
3689This is usually something to be avoided when writing clear code.
3690
3691See also L<C<map>|/map BLOCK LIST> for a list composed of the results of
3692the BLOCK or EXPR.
3693
3694=item hex EXPR
3695X<hex> X<hexadecimal>
3696
3697=item hex
3698
3699=for Pod::Functions convert a hexadecimal string to a number
3700
3701Interprets EXPR as a hex string and returns the corresponding numeric value.
3702If EXPR is omitted, uses L<C<$_>|perlvar/$_>.
3703
3704    print hex '0xAf'; # prints '175'
3705    print hex 'aF';   # same
3706    $valid_input =~ /\A(?:0?[xX])?(?:_?[0-9a-fA-F])*\z/
3707
3708A hex string consists of hex digits and an optional C<0x> or C<x> prefix.
3709Each hex digit may be preceded by a single underscore, which will be ignored.
3710Any other character triggers a warning and causes the rest of the string
3711to be ignored (even leading whitespace, unlike L<C<oct>|/oct EXPR>).
3712Only integers can be represented, and integer overflow triggers a warning.
3713
3714To convert strings that might start with any of C<0>, C<0x>, or C<0b>,
3715see L<C<oct>|/oct EXPR>.  To present something as hex, look into
3716L<C<printf>|/printf FILEHANDLE FORMAT, LIST>,
3717L<C<sprintf>|/sprintf FORMAT, LIST>, and
3718L<C<unpack>|/unpack TEMPLATE,EXPR>.
3719
3720=item import LIST
3721X<import>
3722
3723=for Pod::Functions patch a module's namespace into your own
3724
3725There is no builtin L<C<import>|/import LIST> function.  It is just an
3726ordinary method (subroutine) defined (or inherited) by modules that wish
3727to export names to another module.  The
3728L<C<use>|/use Module VERSION LIST> function calls the
3729L<C<import>|/import LIST> method for the package used.  See also
3730L<C<use>|/use Module VERSION LIST>, L<perlmod>, and L<Exporter>.
3731
3732=item index STR,SUBSTR,POSITION
3733X<index> X<indexOf> X<InStr>
3734
3735=item index STR,SUBSTR
3736
3737=for Pod::Functions find a substring within a string
3738
3739The index function searches for one string within another, but without
3740the wildcard-like behavior of a full regular-expression pattern match.
3741It returns the position of the first occurrence of SUBSTR in STR at
3742or after POSITION.  If POSITION is omitted, starts searching from the
3743beginning of the string.  POSITION before the beginning of the string
3744or after its end is treated as if it were the beginning or the end,
3745respectively.  POSITION and the return value are based at zero.
3746If the substring is not found, L<C<index>|/index STR,SUBSTR,POSITION>
3747returns -1.
3748
3749Find characters or strings:
3750
3751    index("Perl is great", "P");     # Returns 0
3752    index("Perl is great", "g");     # Returns 8
3753    index("Perl is great", "great"); # Also returns 8
3754
3755Attempting to find something not there:
3756
3757    index("Perl is great", "Z");     # Returns -1 (not found)
3758
3759Using an offset to find the I<second> occurrence:
3760
3761    index("Perl is great", "e", 5);  # Returns 10
3762
3763=item int EXPR
3764X<int> X<integer> X<truncate> X<trunc> X<floor>
3765
3766=item int
3767
3768=for Pod::Functions get the integer portion of a number
3769
3770Returns the integer portion of EXPR.  If EXPR is omitted, uses
3771L<C<$_>|perlvar/$_>.
3772You should not use this function for rounding: one because it truncates
3773towards C<0>, and two because machine representations of floating-point
3774numbers can sometimes produce counterintuitive results.  For example,
3775C<int(-6.725/0.025)> produces -268 rather than the correct -269; that's
3776because it's really more like -268.99999999999994315658 instead.  Usually,
3777the L<C<sprintf>|/sprintf FORMAT, LIST>,
3778L<C<printf>|/printf FILEHANDLE FORMAT, LIST>, or the
3779L<C<POSIX::floor>|POSIX/C<floor>> and L<C<POSIX::ceil>|POSIX/C<ceil>>
3780functions will serve you better than will L<C<int>|/int EXPR>.
3781
3782=item ioctl FILEHANDLE,FUNCTION,SCALAR
3783X<ioctl>
3784
3785=for Pod::Functions system-dependent device control system call
3786
3787Implements the L<ioctl(2)> function.  You'll probably first have to say
3788
3789    require "sys/ioctl.ph";  # probably in
3790                             # $Config{archlib}/sys/ioctl.ph
3791
3792to get the correct function definitions.  If F<sys/ioctl.ph> doesn't
3793exist or doesn't have the correct definitions you'll have to roll your
3794own, based on your C header files such as F<< <sys/ioctl.h> >>.
3795(There is a Perl script called B<h2ph> that comes with the Perl kit that
3796may help you in this, but it's nontrivial.)  SCALAR will be read and/or
3797written depending on the FUNCTION; a C pointer to the string value of SCALAR
3798will be passed as the third argument of the actual
3799L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR> call.  (If SCALAR
3800has no string value but does have a numeric value, that value will be
3801passed rather than a pointer to the string value.  To guarantee this to be
3802true, add a C<0> to the scalar before using it.)  The
3803L<C<pack>|/pack TEMPLATE,LIST> and L<C<unpack>|/unpack TEMPLATE,EXPR>
3804functions may be needed to manipulate the values of structures used by
3805L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR>.
3806
3807The return value of L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR> (and
3808L<C<fcntl>|/fcntl FILEHANDLE,FUNCTION,SCALAR>) is as follows:
3809
3810    if OS returns:      then Perl returns:
3811        -1               undefined value
3812         0              string "0 but true"
3813    anything else           that number
3814
3815Thus Perl returns true on success and false on failure, yet you can
3816still easily determine the actual value returned by the operating
3817system:
3818
3819    my $retval = ioctl(...) || -1;
3820    printf "System returned %d\n", $retval;
3821
3822The special string C<"0 but true"> is exempt from
3823L<C<Argument "..." isn't numeric>|perldiag/Argument "%s" isn't numeric%s>
3824L<warnings> on improper numeric conversions.
3825
3826Portability issues: L<perlport/ioctl>.
3827
3828=item join EXPR,LIST
3829X<join>
3830
3831=for Pod::Functions join a list into a string using a separator
3832
3833Joins the separate strings of LIST into a single string with fields
3834separated by the value of EXPR, and returns that new string.  Example:
3835
3836   my $rec = join(':', $login,$passwd,$uid,$gid,$gcos,$home,$shell);
3837
3838Beware that unlike L<C<split>|/split E<sol>PATTERNE<sol>,EXPR,LIMIT>,
3839L<C<join>|/join EXPR,LIST> doesn't take a pattern as its first argument.
3840Compare L<C<split>|/split E<sol>PATTERNE<sol>,EXPR,LIMIT>.
3841
3842=item keys HASH
3843X<keys> X<key>
3844
3845=item keys ARRAY
3846
3847=for Pod::Functions retrieve list of indices from a hash
3848
3849Called in list context, returns a list consisting of all the keys of the
3850named hash, or in Perl 5.12 or later only, the indices of an array.  Perl
3851releases prior to 5.12 will produce a syntax error if you try to use an
3852array argument.  In scalar context, returns the number of keys or indices.
3853
3854Hash entries are returned in an apparently random order.  The actual random
3855order is specific to a given hash; the exact same series of operations
3856on two hashes may result in a different order for each hash.  Any insertion
3857into the hash may change the order, as will any deletion, with the exception
3858that the most recent key returned by L<C<each>|/each HASH> or
3859L<C<keys>|/keys HASH> may be deleted without changing the order.  So
3860long as a given hash is unmodified you may rely on
3861L<C<keys>|/keys HASH>, L<C<values>|/values HASH> and L<C<each>|/each
3862HASH> to repeatedly return the same order
3863as each other.  See L<perlsec/"Algorithmic Complexity Attacks"> for
3864details on why hash order is randomized.  Aside from the guarantees
3865provided here the exact details of Perl's hash algorithm and the hash
3866traversal order are subject to change in any release of Perl.  Tied hashes
3867may behave differently to Perl's hashes with respect to changes in order on
3868insertion and deletion of items.
3869
3870As a side effect, calling L<C<keys>|/keys HASH> resets the internal
3871iterator of the HASH or ARRAY (see L<C<each>|/each HASH>) before
3872yielding the keys.  In
3873particular, calling L<C<keys>|/keys HASH> in void context resets the
3874iterator with no other overhead.
3875
3876Here is yet another way to print your environment:
3877
3878    my @keys = keys %ENV;
3879    my @values = values %ENV;
3880    while (@keys) {
3881        print pop(@keys), '=', pop(@values), "\n";
3882    }
3883
3884or how about sorted by key:
3885
3886    foreach my $key (sort(keys %ENV)) {
3887        print $key, '=', $ENV{$key}, "\n";
3888    }
3889
3890The returned values are copies of the original keys in the hash, so
3891modifying them will not affect the original hash.  Compare
3892L<C<values>|/values HASH>.
3893
3894To sort a hash by value, you'll need to use a
3895L<C<sort>|/sort SUBNAME LIST> function.  Here's a descending numeric
3896sort of a hash by its values:
3897
3898    foreach my $key (sort { $hash{$b} <=> $hash{$a} } keys %hash) {
3899        printf "%4d %s\n", $hash{$key}, $key;
3900    }
3901
3902Used as an lvalue, L<C<keys>|/keys HASH> allows you to increase the
3903number of hash buckets
3904allocated for the given hash.  This can gain you a measure of efficiency if
3905you know the hash is going to get big.  (This is similar to pre-extending
3906an array by assigning a larger number to $#array.)  If you say
3907
3908    keys %hash = 200;
3909
3910then C<%hash> will have at least 200 buckets allocated for it--256 of them,
3911in fact, since it rounds up to the next power of two.  These
3912buckets will be retained even if you do C<%hash = ()>, use C<undef
3913%hash> if you want to free the storage while C<%hash> is still in scope.
3914You can't shrink the number of buckets allocated for the hash using
3915L<C<keys>|/keys HASH> in this way (but you needn't worry about doing
3916this by accident, as trying has no effect).  C<keys @array> in an lvalue
3917context is a syntax error.
3918
3919Starting with Perl 5.14, an experimental feature allowed
3920L<C<keys>|/keys HASH> to take a scalar expression. This experiment has
3921been deemed unsuccessful, and was removed as of Perl 5.24.
3922
3923To avoid confusing would-be users of your code who are running earlier
3924versions of Perl with mysterious syntax errors, put this sort of thing at
3925the top of your file to signal that your code will work I<only> on Perls of
3926a recent vintage:
3927
3928    use v5.12;	# so keys/values/each work on arrays
3929
3930See also L<C<each>|/each HASH>, L<C<values>|/values HASH>, and
3931L<C<sort>|/sort SUBNAME LIST>.
3932
3933=item kill SIGNAL, LIST
3934
3935=item kill SIGNAL
3936X<kill> X<signal>
3937
3938=for Pod::Functions send a signal to a process or process group
3939
3940Sends a signal to a list of processes.  Returns the number of arguments
3941that were successfully used to signal (which is not necessarily the same
3942as the number of processes actually killed, e.g. where a process group is
3943killed).
3944
3945    my $cnt = kill 'HUP', $child1, $child2;
3946    kill 'KILL', @goners;
3947
3948SIGNAL may be either a signal name (a string) or a signal number.  A signal
3949name may start with a C<SIG> prefix, thus C<FOO> and C<SIGFOO> refer to the
3950same signal.  The string form of SIGNAL is recommended for portability because
3951the same signal may have different numbers in different operating systems.
3952
3953A list of signal names supported by the current platform can be found in
3954C<$Config{sig_name}>, which is provided by the L<C<Config>|Config>
3955module.  See L<Config> for more details.
3956
3957A negative signal name is the same as a negative signal number, killing process
3958groups instead of processes.  For example, C<kill '-KILL', $pgrp> and
3959C<kill -9, $pgrp> will send C<SIGKILL> to
3960the entire process group specified.  That
3961means you usually want to use positive not negative signals.
3962
3963If SIGNAL is either the number 0 or the string C<ZERO> (or C<SIGZERO>),
3964no signal is sent to the process, but L<C<kill>|/kill SIGNAL, LIST>
3965checks whether it's I<possible> to send a signal to it
3966(that means, to be brief, that the process is owned by the same user, or we are
3967the super-user).  This is useful to check that a child process is still
3968alive (even if only as a zombie) and hasn't changed its UID.  See
3969L<perlport> for notes on the portability of this construct.
3970
3971The behavior of kill when a I<PROCESS> number is zero or negative depends on
3972the operating system.  For example, on POSIX-conforming systems, zero will
3973signal the current process group, -1 will signal all processes, and any
3974other negative PROCESS number will act as a negative signal number and
3975kill the entire process group specified.
3976
3977If both the SIGNAL and the PROCESS are negative, the results are undefined.
3978A warning may be produced in a future version.
3979
3980See L<perlipc/"Signals"> for more details.
3981
3982On some platforms such as Windows where the L<fork(2)> system call is not
3983available, Perl can be built to emulate L<C<fork>|/fork> at the
3984interpreter level.
3985This emulation has limitations related to kill that have to be considered,
3986for code running on Windows and in code intended to be portable.
3987
3988See L<perlfork> for more details.
3989
3990If there is no I<LIST> of processes, no signal is sent, and the return
3991value is 0.  This form is sometimes used, however, because it causes
3992tainting checks to be run, if your perl support taint checks.  But see
3993L<perlsec/Laundering and Detecting Tainted Data>.
3994
3995Portability issues: L<perlport/kill>.
3996
3997=item last LABEL
3998X<last> X<break>
3999
4000=item last EXPR
4001
4002=item last
4003
4004=for Pod::Functions exit a block prematurely
4005
4006The L<C<last>|/last LABEL> command is like the C<break> statement in C
4007(as used in
4008loops); it immediately exits the loop in question.  If the LABEL is
4009omitted, the command refers to the innermost enclosing
4010loop.  The C<last EXPR> form, available starting in Perl
40115.18.0, allows a label name to be computed at run time,
4012and is otherwise identical to C<last LABEL>.  The
4013L<C<continue>|/continue BLOCK> block, if any, is not executed:
4014
4015    LINE: while (<STDIN>) {
4016        last LINE if /^$/;  # exit when done with header
4017        #...
4018    }
4019
4020L<C<last>|/last LABEL> cannot return a value from a block that typically
4021returns a value, such as C<eval {}>, C<sub {}>, or C<do {}>. It will perform
4022its flow control behavior, which precludes any return value. It should not be
4023used to exit a L<C<grep>|/grep BLOCK LIST> or L<C<map>|/map BLOCK LIST>
4024operation.
4025
4026Note that a block by itself is semantically identical to a loop
4027that executes once.  Thus L<C<last>|/last LABEL> can be used to effect
4028an early exit out of such a block.
4029
4030See also L<C<continue>|/continue BLOCK> for an illustration of how
4031L<C<last>|/last LABEL>, L<C<next>|/next LABEL>, and
4032L<C<redo>|/redo LABEL> work.
4033
4034Unlike most named operators, this has the same precedence as assignment.
4035It is also exempt from the looks-like-a-function rule, so
4036C<last ("foo")."bar"> will cause "bar" to be part of the argument to
4037L<C<last>|/last LABEL>.
4038
4039=item lc EXPR
4040X<lc> X<lowercase>
4041
4042=item lc
4043
4044=for Pod::Functions return lower-case version of a string
4045
4046Returns a lowercased version of EXPR.  If EXPR is omitted, uses
4047L<C<$_>|perlvar/$_>.
4048
4049    my $str = lc("Perl is GREAT"); # "perl is great"
4050
4051What gets returned depends on several factors:
4052
4053=over
4054
4055=item If C<use bytes> is in effect:
4056
4057The results follow ASCII rules.  Only the characters C<A-Z> change,
4058to C<a-z> respectively.
4059
4060=item Otherwise, if C<use locale> for C<LC_CTYPE> is in effect:
4061
4062Respects current C<LC_CTYPE> locale for code points < 256; and uses Unicode
4063rules for the remaining code points (this last can only happen if
4064the UTF8 flag is also set).  See L<perllocale>.
4065
4066Starting in v5.20, Perl uses full Unicode rules if the locale is
4067UTF-8.  Otherwise, there is a deficiency in this scheme, which is that
4068case changes that cross the 255/256
4069boundary are not well-defined.  For example, the lower case of LATIN CAPITAL
4070LETTER SHARP S (U+1E9E) in Unicode rules is U+00DF (on ASCII
4071platforms).   But under C<use locale> (prior to v5.20 or not a UTF-8
4072locale), the lower case of U+1E9E is
4073itself, because 0xDF may not be LATIN SMALL LETTER SHARP S in the
4074current locale, and Perl has no way of knowing if that character even
4075exists in the locale, much less what code point it is.  Perl returns
4076a result that is above 255 (almost always the input character unchanged),
4077for all instances (and there aren't many) where the 255/256 boundary
4078would otherwise be crossed; and starting in v5.22, it raises a
4079L<locale|perldiag/Can't do %s("%s") on non-UTF-8 locale; resolved to "%s".> warning.
4080
4081=item Otherwise, If EXPR has the UTF8 flag set:
4082
4083Unicode rules are used for the case change.
4084
4085=item Otherwise, if C<use feature 'unicode_strings'> or C<use locale ':not_characters'> is in effect:
4086
4087Unicode rules are used for the case change.
4088
4089=item Otherwise:
4090
4091ASCII rules are used for the case change.  The lowercase of any character
4092outside the ASCII range is the character itself.
4093
4094=back
4095
4096B<Note:> This is the internal function implementing the
4097L<C<\L>|perlop/"Quote and Quote-like Operators"> escape in double-quoted
4098strings.
4099
4100    my $str = "Perl is \LGREAT\E"; # "Perl is great"
4101
4102=item lcfirst EXPR
4103X<lcfirst> X<lowercase>
4104
4105=item lcfirst
4106
4107=for Pod::Functions return a string with just the next letter in lower case
4108
4109Returns the value of EXPR with the first character lowercased.  This
4110is the internal function implementing the C<\l> escape in
4111double-quoted strings.
4112
4113If EXPR is omitted, uses L<C<$_>|perlvar/$_>.
4114
4115This function behaves the same way under various pragmas, such as in a locale,
4116as L<C<lc>|/lc EXPR> does.
4117
4118=item length EXPR
4119X<length> X<size>
4120
4121=item length
4122
4123=for Pod::Functions return the number of characters in a string
4124
4125Returns the length in I<characters> of the value of EXPR.  If EXPR is
4126omitted, returns the length of L<C<$_>|perlvar/$_>.  If EXPR is
4127undefined, returns L<C<undef>|/undef EXPR>.
4128
4129This function cannot be used on an entire array or hash to find out how
4130many elements these have.  For that, use C<scalar @array> and C<scalar keys
4131%hash>, respectively.
4132
4133Like all Perl character operations, L<C<length>|/length EXPR> normally
4134deals in logical
4135characters, not physical bytes.  For how many bytes a string encoded as
4136UTF-8 would take up, use C<length(Encode::encode('UTF-8', EXPR))>
4137(you'll have to C<use Encode> first).  See L<Encode> and L<perlunicode>.
4138
4139=item __LINE__
4140X<__LINE__>
4141
4142=for Pod::Functions the current source line number
4143
4144A special token that compiles to the current line number.
4145It can be altered by the mechanism described at
4146L<perlsyn/"Plain Old Comments (Not!)">.
4147
4148=item link OLDFILE,NEWFILE
4149X<link>
4150
4151=for Pod::Functions create a hard link in the filesystem
4152
4153Creates a new filename linked to the old filename.  Returns true for
4154success, false otherwise.
4155
4156Portability issues: L<perlport/link>.
4157
4158=item listen SOCKET,QUEUESIZE
4159X<listen>
4160
4161=for Pod::Functions register your socket as a server
4162
4163Does the same thing that the L<listen(2)> system call does.  Returns true if
4164it succeeded, false otherwise.  See the example in
4165L<perlipc/"Sockets: Client/Server Communication">.
4166
4167=item local EXPR
4168X<local>
4169
4170=for Pod::Functions create a temporary value for a global variable (dynamic scoping)
4171
4172You really probably want to be using L<C<my>|/my VARLIST> instead,
4173because L<C<local>|/local EXPR> isn't what most people think of as
4174"local".  See L<perlsub/"Private Variables via my()"> for details.
4175
4176A local modifies the listed variables to be local to the enclosing
4177block, file, or eval.  If more than one value is listed, the list must
4178be placed in parentheses.  See L<perlsub/"Temporary Values via local()">
4179for details, including issues with tied arrays and hashes.
4180
4181Like L<C<my>|/my VARLIST>, L<C<state>|/state VARLIST>, and
4182L<C<our>|/our VARLIST>, L<C<local>|/local EXPR> can operate on a variable
4183anywhere it appears in an expression (aside from interpolation in strings).
4184Unlike the other declarations, the effect of L<C<local>|/local EXPR> happens
4185at runtime, and so it will apply to additional uses of the same variable
4186executed after the declaration, even within the same statement. Note that
4187this does not include uses within an expression assigned to the variable
4188when it is localized, because the assigned expression is evaluated before
4189the localization.
4190
4191    package main;
4192    our $x = 2;
4193    {
4194      foo($x, local $x = $x + 1, $x); # foo() receives (2, 3, 3)
4195      # $main::x is 3 within the call to foo()
4196    }
4197    foo($x); # foo() receives (2) and $main::x is 2
4198
4199The C<delete local EXPR> construct can also be used to localize the deletion
4200of array/hash elements to the current block.
4201See L<perlsub/"Localized deletion of elements of composite types">.
4202
4203=item localtime EXPR
4204X<localtime> X<ctime>
4205
4206=item localtime
4207
4208=for Pod::Functions convert UNIX time into record or string using local time
4209
4210Converts a time as returned by the time function to a 9-element list
4211with the time analyzed for the local time zone.  Typically used as
4212follows:
4213
4214    #     0    1    2     3     4    5     6     7     8
4215    my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
4216                                                localtime(time);
4217
4218All list elements are numeric and come straight out of the C 'struct
4219tm'.  C<$sec>, C<$min>, and C<$hour> are the seconds, minutes, and hours
4220of the specified time.
4221
4222C<$mday> is the day of the month and C<$mon> the month in
4223the range C<0..11>, with 0 indicating January and 11 indicating December.
4224This makes it easy to get a month name from a list:
4225
4226    my @abbr = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
4227    print "$abbr[$mon] $mday";
4228    # $mon=9, $mday=18 gives "Oct 18"
4229
4230C<$year> contains the number of years since 1900.  To get the full
4231year write:
4232
4233    $year += 1900;
4234
4235To get the last two digits of the year (e.g., "01" in 2001) do:
4236
4237    $year = sprintf("%02d", $year % 100);
4238
4239C<$wday> is the day of the week, with 0 indicating Sunday and 3 indicating
4240Wednesday.  C<$yday> is the day of the year, in the range C<0..364>
4241(or C<0..365> in leap years.)
4242
4243C<$isdst> is true if the specified time occurs when Daylight Saving
4244Time is in effect, false otherwise.
4245
4246If EXPR is omitted, L<C<localtime>|/localtime EXPR> uses the current
4247time (as returned by L<C<time>|/time>).
4248
4249In scalar context, L<C<localtime>|/localtime EXPR> returns the
4250L<ctime(3)> value:
4251
4252 my $now_string = localtime;  # e.g., "Thu Oct 13 04:54:34 1994"
4253
4254This scalar value is always in English, and is B<not> locale-dependent.
4255To get similar but locale-dependent date strings, try for example:
4256
4257 use POSIX qw(strftime);
4258 my $now_string = strftime "%a %b %e %H:%M:%S %Y", localtime;
4259 # or for GMT formatted appropriately for your locale:
4260 my $now_string = strftime "%a %b %e %H:%M:%S %Y", gmtime;
4261
4262C<$now_string> will be formatted according to the current LC_TIME locale
4263the program or thread is running in.  See L<perllocale> for how to set
4264up and change that locale.  Note that C<%a> and C<%b>, the short forms
4265of the day of the week and the month of the year, may not necessarily be
4266three characters wide.
4267
4268The L<Time::gmtime> and L<Time::localtime> modules provide a convenient,
4269by-name access mechanism to the L<C<gmtime>|/gmtime EXPR> and
4270L<C<localtime>|/localtime EXPR> functions, respectively.
4271
4272For a comprehensive date and time representation look at the
4273L<DateTime> module on CPAN.
4274
4275For GMT instead of local time use the L<C<gmtime>|/gmtime EXPR> builtin.
4276
4277See also the L<C<Time::Local>|Time::Local> module (for converting
4278seconds, minutes, hours, and such back to the integer value returned by
4279L<C<time>|/time>), and the L<POSIX> module's
4280L<C<mktime>|POSIX/C<mktime>> function.
4281
4282Portability issues: L<perlport/localtime>.
4283
4284=item lock THING
4285X<lock>
4286
4287=for Pod::Functions +5.005 get a thread lock on a variable, subroutine, or method
4288
4289This function places an advisory lock on a shared variable or referenced
4290object contained in I<THING> until the lock goes out of scope.
4291
4292The value returned is the scalar itself, if the argument is a scalar, or a
4293reference, if the argument is a hash, array or subroutine.
4294
4295L<C<lock>|/lock THING> is a "weak keyword"; this means that if you've
4296defined a function
4297by this name (before any calls to it), that function will be called
4298instead.  If you are not under C<use threads::shared> this does nothing.
4299See L<threads::shared>.
4300
4301=item log EXPR
4302X<log> X<logarithm> X<e> X<ln> X<base>
4303
4304=item log
4305
4306=for Pod::Functions retrieve the natural logarithm for a number
4307
4308Returns the natural logarithm (base I<e>) of EXPR.  If EXPR is omitted,
4309returns the log of L<C<$_>|perlvar/$_>.  To get the
4310log of another base, use basic algebra:
4311The base-N log of a number is equal to the natural log of that number
4312divided by the natural log of N.  For example:
4313
4314    sub log10 {
4315        my $n = shift;
4316        return log($n)/log(10);
4317    }
4318
4319See also L<C<exp>|/exp EXPR> for the inverse operation.
4320
4321=item lstat FILEHANDLE
4322X<lstat>
4323
4324=item lstat EXPR
4325
4326=item lstat DIRHANDLE
4327
4328=item lstat
4329
4330=for Pod::Functions stat a symbolic link
4331
4332Does the same thing as the L<C<stat>|/stat FILEHANDLE> function
4333(including setting the special C<_> filehandle) but stats a symbolic
4334link instead of the file the symbolic link points to.  If symbolic links
4335are unimplemented on your system, a normal L<C<stat>|/stat FILEHANDLE>
4336is done.  For much more detailed information, please see the
4337documentation for L<C<stat>|/stat FILEHANDLE>.
4338
4339If EXPR is omitted, stats L<C<$_>|perlvar/$_>.
4340
4341Portability issues: L<perlport/lstat>.
4342
4343=item m//
4344
4345=for Pod::Functions match a string with a regular expression pattern
4346
4347The match operator.  See L<perlop/"Regexp Quote-Like Operators">.
4348
4349=item map BLOCK LIST
4350X<map>
4351
4352=item map EXPR,LIST
4353
4354=for Pod::Functions apply a change to a list to get back a new list with the changes
4355
4356Evaluates the BLOCK or EXPR for each element of LIST (locally setting
4357L<C<$_>|perlvar/$_> to each element) and composes a list of the results of
4358each such evaluation.  Each element of LIST may produce zero, one, or more
4359elements in the generated list, so the number of elements in the generated
4360list may differ from that in LIST.  In scalar context, returns the total
4361number of elements so generated.  In list context, returns the generated list.
4362
4363    my @chars = map(chr, @numbers);
4364
4365translates a list of numbers to the corresponding characters.
4366
4367    my @squares = map { $_ * $_ } @numbers;
4368
4369translates a list of numbers to their squared values.
4370
4371    my @squares = map { $_ > 5 ? ($_ * $_) : () } @numbers;
4372
4373shows that number of returned elements can differ from the number of
4374input elements.  To omit an element, return an empty list ().
4375This could also be achieved by writing
4376
4377    my @squares = map { $_ * $_ } grep { $_ > 5 } @numbers;
4378
4379which makes the intention more clear.
4380
4381Map always returns a list, which can be
4382assigned to a hash such that the elements
4383become key/value pairs.  See L<perldata> for more details.
4384
4385    my %hash = map { get_a_key_for($_) => $_ } @array;
4386
4387is just a funny way to write
4388
4389    my %hash;
4390    foreach (@array) {
4391        $hash{get_a_key_for($_)} = $_;
4392    }
4393
4394Note that L<C<$_>|perlvar/$_> is an alias to the list value, so it can
4395be used to modify the elements of the LIST.  While this is useful and
4396supported, it can cause bizarre results if the elements of LIST are not
4397variables.  Using a regular C<foreach> loop for this purpose would be
4398clearer in most cases.  See also L<C<grep>|/grep BLOCK LIST> for a
4399list composed of those items of the original list for which the BLOCK
4400or EXPR evaluates to true.
4401
4402C<{> starts both hash references and blocks, so C<map { ...> could be either
4403the start of map BLOCK LIST or map EXPR, LIST.  Because Perl doesn't look
4404ahead for the closing C<}> it has to take a guess at which it's dealing with
4405based on what it finds just after the
4406C<{>.  Usually it gets it right, but if it
4407doesn't it won't realize something is wrong until it gets to the C<}> and
4408encounters the missing (or unexpected) comma.  The syntax error will be
4409reported close to the C<}>, but you'll need to change something near the C<{>
4410such as using a unary C<+> or semicolon to give Perl some help:
4411
4412 my %hash = map {  "\L$_" => 1  } @array # perl guesses EXPR. wrong
4413 my %hash = map { +"\L$_" => 1  } @array # perl guesses BLOCK. right
4414 my %hash = map {; "\L$_" => 1  } @array # this also works
4415 my %hash = map { ("\L$_" => 1) } @array # as does this
4416 my %hash = map {  lc($_) => 1  } @array # and this.
4417 my %hash = map +( lc($_) => 1 ), @array # this is EXPR and works!
4418
4419 my %hash = map  ( lc($_), 1 ),   @array # evaluates to (1, @array)
4420
4421or to force an anon hash constructor use C<+{>:
4422
4423    my @hashes = map +{ lc($_) => 1 }, @array # EXPR, so needs
4424                                              # comma at end
4425
4426to get a list of anonymous hashes each with only one entry apiece.
4427
4428=item method NAME BLOCK
4429X<method>
4430
4431=item method NAME : ATTRS BLOCK
4432
4433=for Pod::Functions declare a method of a class
4434
4435Creates a new named method in the scope of the class that it appears within.
4436This is only valid inside a L<C<class>|/class NAMESPACE> declaration.
4437
4438=item mkdir FILENAME,MODE
4439X<mkdir> X<md> X<directory, create>
4440
4441=item mkdir FILENAME
4442
4443=item mkdir
4444
4445=for Pod::Functions create a directory
4446
4447Creates the directory specified by FILENAME, with permissions
4448specified by MODE (as modified by L<C<umask>|/umask EXPR>).  If it
4449succeeds it returns true; otherwise it returns false and sets
4450L<C<$!>|perlvar/$!> (errno).
4451MODE defaults to 0777 if omitted, and FILENAME defaults
4452to L<C<$_>|perlvar/$_> if omitted.
4453
4454In general, it is better to create directories with a permissive MODE
4455and let the user modify that with their L<C<umask>|/umask EXPR> than it
4456is to supply
4457a restrictive MODE and give the user no way to be more permissive.
4458The exceptions to this rule are when the file or directory should be
4459kept private (mail files, for instance).  The documentation for
4460L<C<umask>|/umask EXPR> discusses the choice of MODE in more detail.
4461If bits in MODE other than the permission bits are set, the result may
4462be implementation defined, per POSIX 1003.1-2008.
4463
4464Note that according to the POSIX 1003.1-1996 the FILENAME may have any
4465number of trailing slashes.  Some operating and filesystems do not get
4466this right, so Perl automatically removes all trailing slashes to keep
4467everyone happy.
4468
4469To recursively create a directory structure, look at
4470the L<C<make_path>|File::Path/make_path( $dir1, $dir2, .... )> function
4471of the L<File::Path> module.
4472
4473=item msgctl ID,CMD,ARG
4474X<msgctl>
4475
4476=for Pod::Functions SysV IPC message control operations
4477
4478Calls the System V IPC function L<msgctl(2)>.  You'll probably have to say
4479
4480    use IPC::SysV;
4481
4482first to get the correct constant definitions.  If CMD is C<IPC_STAT>,
4483then ARG must be a variable that will hold the returned C<msqid_ds>
4484structure.  Returns like L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR>:
4485the undefined value for error, C<"0 but true"> for zero, or the actual
4486return value otherwise.  See also L<perlipc/"SysV IPC"> and the
4487documentation for L<C<IPC::SysV>|IPC::SysV> and
4488L<C<IPC::Semaphore>|IPC::Semaphore>.
4489
4490Portability issues: L<perlport/msgctl>.
4491
4492=item msgget KEY,FLAGS
4493X<msgget>
4494
4495=for Pod::Functions get SysV IPC message queue
4496
4497Calls the System V IPC function L<msgget(2)>.  Returns the message queue
4498id, or L<C<undef>|/undef EXPR> on error.  See also L<perlipc/"SysV IPC">
4499and the documentation for L<C<IPC::SysV>|IPC::SysV> and
4500L<C<IPC::Msg>|IPC::Msg>.
4501
4502Portability issues: L<perlport/msgget>.
4503
4504=item msgrcv ID,VAR,SIZE,TYPE,FLAGS
4505X<msgrcv>
4506
4507=for Pod::Functions receive a SysV IPC message from a message queue
4508
4509Calls the System V IPC function msgrcv to receive a message from
4510message queue ID into variable VAR with a maximum message size of
4511SIZE.  Note that when a message is received, the message type as a
4512native long integer will be the first thing in VAR, followed by the
4513actual message.  This packing may be opened with C<unpack("l! a*")>.
4514Taints the variable.  Returns true if successful, false
4515on error.  See also L<perlipc/"SysV IPC"> and the documentation for
4516L<C<IPC::SysV>|IPC::SysV> and L<C<IPC::Msg>|IPC::Msg>.
4517
4518Portability issues: L<perlport/msgrcv>.
4519
4520=item msgsnd ID,MSG,FLAGS
4521X<msgsnd>
4522
4523=for Pod::Functions send a SysV IPC message to a message queue
4524
4525Calls the System V IPC function msgsnd to send the message MSG to the
4526message queue ID.  MSG must begin with the native long integer message
4527type, followed by the message itself.  This kind of packing can be achieved
4528with C<pack("l! a*", $type, $message)>.  Returns true if successful,
4529false on error.  See also L<perlipc/"SysV IPC"> and the documentation
4530for L<C<IPC::SysV>|IPC::SysV> and L<C<IPC::Msg>|IPC::Msg>.
4531
4532Portability issues: L<perlport/msgsnd>.
4533
4534=item my VARLIST
4535X<my>
4536
4537=item my TYPE VARLIST
4538
4539=item my VARLIST : ATTRS
4540
4541=item my TYPE VARLIST : ATTRS
4542
4543=for Pod::Functions declare and assign a local variable (lexical scoping)
4544
4545A L<C<my>|/my VARLIST> declares the listed variables to be local
4546(lexically) to the enclosing block, file, or L<C<eval>|/eval EXPR>.  If
4547more than one variable is listed, the list must be placed in
4548parentheses.
4549
4550Note that with a parenthesised list, L<C<undef>|/undef EXPR> can be used
4551as a dummy placeholder, for example to skip assignment of initial
4552values:
4553
4554    my ( undef, $min, $hour ) = localtime;
4555
4556Like L<C<state>|/state VARLIST>, L<C<local>|/local EXPR>, and
4557L<C<our>|/our VARLIST>, L<C<my>|/my VARLIST> can operate on a variable
4558anywhere it appears in an expression (aside from interpolation in strings).
4559The declaration will not apply to additional uses of the same variable until
4560the next statement. This means additional uses of that variable within the
4561same statement will act as they would have before that declaration occurred,
4562or result in a strict 'vars' error, as appropriate.
4563
4564    package main;
4565    our $x = 2;
4566    foo($x, my $x = $x + 1, $x); # foo() receives (2, 3, 2)
4567    foo($x, $main::x);           # foo() receives (3, 2)
4568
4569Redeclaring a variable in the same scope or statement will "shadow" the
4570previous declaration, creating a new instance and preventing access to
4571the previous one. This is usually undesired and, if warnings are enabled,
4572will result in a warning in the C<shadow> category.
4573
4574The exact semantics and interface of TYPE and ATTRS are still
4575evolving.  TYPE may be a bareword, a constant declared
4576with L<C<use constant>|constant>, or L<C<__PACKAGE__>|/__PACKAGE__>.  It
4577is
4578currently bound to the use of the L<fields> pragma,
4579and attributes are handled using the L<attributes> pragma, or starting
4580from Perl 5.8.0 also via the L<Attribute::Handlers> module.  See
4581L<perlsub/"Private Variables via my()"> for details.
4582
4583=item next LABEL
4584X<next> X<continue>
4585
4586=item next EXPR
4587
4588=item next
4589
4590=for Pod::Functions iterate a block prematurely
4591
4592The L<C<next>|/next LABEL> command is like the C<continue> statement in
4593C; it starts the next iteration of the loop:
4594
4595    LINE: while (<STDIN>) {
4596        next LINE if /^#/;  # discard comments
4597        #...
4598    }
4599
4600Note that if there were a L<C<continue>|/continue BLOCK> block on the
4601above, it would get
4602executed even on discarded lines.  If LABEL is omitted, the command
4603refers to the innermost enclosing loop.  The C<next EXPR> form, available
4604as of Perl 5.18.0, allows a label name to be computed at run time, being
4605otherwise identical to C<next LABEL>.
4606
4607L<C<next>|/next LABEL> cannot return a value from a block that typically
4608returns a value, such as C<eval {}>, C<sub {}>, or C<do {}>. It will perform
4609its flow control behavior, which precludes any return value. It should not be
4610used to exit a L<C<grep>|/grep BLOCK LIST> or L<C<map>|/map BLOCK LIST>
4611operation.
4612
4613Note that a block by itself is semantically identical to a loop
4614that executes once.  Thus L<C<next>|/next LABEL> will exit such a block
4615early.
4616
4617See also L<C<continue>|/continue BLOCK> for an illustration of how
4618L<C<last>|/last LABEL>, L<C<next>|/next LABEL>, and
4619L<C<redo>|/redo LABEL> work.
4620
4621Unlike most named operators, this has the same precedence as assignment.
4622It is also exempt from the looks-like-a-function rule, so
4623C<next ("foo")."bar"> will cause "bar" to be part of the argument to
4624L<C<next>|/next LABEL>.
4625
4626=item no MODULE VERSION LIST
4627X<no declarations>
4628X<unimporting>
4629
4630=item no MODULE VERSION
4631
4632=item no MODULE LIST
4633
4634=item no MODULE
4635
4636=item no VERSION
4637
4638=for Pod::Functions unimport some module symbols or semantics at compile time
4639
4640See the L<C<use>|/use Module VERSION LIST> function, of which
4641L<C<no>|/no MODULE VERSION LIST> is the opposite.
4642
4643=item oct EXPR
4644X<oct> X<octal> X<hex> X<hexadecimal> X<binary> X<bin>
4645
4646=item oct
4647
4648=for Pod::Functions convert a string to an octal number
4649
4650Interprets EXPR as an octal string and returns the corresponding
4651value.  An octal string consists of octal digits and, as of Perl 5.33.5,
4652an optional C<0o> or C<o> prefix.  Each octal digit may be preceded by
4653a single underscore, which will be ignored.
4654(If EXPR happens to start off with C<0x> or C<x>, interprets it as a
4655hex string.  If EXPR starts off with C<0b> or C<b>, it is interpreted as a
4656binary string.  Leading whitespace is ignored in all three cases.)
4657The following will handle decimal, binary, octal, and hex in standard
4658Perl notation:
4659
4660    $val = oct($val) if $val =~ /^0/;
4661
4662If EXPR is omitted, uses L<C<$_>|perlvar/$_>.   To go the other way
4663(produce a number in octal), use L<C<sprintf>|/sprintf FORMAT, LIST> or
4664L<C<printf>|/printf FILEHANDLE FORMAT, LIST>:
4665
4666    my $dec_perms = (stat("filename"))[2] & 07777;
4667    my $oct_perm_str = sprintf "%o", $perms;
4668
4669The L<C<oct>|/oct EXPR> function is commonly used when a string such as
4670C<644> needs
4671to be converted into a file mode, for example.  Although Perl
4672automatically converts strings into numbers as needed, this automatic
4673conversion assumes base 10.
4674
4675Leading white space is ignored without warning, as too are any trailing
4676non-digits, such as a decimal point (L<C<oct>|/oct EXPR> only handles
4677non-negative integers, not negative integers or floating point).
4678
4679=item open FILEHANDLE,MODE,EXPR
4680X<open> X<pipe> X<file, open> X<fopen>
4681
4682=item open FILEHANDLE,MODE,EXPR,LIST
4683
4684=item open FILEHANDLE,MODE,REFERENCE
4685
4686=item open FILEHANDLE,EXPR
4687
4688=item open FILEHANDLE
4689
4690=for Pod::Functions open a file, pipe, or descriptor
4691
4692Associates an internal FILEHANDLE with the external file specified by
4693EXPR. That filehandle will subsequently allow you to perform
4694I/O operations on that file, such as reading from it or writing to it.
4695
4696Instead of a filename, you may specify an external command
4697(plus an optional argument list) or a scalar reference, in order to open
4698filehandles on commands or in-memory scalars, respectively.
4699
4700A thorough reference to C<open> follows. For a gentler introduction to
4701the basics of C<open>, see also the L<perlopentut> manual page.
4702
4703=over
4704
4705=item Working with files
4706
4707Most often, C<open> gets invoked with three arguments: the required
4708FILEHANDLE (usually an empty scalar variable), followed by MODE (usually
4709a literal describing the I/O mode the filehandle will use), and then the
4710filename  that the new filehandle will refer to.
4711
4712=over
4713
4714=item Simple examples
4715
4716Reading from a file:
4717
4718    open(my $fh, "<", "input.txt")
4719        or die "Can't open < input.txt: $!";
4720
4721    # Process every line in input.txt
4722    while (my $line = readline($fh)) {
4723        #
4724        # ... do something interesting with $line here ...
4725        #
4726    }
4727
4728or writing to one:
4729
4730    open(my $fh, ">", "output.txt")
4731        or die "Can't open > output.txt: $!";
4732
4733    print $fh "This line gets printed into output.txt.\n";
4734
4735For a summary of common filehandle operations such as these, see
4736L<perlintro/Files and I/O>.
4737
4738=item About filehandles
4739
4740The first argument to C<open>, labeled FILEHANDLE in this reference, is
4741usually a scalar variable. (Exceptions exist, described in "Other
4742considerations", below.) If the call to C<open> succeeds, then the
4743expression provided as FILEHANDLE will get assigned an open
4744I<filehandle>. That filehandle provides an internal reference to the
4745specified external file, conveniently stored in a Perl variable, and
4746ready for I/O operations such as reading and writing.
4747
4748=item About modes
4749
4750When calling C<open> with three or more arguments, the second argument
4751-- labeled MODE here -- defines the I<open mode>. MODE is usually a
4752literal string comprising special characters that define the intended
4753I/O role of the filehandle being created: whether it's read-only, or
4754read-and-write, and so on.
4755
4756If MODE is C<< < >>, the file is opened for input (read-only).
4757If MODE is C<< > >>, the file is opened for output, with existing files
4758first being truncated ("clobbered") and nonexisting files newly created.
4759If MODE is C<<< >> >>>, the file is opened for appending, again being
4760created if necessary.
4761
4762You can put a C<+> in front of the C<< > >> or C<< < >> to
4763indicate that you want both read and write access to the file; thus
4764C<< +< >> is almost always preferred for read/write updates--the
4765C<< +> >> mode would clobber the file first.  You can't usually use
4766either read-write mode for updating textfiles, since they have
4767variable-length records.  See the B<-i> switch in
4768L<perlrun|perlrun/-i[extension]> for a better approach.  The file is
4769created with permissions of C<0666> modified by the process's
4770L<C<umask>|/umask EXPR> value.
4771
4772These various prefixes correspond to the L<fopen(3)> modes of C<r>,
4773C<r+>, C<w>, C<w+>, C<a>, and C<a+>.
4774
4775More examples of different modes in action:
4776
4777 # Open a file for concatenation
4778 open(my $log, ">>", "/usr/spool/news/twitlog")
4779     or warn "Couldn't open log file; discarding input";
4780
4781 # Open a file for reading and writing
4782 open(my $dbase, "+<", "dbase.mine")
4783     or die "Can't open 'dbase.mine' for update: $!";
4784
4785=item Checking the return value
4786
4787Open returns nonzero on success, the undefined value otherwise.  If the
4788C<open> involved a pipe, the return value happens to be the pid of the
4789subprocess.
4790
4791When opening a file, it's seldom a good idea to continue if the request
4792failed, so C<open> is frequently used with L<C<die>|/die LIST>. Even if
4793you want your code to do something other than C<die> on a failed open,
4794you should still always check the return value from opening a file.
4795
4796=back
4797
4798=item Specifying I/O layers in MODE
4799
4800You can use the three-argument form of open to specify
4801I/O layers (sometimes referred to as "disciplines") to apply to the new
4802filehandle. These affect how the input and output are processed (see
4803L<open> and
4804L<PerlIO> for more details).  For example:
4805
4806    # loads PerlIO::encoding automatically
4807    open(my $fh, "<:encoding(UTF-8)", $filename)
4808        || die "Can't open UTF-8 encoded $filename: $!";
4809
4810This opens the UTF8-encoded file containing Unicode characters;
4811see L<perluniintro>.  Note that if layers are specified in the
4812three-argument form, then default layers stored in
4813L<C<${^OPEN}>|perlvar/${^OPEN}>
4814(usually set by the L<open> pragma or the switch C<-CioD>) are ignored.
4815Those layers will also be ignored if you specify a colon with no name
4816following it.  In that case the default layer for the operating system
4817(:raw on Unix, :crlf on Windows) is used.
4818
4819On some systems (in general, DOS- and Windows-based systems)
4820L<C<binmode>|/binmode FILEHANDLE, LAYER> is necessary when you're not
4821working with a text file.  For the sake of portability it is a good idea
4822always to use it when appropriate, and never to use it when it isn't
4823appropriate.  Also, people can set their I/O to be by default
4824UTF8-encoded Unicode, not bytes.
4825
4826=item Using C<undef> for temporary files
4827
4828As a special case the three-argument form with a read/write mode and the third
4829argument being L<C<undef>|/undef EXPR>:
4830
4831    open(my $tmp, "+>", undef) or die ...
4832
4833opens a filehandle to a newly created empty anonymous temporary file.
4834(This happens under any mode, which makes C<< +> >> the only useful and
4835sensible mode to use.)  You will need to
4836L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE> to do the reading.
4837
4838
4839=item Opening a filehandle into an in-memory scalar
4840
4841You can open filehandles directly to Perl scalars instead of a file or
4842other resource external to the program. To do so, provide a reference to
4843that scalar as the third argument to C<open>, like so:
4844
4845 open(my $memory, ">", \$var)
4846     or die "Can't open memory file: $!";
4847 print $memory "foo!\n";    # output will appear in $var
4848
4849To (re)open C<STDOUT> or C<STDERR> as an in-memory file, close it first:
4850
4851    close STDOUT;
4852    open(STDOUT, ">", \$variable)
4853	or die "Can't open STDOUT: $!";
4854
4855The scalars for in-memory files are treated as octet strings: unless
4856the file is being opened with truncation the scalar may not contain
4857any code points over 0xFF.
4858
4859Opening in-memory files I<can> fail for a variety of reasons.  As with
4860any other C<open>, check the return value for success.
4861
4862I<Technical note>: This feature works only when Perl is built with
4863PerlIO -- the default, except with older (pre-5.16) Perl installations
4864that were configured to not include it (e.g. via C<Configure
4865-Uuseperlio>). You can see whether your Perl was built with PerlIO by
4866running C<perl -V:useperlio>.  If it says C<'define'>, you have PerlIO;
4867otherwise you don't.
4868
4869See L<perliol> for detailed info on PerlIO.
4870
4871=item Opening a filehandle into a command
4872
4873If MODE is C<|->, then the filename is
4874interpreted as a command to which output is to be piped, and if MODE
4875is C<-|>, the filename is interpreted as a command that pipes
4876output to us.  In the two-argument (and one-argument) form, one should
4877replace dash (C<->) with the command.
4878See L<perlipc/"Using open() for IPC"> for more examples of this.
4879(You are not allowed to L<C<open>|/open FILEHANDLE,MODE,EXPR> to a command
4880that pipes both in I<and> out, but see L<IPC::Open2>, L<IPC::Open3>, and
4881L<perlipc/"Bidirectional Communication with Another Process"> for
4882alternatives.)
4883
4884
4885 open(my $article_fh, "-|", "caesar <$article")  # decrypt
4886                                                 # article
4887     or die "Can't start caesar: $!";
4888
4889 open(my $article_fh, "caesar <$article |")      # ditto
4890     or die "Can't start caesar: $!";
4891
4892 open(my $out_fh, "|-", "sort >Tmp$$")    # $$ is our process id
4893     or die "Can't start sort: $!";
4894
4895
4896In the form of pipe opens taking three or more arguments, if LIST is specified
4897(extra arguments after the command name) then LIST becomes arguments
4898to the command invoked if the platform supports it.  The meaning of
4899L<C<open>|/open FILEHANDLE,MODE,EXPR> with more than three arguments for
4900non-pipe modes is not yet defined, but experimental "layers" may give
4901extra LIST arguments meaning.
4902
4903If you open a pipe on the command C<-> (that is, specify either C<|-> or C<-|>
4904with the one- or two-argument forms of
4905L<C<open>|/open FILEHANDLE,MODE,EXPR>), an implicit L<C<fork>|/fork> is done,
4906so L<C<open>|/open FILEHANDLE,MODE,EXPR> returns twice: in the parent process
4907it returns the pid
4908of the child process, and in the child process it returns (a defined) C<0>.
4909Use C<defined($pid)> or C<//> to determine whether the open was successful.
4910
4911For example, use either
4912
4913   my $child_pid = open(my $from_kid, "-|")
4914        // die "Can't fork: $!";
4915
4916or
4917
4918   my $child_pid = open(my $to_kid,   "|-")
4919        // die "Can't fork: $!";
4920
4921followed by
4922
4923    if ($child_pid) {
4924	# am the parent:
4925	# either write $to_kid or else read $from_kid
4926	...
4927       waitpid $child_pid, 0;
4928    } else {
4929	# am the child; use STDIN/STDOUT normally
4930	...
4931	exit;
4932    }
4933
4934The filehandle behaves normally for the parent, but I/O to that
4935filehandle is piped from/to the STDOUT/STDIN of the child process.
4936In the child process, the filehandle isn't opened--I/O happens from/to
4937the new STDOUT/STDIN.  Typically this is used like the normal
4938piped open when you want to exercise more control over just how the
4939pipe command gets executed, such as when running setuid and
4940you don't want to have to scan shell commands for metacharacters.
4941
4942The following blocks are more or less equivalent:
4943
4944    open(my $fh, "|tr '[a-z]' '[A-Z]'");
4945    open(my $fh, "|-", "tr '[a-z]' '[A-Z]'");
4946    open(my $fh, "|-") || exec 'tr', '[a-z]', '[A-Z]';
4947    open(my $fh, "|-", "tr", '[a-z]', '[A-Z]');
4948
4949    open(my $fh, "cat -n '$file'|");
4950    open(my $fh, "-|", "cat -n '$file'");
4951    open(my $fh, "-|") || exec "cat", "-n", $file;
4952    open(my $fh, "-|", "cat", "-n", $file);
4953
4954The last two examples in each block show the pipe as "list form", which
4955is not yet supported on all platforms. (If your platform has a real
4956L<C<fork>|/fork>, such as Linux and macOS, you can use the list form; it
4957also works on Windows with Perl 5.22 or later.) You would want to use
4958the list form of the pipe so you can pass literal arguments to the
4959command without risk of the shell interpreting any shell metacharacters
4960in them. However, this also bars you from opening pipes to commands that
4961intentionally contain shell metacharacters, such as:
4962
4963    open(my $fh, "|cat -n | expand -4 | lpr")
4964    	|| die "Can't open pipeline to lpr: $!";
4965
4966See L<perlipc/"Safe Pipe Opens"> for more examples of this.
4967
4968=item Duping filehandles
4969
4970You may also, in the Bourne shell tradition, specify an EXPR beginning
4971with C<< >& >>, in which case the rest of the string is interpreted
4972as the name of a filehandle (or file descriptor, if numeric) to be
4973duped (as in L<dup(2)>) and opened.  You may use C<&> after C<< > >>,
4974C<<< >> >>>, C<< < >>, C<< +> >>, C<<< +>> >>>, and C<< +< >>.
4975The mode you specify should match the mode of the original filehandle.
4976(Duping a filehandle does not take into account any existing contents
4977of IO buffers.)  If you use the three-argument
4978form, then you can pass either a
4979number, the name of a filehandle, or the normal "reference to a glob".
4980
4981Here is a script that saves, redirects, and restores C<STDOUT> and
4982C<STDERR> using various methods:
4983
4984    #!/usr/bin/perl
4985    open(my $oldout, ">&STDOUT")
4986        or die "Can't dup STDOUT: $!";
4987    open(OLDERR,     ">&", \*STDERR)
4988        or die "Can't dup STDERR: $!";
4989
4990    open(STDOUT, '>', "foo.out")
4991        or die "Can't redirect STDOUT: $!";
4992    open(STDERR, ">&STDOUT")
4993        or die "Can't dup STDOUT: $!";
4994
4995    select STDERR; $| = 1;  # make unbuffered
4996    select STDOUT; $| = 1;  # make unbuffered
4997
4998    print STDOUT "stdout 1\n";  # this works for
4999    print STDERR "stderr 1\n";  # subprocesses too
5000
5001    open(STDOUT, ">&", $oldout)
5002        or die "Can't dup \$oldout: $!";
5003    open(STDERR, ">&OLDERR")
5004        or die "Can't dup OLDERR: $!";
5005
5006    print STDOUT "stdout 2\n";
5007    print STDERR "stderr 2\n";
5008
5009If you specify C<< '<&=X' >>, where C<X> is a file descriptor number
5010or a filehandle, then Perl will do an equivalent of C's L<fdopen(3)> of
5011that file descriptor (and not call L<dup(2)>); this is more
5012parsimonious of file descriptors.  For example:
5013
5014    # open for input, reusing the fileno of $fd
5015    open(my $fh, "<&=", $fd)
5016
5017or
5018
5019    open(my $fh, "<&=$fd")
5020
5021or
5022
5023    # open for append, using the fileno of $oldfh
5024    open(my $fh, ">>&=", $oldfh)
5025
5026Being parsimonious on filehandles is also useful (besides being
5027parsimonious) for example when something is dependent on file
5028descriptors, like for example locking using
5029L<C<flock>|/flock FILEHANDLE,OPERATION>.  If you do just
5030C<< open(my $A, ">>&", $B) >>, the filehandle C<$A> will not have the
5031same file descriptor as C<$B>, and therefore C<flock($A)> will not
5032C<flock($B)> nor vice versa.  But with C<< open(my $A, ">>&=", $B) >>,
5033the filehandles will share the same underlying system file descriptor.
5034
5035Note that under Perls older than 5.8.0, Perl uses the standard C library's'
5036L<fdopen(3)> to implement the C<=> functionality.  On many Unix systems,
5037L<fdopen(3)> fails when file descriptors exceed a certain value, typically 255.
5038For Perls 5.8.0 and later, PerlIO is (most often) the default.
5039
5040=item Legacy usage
5041
5042This section describes ways to call C<open> outside of best practices;
5043you may encounter these uses in older code. Perl does not consider their
5044use deprecated, exactly, but neither is it recommended in new code, for
5045the sake of clarity and readability.
5046
5047=over
5048
5049=item Specifying mode and filename as a single argument
5050
5051In the one- and two-argument forms of the call, the mode and filename
5052should be concatenated (in that order), preferably separated by white
5053space.  You can--but shouldn't--omit the mode in these forms when that mode
5054is C<< < >>.  It is safe to use the two-argument form of
5055L<C<open>|/open FILEHANDLE,MODE,EXPR> if the filename argument is a known literal.
5056
5057 open(my $dbase, "+<dbase.mine")          # ditto
5058     or die "Can't open 'dbase.mine' for update: $!";
5059
5060In the two-argument (and one-argument) form, opening C<< <- >>
5061or C<-> opens STDIN and opening C<< >- >> opens STDOUT.
5062
5063New code should favor the three-argument form of C<open> over this older
5064form. Declaring the mode and the filename as two distinct arguments
5065avoids any confusion between the two.
5066
5067=item Assigning a filehandle to a bareword
5068
5069An older style is to use a bareword as the filehandle, as
5070
5071    open(FH, "<", "input.txt")
5072       or die "Can't open < input.txt: $!";
5073
5074Then you can use C<FH> as the filehandle, in C<< close FH >> and C<<
5075<FH> >> and so on.  Note that it's a global variable, so this form is
5076not recommended when dealing with filehandles other than Perl's built-in ones
5077(e.g. STDOUT and STDIN).  In fact, using a bareword for the filehandle is
5078an error when the
5079L<C<"bareword_filehandles"> feature|feature/"The 'bareword_filehandles' feature">
5080has been disabled.  This feature is disabled automatically when in the
5081scope of C<use v5.36.0> or later.
5082
5083=item Calling C<open> with one argument via global variables
5084
5085As a shortcut, a one-argument call takes the filename from the global
5086scalar variable of the same name as the bareword filehandle:
5087
5088    $ARTICLE = 100;
5089    open(ARTICLE)
5090        or die "Can't find article $ARTICLE: $!\n";
5091
5092Here C<$ARTICLE> must be a global scalar variable in the same package
5093as the filehandle - not one declared with L<C<my>|/my VARLIST> or
5094L<C<state>|/state VARLIST>.
5095
5096=back
5097
5098=item Other considerations
5099
5100=over
5101
5102=item Automatic filehandle closure
5103
5104The filehandle will be closed when its reference count reaches zero. If
5105it is a lexically scoped variable declared with L<C<my>|/my VARLIST>,
5106that usually means the end of the enclosing scope.  However, this
5107automatic close does not check for errors, so it is better to explicitly
5108close filehandles, especially those used for writing:
5109
5110    close($handle)
5111       || warn "close failed: $!";
5112
5113=item Automatic pipe flushing
5114
5115Perl will attempt to flush all files opened for
5116output before any operation that may do a fork, but this may not be
5117supported on some platforms (see L<perlport>).  To be safe, you may need
5118to set L<C<$E<verbar>>|perlvar/$E<verbar>> (C<$AUTOFLUSH> in L<English>)
5119or call the C<autoflush> method of L<C<IO::Handle>|IO::Handle/METHODS>
5120on any open handles.
5121
5122On systems that support a close-on-exec flag on files, the flag will
5123be set for the newly opened file descriptor as determined by the value
5124of L<C<$^F>|perlvar/$^F>.  See L<perlvar/$^F>.
5125
5126Closing any piped filehandle causes the parent process to wait for the
5127child to finish, then returns the status value in L<C<$?>|perlvar/$?> and
5128L<C<${^CHILD_ERROR_NATIVE}>|perlvar/${^CHILD_ERROR_NATIVE}>.
5129
5130=item Direct versus by-reference assignment of filehandles
5131
5132If FILEHANDLE -- the first argument in a call to C<open> -- is an
5133undefined scalar variable (or array or hash element), a new filehandle
5134is autovivified, meaning that the variable is assigned a reference to a
5135newly allocated anonymous filehandle.  Otherwise if FILEHANDLE is an
5136expression, its value is the real filehandle.  (This is considered a
5137symbolic reference, so C<use strict "refs"> should I<not> be in effect.)
5138
5139=item Whitespace and special characters in the filename argument
5140
5141The filename passed to the one- and two-argument forms of
5142L<C<open>|/open FILEHANDLE,MODE,EXPR> will
5143have leading and trailing whitespace deleted and normal
5144redirection characters honored.  This property, known as "magic open",
5145can often be used to good effect.  A user could specify a filename of
5146F<"rsh cat file |">, or you could change certain filenames as needed:
5147
5148    $filename =~ s/(.*\.gz)\s*$/gzip -dc < $1|/;
5149    open(my $fh, $filename)
5150        or die "Can't open $filename: $!";
5151
5152Use the three-argument form to open a file with arbitrary weird characters in it,
5153
5154    open(my $fh, "<", $file)
5155    	|| die "Can't open $file: $!";
5156
5157otherwise it's necessary to protect any leading and trailing whitespace:
5158
5159    $file =~ s#^(\s)#./$1#;
5160    open(my $fh, "< $file\0")
5161    	|| die "Can't open $file: $!";
5162
5163(this may not work on some bizarre filesystems).  One should
5164conscientiously choose between the I<magic> and I<three-argument> form
5165of L<C<open>|/open FILEHANDLE,MODE,EXPR>:
5166
5167    open(my $in, $ARGV[0]) || die "Can't open $ARGV[0]: $!";
5168
5169will allow the user to specify an argument of the form C<"rsh cat file |">,
5170but will not work on a filename that happens to have a trailing space, while
5171
5172    open(my $in, "<", $ARGV[0])
5173    	|| die "Can't open $ARGV[0]: $!";
5174
5175will have exactly the opposite restrictions. (However, some shells
5176support the syntax C<< perl your_program.pl <( rsh cat file ) >>, which
5177produces a filename that can be opened normally.)
5178
5179=item Invoking C-style C<open>
5180
5181If you want a "real" C L<open(2)>, then you should use the
5182L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE> function, which involves
5183no such magic (but uses different filemodes than Perl
5184L<C<open>|/open FILEHANDLE,MODE,EXPR>, which corresponds to C L<fopen(3)>).
5185This is another way to protect your filenames from interpretation.  For
5186example:
5187
5188    use IO::Handle;
5189    sysopen(my $fh, $path, O_RDWR|O_CREAT|O_EXCL)
5190        or die "Can't open $path: $!";
5191    $fh->autoflush(1);
5192    print $fh "stuff $$\n";
5193    seek($fh, 0, 0);
5194    print "File contains: ", readline($fh);
5195
5196See L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE> for some details about
5197mixing reading and writing.
5198
5199=item Portability issues
5200
5201See L<perlport/open>.
5202
5203=back
5204
5205=back
5206
5207
5208=item opendir DIRHANDLE,EXPR
5209X<opendir>
5210
5211=for Pod::Functions open a directory
5212
5213Opens a directory named EXPR for processing by
5214L<C<readdir>|/readdir DIRHANDLE>, L<C<telldir>|/telldir DIRHANDLE>,
5215L<C<seekdir>|/seekdir DIRHANDLE,POS>,
5216L<C<rewinddir>|/rewinddir DIRHANDLE>, and
5217L<C<closedir>|/closedir DIRHANDLE>.  Returns true if successful.
5218DIRHANDLE may be an expression whose value can be used as an indirect
5219dirhandle, usually the real dirhandle name.  If DIRHANDLE is an undefined
5220scalar variable (or array or hash element), the variable is assigned a
5221reference to a new anonymous dirhandle; that is, it's autovivified.
5222Dirhandles are the same objects as filehandles; an I/O object can only
5223be open as one of these handle types at once.
5224
5225See the example at L<C<readdir>|/readdir DIRHANDLE>.
5226
5227=item ord EXPR
5228X<ord> X<encoding>
5229
5230=item ord
5231
5232=for Pod::Functions find a character's code point
5233
5234Returns the code point of the first character of EXPR.
5235If EXPR is an empty string, returns 0.  If EXPR is omitted, uses
5236L<C<$_>|perlvar/$_>.
5237(Note I<character>, not byte.)
5238
5239For the reverse, see L<C<chr>|/chr NUMBER>.
5240See L<perlunicode> for more about Unicode.
5241
5242=item our VARLIST
5243X<our> X<global>
5244
5245=item our TYPE VARLIST
5246
5247=item our VARLIST : ATTRS
5248
5249=item our TYPE VARLIST : ATTRS
5250
5251=for Pod::Functions +5.6.0 declare and assign a package variable (lexical scoping)
5252
5253L<C<our>|/our VARLIST> makes a lexical alias to a package (i.e. global)
5254variable of the same name in the current package for use within the
5255current lexical scope.
5256
5257L<C<our>|/our VARLIST> has the same scoping rules as
5258L<C<my>|/my VARLIST> or L<C<state>|/state VARLIST>, meaning that it is
5259only valid within a lexical scope.  Unlike L<C<my>|/my VARLIST> and
5260L<C<state>|/state VARLIST>, which both declare new (lexical) variables,
5261L<C<our>|/our VARLIST> only creates an alias to an existing variable: a
5262package variable of the same name.
5263
5264This means that when C<use strict 'vars'> is in effect, L<C<our>|/our
5265VARLIST> lets you use a package variable without qualifying it with the
5266package name, but only within the lexical scope of the
5267L<C<our>|/our VARLIST> declaration.  This applies immediately--even
5268within the same statement.
5269
5270    package Foo;
5271    use v5.36;  # which implies "use strict;"
5272
5273    $Foo::foo = 23;
5274
5275    {
5276        our $foo;   # alias to $Foo::foo
5277        print $foo; # prints 23
5278    }
5279
5280    print $Foo::foo; # prints 23
5281
5282    print $foo; # ERROR: requires explicit package name
5283
5284This works even if the package variable has not been used before, as
5285package variables spring into existence when first used.
5286
5287    package Foo;
5288    use v5.36;
5289
5290    our $foo = 23;   # just like $Foo::foo = 23
5291
5292    print $Foo::foo; # prints 23
5293
5294Because the variable becomes legal immediately under C<use strict 'vars'>, so
5295long as there is no variable with that name is already in scope, you can then
5296reference the package variable again even within the same statement.
5297
5298    package Foo;
5299    use v5.36;
5300
5301    my  $foo = $foo; # error, undeclared $foo on right-hand side
5302    our $foo = $foo; # no errors
5303
5304If more than one variable is listed, the list must be placed
5305in parentheses.
5306
5307    our($bar, $baz);
5308
5309Like L<C<my>|/my VARLIST>, L<C<state>|/state VARLIST>, and
5310L<C<local>|/local EXPR>, L<C<our>|/our VARLIST> can operate on a variable
5311anywhere it appears in an expression (aside from interpolation in strings).
5312The declaration will not apply to additional uses of the same variable until
5313the next statement. This means additional uses of that variable within the
5314same statement will act as they would have before that declaration occurred,
5315with the exception that it will still satisfy strict 'vars' and interpret that
5316variable as the newly aliased package variable if it was not yet declared in
5317that scope.
5318
5319    package main;
5320    my $x = 2;
5321    foo($x, our $x = $x + 1, $x); # foo() receives (2, 3, 2)
5322    foo($x, our $z = 5, $z);      # foo() receives (3, 5, 5)
5323
5324An L<C<our>|/our VARLIST> declaration declares an alias for a package
5325variable that will be visible
5326across its entire lexical scope, even across package boundaries.  The
5327package in which the variable is entered is determined at the point
5328of the declaration, not at the point of use.  This means the following
5329behavior holds:
5330
5331    package Foo;
5332    our $bar;      # declares $Foo::bar for rest of lexical scope
5333    $bar = 20;
5334
5335    package Bar;
5336    print $bar;    # prints 20, as it refers to $Foo::bar
5337
5338Multiple L<C<our>|/our VARLIST> declarations with the same name in the
5339same lexical
5340scope are allowed if they are in different packages.  If they happen
5341to be in the same package, Perl will emit warnings if you have asked
5342for them, just like multiple L<C<my>|/my VARLIST> declarations.  Unlike
5343a second L<C<my>|/my VARLIST> declaration, which will bind the name to a
5344fresh variable, a second L<C<our>|/our VARLIST> declaration in the same
5345package, in the same scope, is merely redundant.
5346
5347    use warnings;
5348    package Foo;
5349    our $bar;      # declares $Foo::bar for rest of lexical scope
5350    $bar = 20;
5351
5352    package Bar;
5353    our $bar = 30; # declares $Bar::bar for rest of lexical scope
5354    print $bar;    # prints 30
5355
5356    our $bar;      # emits warning but has no other effect
5357    print $bar;    # still prints 30
5358
5359An L<C<our>|/our VARLIST> declaration may also have a list of attributes
5360associated with it.
5361
5362The exact semantics and interface of TYPE and ATTRS are still
5363evolving.  TYPE is currently bound to the use of the L<fields> pragma,
5364and attributes are handled using the L<attributes> pragma, or, starting
5365from Perl 5.8.0, also via the L<Attribute::Handlers> module.  See
5366L<perlsub/"Private Variables via my()"> for details.
5367
5368Note that with a parenthesised list, L<C<undef>|/undef EXPR> can be used
5369as a dummy placeholder, for example to skip assignment of initial
5370values:
5371
5372    our ( undef, $min, $hour ) = localtime;
5373
5374L<C<our>|/our VARLIST> differs from L<C<use vars>|vars>, which allows
5375use of an unqualified name I<only> within the affected package, but
5376across scopes.
5377
5378=item pack TEMPLATE,LIST
5379X<pack>
5380
5381=for Pod::Functions convert a list into a binary representation
5382
5383Takes a LIST of values and converts it into a string using the rules
5384given by the TEMPLATE.  The resulting string is the concatenation of
5385the converted values.  Typically, each converted value looks
5386like its machine-level representation.  For example, on 32-bit machines
5387an integer may be represented by a sequence of 4 bytes, which  will in
5388Perl be presented as a string that's 4 characters long.
5389
5390See L<perlpacktut> for an introduction to this function.
5391
5392The TEMPLATE is a sequence of characters that give the order and type
5393of values, as follows:
5394
5395    a  A string with arbitrary binary data, will be null padded.
5396    A  A text (ASCII) string, will be space padded.
5397    Z  A null-terminated (ASCIZ) string, will be null padded.
5398
5399    b  A bit string (ascending bit order inside each byte,
5400       like vec()).
5401    B  A bit string (descending bit order inside each byte).
5402    h  A hex string (low nybble first).
5403    H  A hex string (high nybble first).
5404
5405    c  A signed char (8-bit) value.
5406    C  An unsigned char (octet) value.
5407    W  An unsigned char value (can be greater than 255).
5408
5409    s  A signed short (16-bit) value.
5410    S  An unsigned short value.
5411
5412    l  A signed long (32-bit) value.
5413    L  An unsigned long value.
5414
5415    q  A signed quad (64-bit) value.
5416    Q  An unsigned quad value.
5417         (Quads are available only if your system supports 64-bit
5418          integer values _and_ if Perl has been compiled to support
5419          those.  Raises an exception otherwise.)
5420
5421    i  A signed integer value.
5422    I  An unsigned integer value.
5423         (This 'integer' is _at_least_ 32 bits wide.  Its exact
5424          size depends on what a local C compiler calls 'int'.)
5425
5426    n  An unsigned short (16-bit) in "network" (big-endian) order.
5427    N  An unsigned long (32-bit) in "network" (big-endian) order.
5428    v  An unsigned short (16-bit) in "VAX" (little-endian) order.
5429    V  An unsigned long (32-bit) in "VAX" (little-endian) order.
5430
5431    j  A Perl internal signed integer value (IV).
5432    J  A Perl internal unsigned integer value (UV).
5433
5434    f  A single-precision float in native format.
5435    d  A double-precision float in native format.
5436
5437    F  A Perl internal floating-point value (NV) in native format
5438    D  A float of long-double precision in native format.
5439         (Long doubles are available only if your system supports
5440          long double values. Raises an exception otherwise.
5441          Note that there are different long double formats.)
5442
5443    p  A pointer to a null-terminated string.
5444    P  A pointer to a structure (fixed-length string).
5445
5446    u  A uuencoded string.
5447    U  A Unicode character number.  Encodes to a character in char-
5448       acter mode and UTF-8 (or UTF-EBCDIC in EBCDIC platforms) in
5449       byte mode.  Also on EBCDIC platforms, the character number will
5450       be the native EBCDIC value for character numbers below 256.
5451       This allows most programs using this feature to not have to
5452       care which type of platform they are running on.
5453
5454    w  A BER compressed integer (not an ASN.1 BER, see perlpacktut
5455       for details).  Its bytes represent an unsigned integer in
5456       base 128, most significant digit first, with as few digits
5457       as possible.  Bit eight (the high bit) is set on each byte
5458       except the last.
5459
5460    x  A null byte (a.k.a ASCII NUL, "\000", chr(0))
5461    X  Back up a byte.
5462    @  Null-fill or truncate to absolute position, counted from the
5463       start of the innermost ()-group.
5464    .  Null-fill or truncate to absolute position specified by
5465       the value.
5466    (  Start of a ()-group.
5467
5468One or more modifiers below may optionally follow certain letters in the
5469TEMPLATE (the second column lists letters for which the modifier is valid):
5470
5471    !   sSlLiI     Forces native (short, long, int) sizes instead
5472                   of fixed (16-/32-bit) sizes.
5473
5474    !   xX         Make x and X act as alignment commands.
5475
5476    !   nNvV       Treat integers as signed instead of unsigned.
5477
5478    !   @.         Specify position as byte offset in the internal
5479                   representation of the packed string.  Efficient
5480                   but dangerous.
5481
5482    >   sSiIlLqQ   Force big-endian byte-order on the type.
5483        jJfFdDpP   (The "big end" touches the construct.)
5484
5485    <   sSiIlLqQ   Force little-endian byte-order on the type.
5486        jJfFdDpP   (The "little end" touches the construct.)
5487
5488The C<< > >> and C<< < >> modifiers can also be used on C<()> groups
5489to force a particular byte-order on all components in that group,
5490including all its subgroups.
5491
5492=begin comment
5493
5494Larry recalls that the hex and bit string formats (H, h, B, b) were added to
5495pack for processing data from NASA's Magellan probe.  Magellan was in an
5496elliptical orbit, using the antenna for the radar mapping when close to
5497Venus and for communicating data back to Earth for the rest of the orbit.
5498There were two transmission units, but one of these failed, and then the
5499other developed a fault whereby it would randomly flip the sense of all the
5500bits. It was easy to automatically detect complete records with the correct
5501sense, and complete records with all the bits flipped. However, this didn't
5502recover the records where the sense flipped midway. A colleague of Larry's
5503was able to pretty much eyeball where the records flipped, so they wrote an
5504editor named kybble (a pun on the dog food Kibbles 'n Bits) to enable him to
5505manually correct the records and recover the data. For this purpose pack
5506gained the hex and bit string format specifiers.
5507
5508git shows that they were added to perl 3.0 in patch #44 (Jan 1991, commit
550927e2fb84680b9cc1), but the patch description makes no mention of their
5510addition, let alone the story behind them.
5511
5512=end comment
5513
5514The following rules apply:
5515
5516=over
5517
5518=item *
5519
5520Each letter may optionally be followed by a number indicating the repeat
5521count.  A numeric repeat count may optionally be enclosed in brackets, as
5522in C<pack("C[80]", @arr)>.  The repeat count gobbles that many values from
5523the LIST when used with all format types other than C<a>, C<A>, C<Z>, C<b>,
5524C<B>, C<h>, C<H>, C<@>, C<.>, C<x>, C<X>, and C<P>, where it means
5525something else, described below.  Supplying a C<*> for the repeat count
5526instead of a number means to use however many items are left, except for:
5527
5528=over
5529
5530=item *
5531
5532C<@>, C<x>, and C<X>, where it is equivalent to C<0>.
5533
5534=item *
5535
5536<.>, where it means relative to the start of the string.
5537
5538=item *
5539
5540C<u>, where it is equivalent to 1 (or 45, which here is equivalent).
5541
5542=back
5543
5544One can replace a numeric repeat count with a template letter enclosed in
5545brackets to use the packed byte length of the bracketed template for the
5546repeat count.
5547
5548For example, the template C<x[L]> skips as many bytes as in a packed long,
5549and the template C<"$t X[$t] $t"> unpacks twice whatever $t (when
5550variable-expanded) unpacks.  If the template in brackets contains alignment
5551commands (such as C<x![d]>), its packed length is calculated as if the
5552start of the template had the maximal possible alignment.
5553
5554When used with C<Z>, a C<*> as the repeat count is guaranteed to add a
5555trailing null byte, so the resulting string is always one byte longer than
5556the byte length of the item itself.
5557
5558When used with C<@>, the repeat count represents an offset from the start
5559of the innermost C<()> group.
5560
5561When used with C<.>, the repeat count determines the starting position to
5562calculate the value offset as follows:
5563
5564=over
5565
5566=item *
5567
5568If the repeat count is C<0>, it's relative to the current position.
5569
5570=item *
5571
5572If the repeat count is C<*>, the offset is relative to the start of the
5573packed string.
5574
5575=item *
5576
5577And if it's an integer I<n>, the offset is relative to the start of the
5578I<n>th innermost C<( )> group, or to the start of the string if I<n> is
5579bigger then the group level.
5580
5581=back
5582
5583The repeat count for C<u> is interpreted as the maximal number of bytes
5584to encode per line of output, with 0, 1 and 2 replaced by 45.  The repeat
5585count should not be more than 65.
5586
5587=item *
5588
5589The C<a>, C<A>, and C<Z> types gobble just one value, but pack it as a
5590string of length count, padding with nulls or spaces as needed.  When
5591unpacking, C<A> strips trailing whitespace and nulls, C<Z> strips everything
5592after the first null, and C<a> returns data with no stripping at all.
5593
5594If the value to pack is too long, the result is truncated.  If it's too
5595long and an explicit count is provided, C<Z> packs only C<$count-1> bytes,
5596followed by a null byte.  Thus C<Z> always packs a trailing null, except
5597when the count is 0.
5598
5599=item *
5600
5601Likewise, the C<b> and C<B> formats pack a string that's that many bits long.
5602Each such format generates 1 bit of the result.  These are typically followed
5603by a repeat count like C<B8> or C<B64>.
5604
5605Each result bit is based on the least-significant bit of the corresponding
5606input character, i.e., on C<ord($char)%2>.  In particular, characters C<"0">
5607and C<"1"> generate bits 0 and 1, as do characters C<"\000"> and C<"\001">.
5608
5609Starting from the beginning of the input string, each 8-tuple
5610of characters is converted to 1 character of output.  With format C<b>,
5611the first character of the 8-tuple determines the least-significant bit of a
5612character; with format C<B>, it determines the most-significant bit of
5613a character.
5614
5615If the length of the input string is not evenly divisible by 8, the
5616remainder is packed as if the input string were padded by null characters
5617at the end.  Similarly during unpacking, "extra" bits are ignored.
5618
5619If the input string is longer than needed, remaining characters are ignored.
5620
5621A C<*> for the repeat count uses all characters of the input field.
5622On unpacking, bits are converted to a string of C<0>s and C<1>s.
5623
5624=item *
5625
5626The C<h> and C<H> formats pack a string that many nybbles (4-bit groups,
5627representable as hexadecimal digits, C<"0".."9"> C<"a".."f">) long.
5628
5629For each such format, L<C<pack>|/pack TEMPLATE,LIST> generates 4 bits of result.
5630With non-alphabetical characters, the result is based on the 4 least-significant
5631bits of the input character, i.e., on C<ord($char)%16>.  In particular,
5632characters C<"0"> and C<"1"> generate nybbles 0 and 1, as do bytes
5633C<"\000"> and C<"\001">.  For characters C<"a".."f"> and C<"A".."F">, the result
5634is compatible with the usual hexadecimal digits, so that C<"a"> and
5635C<"A"> both generate the nybble C<0xA==10>.  Use only these specific hex
5636characters with this format.
5637
5638Starting from the beginning of the template to
5639L<C<pack>|/pack TEMPLATE,LIST>, each pair
5640of characters is converted to 1 character of output.  With format C<h>, the
5641first character of the pair determines the least-significant nybble of the
5642output character; with format C<H>, it determines the most-significant
5643nybble.
5644
5645If the length of the input string is not even, it behaves as if padded by
5646a null character at the end.  Similarly, "extra" nybbles are ignored during
5647unpacking.
5648
5649If the input string is longer than needed, extra characters are ignored.
5650
5651A C<*> for the repeat count uses all characters of the input field.  For
5652L<C<unpack>|/unpack TEMPLATE,EXPR>, nybbles are converted to a string of
5653hexadecimal digits.
5654
5655=item *
5656
5657The C<p> format packs a pointer to a null-terminated string.  You are
5658responsible for ensuring that the string is not a temporary value, as that
5659could potentially get deallocated before you got around to using the packed
5660result.  The C<P> format packs a pointer to a structure of the size indicated
5661by the length.  A null pointer is created if the corresponding value for
5662C<p> or C<P> is L<C<undef>|/undef EXPR>; similarly with
5663L<C<unpack>|/unpack TEMPLATE,EXPR>, where a null pointer unpacks into
5664L<C<undef>|/undef EXPR>.
5665
5666If your system has a strange pointer size--meaning a pointer is neither as
5667big as an int nor as big as a long--it may not be possible to pack or
5668unpack pointers in big- or little-endian byte order.  Attempting to do
5669so raises an exception.
5670
5671=item *
5672
5673The C</> template character allows packing and unpacking of a sequence of
5674items where the packed structure contains a packed item count followed by
5675the packed items themselves.  This is useful when the structure you're
5676unpacking has encoded the sizes or repeat counts for some of its fields
5677within the structure itself as separate fields.
5678
5679For L<C<pack>|/pack TEMPLATE,LIST>, you write
5680I<length-item>C</>I<sequence-item>, and the
5681I<length-item> describes how the length value is packed.  Formats likely
5682to be of most use are integer-packing ones like C<n> for Java strings,
5683C<w> for ASN.1 or SNMP, and C<N> for Sun XDR.
5684
5685For L<C<pack>|/pack TEMPLATE,LIST>, I<sequence-item> may have a repeat
5686count, in which case
5687the minimum of that and the number of available items is used as the argument
5688for I<length-item>.  If it has no repeat count or uses a '*', the number
5689of available items is used.
5690
5691For L<C<unpack>|/unpack TEMPLATE,EXPR>, an internal stack of integer
5692arguments unpacked so far is
5693used.  You write C</>I<sequence-item> and the repeat count is obtained by
5694popping off the last element from the stack.  The I<sequence-item> must not
5695have a repeat count.
5696
5697If I<sequence-item> refers to a string type (C<"A">, C<"a">, or C<"Z">),
5698the I<length-item> is the string length, not the number of strings.  With
5699an explicit repeat count for pack, the packed string is adjusted to that
5700length.  For example:
5701
5702 This code:                             gives this result:
5703
5704 unpack("W/a", "\004Gurusamy")          ("Guru")
5705 unpack("a3/A A*", "007 Bond  J ")      (" Bond", "J")
5706 unpack("a3 x2 /A A*", "007: Bond, J.") ("Bond, J", ".")
5707
5708 pack("n/a* w/a","hello,","world")     "\000\006hello,\005world"
5709 pack("a/W2", ord("a") .. ord("z"))    "2ab"
5710
5711The I<length-item> is not returned explicitly from
5712L<C<unpack>|/unpack TEMPLATE,EXPR>.
5713
5714Supplying a count to the I<length-item> format letter is only useful with
5715C<A>, C<a>, or C<Z>.  Packing with a I<length-item> of C<a> or C<Z> may
5716introduce C<"\000"> characters, which Perl does not regard as legal in
5717numeric strings.
5718
5719=item *
5720
5721The integer types C<s>, C<S>, C<l>, and C<L> may be
5722followed by a C<!> modifier to specify native shorts or
5723longs.  As shown in the example above, a bare C<l> means
5724exactly 32 bits, although the native C<long> as seen by the local C compiler
5725may be larger.  This is mainly an issue on 64-bit platforms.  You can
5726see whether using C<!> makes any difference this way:
5727
5728    printf "format s is %d, s! is %d\n",
5729	length pack("s"), length pack("s!");
5730
5731    printf "format l is %d, l! is %d\n",
5732	length pack("l"), length pack("l!");
5733
5734
5735C<i!> and C<I!> are also allowed, but only for completeness' sake:
5736they are identical to C<i> and C<I>.
5737
5738The actual sizes (in bytes) of native shorts, ints, longs, and long
5739longs on the platform where Perl was built are also available from
5740the command line:
5741
5742    $ perl -V:{short,int,long{,long}}size
5743    shortsize='2';
5744    intsize='4';
5745    longsize='4';
5746    longlongsize='8';
5747
5748or programmatically via the L<C<Config>|Config> module:
5749
5750       use Config;
5751       print $Config{shortsize},    "\n";
5752       print $Config{intsize},      "\n";
5753       print $Config{longsize},     "\n";
5754       print $Config{longlongsize}, "\n";
5755
5756C<$Config{longlongsize}> is undefined on systems without
5757long long support.
5758
5759=item *
5760
5761The integer formats C<s>, C<S>, C<i>, C<I>, C<l>, C<L>, C<j>, and C<J> are
5762inherently non-portable between processors and operating systems because
5763they obey native byteorder and endianness.  For example, a 4-byte integer
57640x12345678 (305419896 decimal) would be ordered natively (arranged in and
5765handled by the CPU registers) into bytes as
5766
5767    0x12 0x34 0x56 0x78  # big-endian
5768    0x78 0x56 0x34 0x12  # little-endian
5769
5770Basically, Intel and VAX CPUs are little-endian, while everybody else,
5771including Motorola m68k/88k, PPC, Sparc, HP PA, Power, and Cray, are
5772big-endian.  Alpha and MIPS can be either: Digital/Compaq uses (well, used)
5773them in little-endian mode, but SGI/Cray uses them in big-endian mode.
5774
5775The names I<big-endian> and I<little-endian> are comic references to the
5776egg-eating habits of the little-endian Lilliputians and the big-endian
5777Blefuscudians from the classic Jonathan Swift satire, I<Gulliver's Travels>.
5778This entered computer lingo via the paper "On Holy Wars and a Plea for
5779Peace" by Danny Cohen, USC/ISI IEN 137, April 1, 1980.
5780
5781Some systems may have even weirder byte orders such as
5782
5783   0x56 0x78 0x12 0x34
5784   0x34 0x12 0x78 0x56
5785
5786These are called mid-endian, middle-endian, mixed-endian, or just weird.
5787
5788You can determine your system endianness with this incantation:
5789
5790   printf("%#02x ", $_) for unpack("W*", pack L=>0x12345678);
5791
5792The byteorder on the platform where Perl was built is also available
5793via L<Config>:
5794
5795    use Config;
5796    print "$Config{byteorder}\n";
5797
5798or from the command line:
5799
5800    $ perl -V:byteorder
5801
5802Byteorders C<"1234"> and C<"12345678"> are little-endian; C<"4321">
5803and C<"87654321"> are big-endian.  Systems with multiarchitecture binaries
5804will have C<"ffff">, signifying that static information doesn't work,
5805one must use runtime probing.
5806
5807For portably packed integers, either use the formats C<n>, C<N>, C<v>,
5808and C<V> or else use the C<< > >> and C<< < >> modifiers described
5809immediately below.  See also L<perlport>.
5810
5811=item *
5812
5813Also floating point numbers have endianness.  Usually (but not always)
5814this agrees with the integer endianness.  Even though most platforms
5815these days use the IEEE 754 binary format, there are differences,
5816especially if the long doubles are involved.  You can see the
5817C<Config> variables C<doublekind> and C<longdblkind> (also C<doublesize>,
5818C<longdblsize>): the "kind" values are enums, unlike C<byteorder>.
5819
5820Portability-wise the best option is probably to keep to the IEEE 754
582164-bit doubles, and of agreed-upon endianness.  Another possibility
5822is the C<"%a">) format of L<C<printf>|/printf FILEHANDLE FORMAT, LIST>.
5823
5824=item *
5825
5826Starting with Perl 5.10.0, integer and floating-point formats, along with
5827the C<p> and C<P> formats and C<()> groups, may all be followed by the
5828C<< > >> or C<< < >> endianness modifiers to respectively enforce big-
5829or little-endian byte-order.  These modifiers are especially useful
5830given how C<n>, C<N>, C<v>, and C<V> don't cover signed integers,
583164-bit integers, or floating-point values.
5832
5833Here are some concerns to keep in mind when using an endianness modifier:
5834
5835=over
5836
5837=item *
5838
5839Exchanging signed integers between different platforms works only
5840when all platforms store them in the same format.  Most platforms store
5841signed integers in two's-complement notation, so usually this is not an issue.
5842
5843=item *
5844
5845The C<< > >> or C<< < >> modifiers can only be used on floating-point
5846formats on big- or little-endian machines.  Otherwise, attempting to
5847use them raises an exception.
5848
5849=item *
5850
5851Forcing big- or little-endian byte-order on floating-point values for
5852data exchange can work only if all platforms use the same
5853binary representation such as IEEE floating-point.  Even if all
5854platforms are using IEEE, there may still be subtle differences.  Being able
5855to use C<< > >> or C<< < >> on floating-point values can be useful,
5856but also dangerous if you don't know exactly what you're doing.
5857It is not a general way to portably store floating-point values.
5858
5859=item *
5860
5861When using C<< > >> or C<< < >> on a C<()> group, this affects
5862all types inside the group that accept byte-order modifiers,
5863including all subgroups.  It is silently ignored for all other
5864types.  You are not allowed to override the byte-order within a group
5865that already has a byte-order modifier suffix.
5866
5867=back
5868
5869=item *
5870
5871Real numbers (floats and doubles) are in native machine format only.
5872Due to the multiplicity of floating-point formats and the lack of a
5873standard "network" representation for them, no facility for interchange has been
5874made.  This means that packed floating-point data written on one machine
5875may not be readable on another, even if both use IEEE floating-point
5876arithmetic (because the endianness of the memory representation is not part
5877of the IEEE spec).  See also L<perlport>.
5878
5879If you know I<exactly> what you're doing, you can use the C<< > >> or C<< < >>
5880modifiers to force big- or little-endian byte-order on floating-point values.
5881
5882Because Perl uses doubles (or long doubles, if configured) internally for
5883all numeric calculation, converting from double into float and thence
5884to double again loses precision, so C<unpack("f", pack("f", $foo)>)
5885will not in general equal $foo.
5886
5887=item *
5888
5889Pack and unpack can operate in two modes: character mode (C<C0> mode) where
5890the packed string is processed per character, and UTF-8 byte mode (C<U0> mode)
5891where the packed string is processed in its UTF-8-encoded Unicode form on
5892a byte-by-byte basis.  Character mode is the default
5893unless the format string starts with C<U>.  You
5894can always switch mode mid-format with an explicit
5895C<C0> or C<U0> in the format.  This mode remains in effect until the next
5896mode change, or until the end of the C<()> group it (directly) applies to.
5897
5898Using C<C0> to get Unicode characters while using C<U0> to get I<non>-Unicode
5899bytes is not necessarily obvious.   Probably only the first of these
5900is what you want:
5901
5902    $ perl -CS -E 'say "\x{3B1}\x{3C9}"' |
5903      perl -CS -ne 'printf "%v04X\n", $_ for unpack("C0A*", $_)'
5904    03B1.03C9
5905    $ perl -CS -E 'say "\x{3B1}\x{3C9}"' |
5906      perl -CS -ne 'printf "%v02X\n", $_ for unpack("U0A*", $_)'
5907    CE.B1.CF.89
5908    $ perl -CS -E 'say "\x{3B1}\x{3C9}"' |
5909      perl -C0 -ne 'printf "%v02X\n", $_ for unpack("C0A*", $_)'
5910    CE.B1.CF.89
5911    $ perl -CS -E 'say "\x{3B1}\x{3C9}"' |
5912      perl -C0 -ne 'printf "%v02X\n", $_ for unpack("U0A*", $_)'
5913    C3.8E.C2.B1.C3.8F.C2.89
5914
5915Those examples also illustrate that you should not try to use
5916L<C<pack>|/pack TEMPLATE,LIST>/L<C<unpack>|/unpack TEMPLATE,EXPR> as a
5917substitute for the L<Encode> module.
5918
5919=item *
5920
5921You must yourself do any alignment or padding by inserting, for example,
5922enough C<"x">es while packing.  There is no way for
5923L<C<pack>|/pack TEMPLATE,LIST> and L<C<unpack>|/unpack TEMPLATE,EXPR>
5924to know where characters are going to or coming from, so they
5925handle their output and input as flat sequences of characters.
5926
5927=item *
5928
5929A C<()> group is a sub-TEMPLATE enclosed in parentheses.  A group may
5930take a repeat count either as postfix, or for
5931L<C<unpack>|/unpack TEMPLATE,EXPR>, also via the C</>
5932template character.  Within each repetition of a group, positioning with
5933C<@> starts over at 0.  Therefore, the result of
5934
5935    pack("@1A((@2A)@3A)", qw[X Y Z])
5936
5937is the string C<"\0X\0\0YZ">.
5938
5939=item *
5940
5941C<x> and C<X> accept the C<!> modifier to act as alignment commands: they
5942jump forward or back to the closest position aligned at a multiple of C<count>
5943characters.  For example, to L<C<pack>|/pack TEMPLATE,LIST> or
5944L<C<unpack>|/unpack TEMPLATE,EXPR> a C structure like
5945
5946    struct {
5947	char   c;    /* one signed, 8-bit character */
5948	double d;
5949	char   cc[2];
5950    }
5951
5952one may need to use the template C<c x![d] d c[2]>.  This assumes that
5953doubles must be aligned to the size of double.
5954
5955For alignment commands, a C<count> of 0 is equivalent to a C<count> of 1;
5956both are no-ops.
5957
5958=item *
5959
5960C<n>, C<N>, C<v> and C<V> accept the C<!> modifier to
5961represent signed 16-/32-bit integers in big-/little-endian order.
5962This is portable only when all platforms sharing packed data use the
5963same binary representation for signed integers; for example, when all
5964platforms use two's-complement representation.
5965
5966=item *
5967
5968Comments can be embedded in a TEMPLATE using C<#> through the end of line.
5969White space can separate pack codes from each other, but modifiers and
5970repeat counts must follow immediately.  Breaking complex templates into
5971individual line-by-line components, suitably annotated, can do as much to
5972improve legibility and maintainability of pack/unpack formats as C</x> can
5973for complicated pattern matches.
5974
5975=item *
5976
5977If TEMPLATE requires more arguments than L<C<pack>|/pack TEMPLATE,LIST>
5978is given, L<C<pack>|/pack TEMPLATE,LIST>
5979assumes additional C<""> arguments.  If TEMPLATE requires fewer arguments
5980than given, extra arguments are ignored.
5981
5982=item *
5983
5984Attempting to pack the special floating point values C<Inf> and C<NaN>
5985(infinity, also in negative, and not-a-number) into packed integer values
5986(like C<"L">) is a fatal error.  The reason for this is that there simply
5987isn't any sensible mapping for these special values into integers.
5988
5989=back
5990
5991Examples:
5992
5993    $foo = pack("WWWW",65,66,67,68);
5994    # foo eq "ABCD"
5995    $foo = pack("W4",65,66,67,68);
5996    # same thing
5997    $foo = pack("W4",0x24b6,0x24b7,0x24b8,0x24b9);
5998    # same thing with Unicode circled letters.
5999    $foo = pack("U4",0x24b6,0x24b7,0x24b8,0x24b9);
6000    # same thing with Unicode circled letters.  You don't get the
6001    # UTF-8 bytes because the U at the start of the format caused
6002    # a switch to U0-mode, so the UTF-8 bytes get joined into
6003    # characters
6004    $foo = pack("C0U4",0x24b6,0x24b7,0x24b8,0x24b9);
6005    # foo eq "\xe2\x92\xb6\xe2\x92\xb7\xe2\x92\xb8\xe2\x92\xb9"
6006    # This is the UTF-8 encoding of the string in the
6007    # previous example
6008
6009    $foo = pack("ccxxcc",65,66,67,68);
6010    # foo eq "AB\0\0CD"
6011
6012    # NOTE: The examples above featuring "W" and "c" are true
6013    # only on ASCII and ASCII-derived systems such as ISO Latin 1
6014    # and UTF-8.  On EBCDIC systems, the first example would be
6015    #      $foo = pack("WWWW",193,194,195,196);
6016
6017    $foo = pack("s2",1,2);
6018    # "\001\000\002\000" on little-endian
6019    # "\000\001\000\002" on big-endian
6020
6021    $foo = pack("a4","abcd","x","y","z");
6022    # "abcd"
6023
6024    $foo = pack("aaaa","abcd","x","y","z");
6025    # "axyz"
6026
6027    $foo = pack("a14","abcdefg");
6028    # "abcdefg\0\0\0\0\0\0\0"
6029
6030    $foo = pack("i9pl", gmtime);
6031    # a real struct tm (on my system anyway)
6032
6033    $utmp_template = "Z8 Z8 Z16 L";
6034    $utmp = pack($utmp_template, @utmp1);
6035    # a struct utmp (BSDish)
6036
6037    @utmp2 = unpack($utmp_template, $utmp);
6038    # "@utmp1" eq "@utmp2"
6039
6040    sub bintodec {
6041        unpack("N", pack("B32", substr("0" x 32 . shift, -32)));
6042    }
6043
6044    $foo = pack('sx2l', 12, 34);
6045    # short 12, two zero bytes padding, long 34
6046    $bar = pack('s@4l', 12, 34);
6047    # short 12, zero fill to position 4, long 34
6048    # $foo eq $bar
6049    $baz = pack('s.l', 12, 4, 34);
6050    # short 12, zero fill to position 4, long 34
6051
6052    $foo = pack('nN', 42, 4711);
6053    # pack big-endian 16- and 32-bit unsigned integers
6054    $foo = pack('S>L>', 42, 4711);
6055    # exactly the same
6056    $foo = pack('s<l<', -42, 4711);
6057    # pack little-endian 16- and 32-bit signed integers
6058    $foo = pack('(sl)<', -42, 4711);
6059    # exactly the same
6060
6061The same template may generally also be used in
6062L<C<unpack>|/unpack TEMPLATE,EXPR>.
6063
6064=item package NAMESPACE
6065
6066=item package NAMESPACE VERSION
6067X<package> X<module> X<namespace> X<version>
6068
6069=item package NAMESPACE BLOCK
6070
6071=item package NAMESPACE VERSION BLOCK
6072X<package> X<module> X<namespace> X<version>
6073
6074=for Pod::Functions declare a separate global namespace
6075
6076Declares the BLOCK or the rest of the compilation unit as being in the
6077given namespace.  The scope of the package declaration is either the
6078supplied code BLOCK or, in the absence of a BLOCK, from the declaration
6079itself through the end of current scope (the enclosing block, file, or
6080L<C<eval>|/eval EXPR>).  That is, the forms without a BLOCK are
6081operative through the end of the current scope, just like the
6082L<C<my>|/my VARLIST>, L<C<state>|/state VARLIST>, and
6083L<C<our>|/our VARLIST> operators.  All unqualified dynamic identifiers
6084in this scope will be in the given namespace, except where overridden by
6085another L<C<package>|/package NAMESPACE> declaration or
6086when they're one of the special identifiers that qualify into C<main::>,
6087like C<STDOUT>, C<ARGV>, C<ENV>, and the punctuation variables.
6088
6089A package statement affects dynamic variables only, including those
6090you've used L<C<local>|/local EXPR> on, but I<not> lexically-scoped
6091variables, which are created with L<C<my>|/my VARLIST>,
6092L<C<state>|/state VARLIST>, or L<C<our>|/our VARLIST>.  Typically it
6093would be the first declaration in a file included by
6094L<C<require>|/require VERSION> or L<C<use>|/use Module VERSION LIST>.
6095You can switch into a
6096package in more than one place, since this only determines which default
6097symbol table the compiler uses for the rest of that block.  You can refer to
6098identifiers in other packages than the current one by prefixing the identifier
6099with the package name and a double colon, as in C<$SomePack::var>
6100or C<ThatPack::INPUT_HANDLE>.  If package name is omitted, the C<main>
6101package is assumed.  That is, C<$::sail> is equivalent to
6102C<$main::sail> (as well as to C<$main'sail>, still seen in ancient
6103code, mostly from Perl 4).
6104
6105If VERSION is provided, L<C<package>|/package NAMESPACE> sets the
6106C<$VERSION> variable in the given
6107namespace to a L<version> object with the VERSION provided.  VERSION must be a
6108"strict" style version number as defined by the L<version> module: a positive
6109decimal number (integer or decimal-fraction) without exponentiation or else a
6110dotted-decimal v-string with a leading 'v' character and at least three
6111components.  You should set C<$VERSION> only once per package.
6112
6113See L<perlmod/"Packages"> for more information about packages, modules,
6114and classes.  See L<perlsub> for other scoping issues.
6115
6116=item __PACKAGE__
6117X<__PACKAGE__>
6118
6119=for Pod::Functions +5.004 the current package
6120
6121A special token that returns the name of the package in which it occurs.
6122
6123=item __CLASS__
6124X<__CLASS__>
6125
6126=for Pod::Functions the class name of the current instance.
6127
6128Invoked within a L<C<method>|/method NAME BLOCK>, or similar location, such as
6129a field initializer expression, this token returns the name of the class of
6130the invoking instance.  Essentially it is equivalent to C<ref($self)> except
6131that it can additionally be used in a field initializer to gain access to
6132class methods, before the instance is fully constructed.
6133
6134    use feature 'class';
6135
6136    class Example1 {
6137	field $f = __CLASS__->default_f;
6138
6139	sub default_f { 10 }
6140    }
6141
6142In a basic class, this value will be the same as
6143L<C<__PACKAGE__>|/__PACKAGE__>.  The distinction can be seen when a subclass
6144is constructed; it will give the class name of the instance being constructed,
6145rather than just the package name that the actual code belongs to.
6146
6147    class Example2 :isa(Example1) {
6148	sub default_f { 20 }
6149    }
6150
6151    my $obj = Example2->new;
6152    # The $f field now has the value 20
6153
6154=item pipe READHANDLE,WRITEHANDLE
6155X<pipe>
6156
6157=for Pod::Functions open a pair of connected filehandles
6158
6159Opens a pair of connected pipes like the corresponding system call.
6160Note that if you set up a loop of piped processes, deadlock can occur
6161unless you are very careful.  In addition, note that Perl's pipes use
6162IO buffering, so you may need to set L<C<$E<verbar>>|perlvar/$E<verbar>>
6163to flush your WRITEHANDLE after each command, depending on the
6164application.
6165
6166Returns true on success.
6167
6168See L<IPC::Open2>, L<IPC::Open3>, and
6169L<perlipc/"Bidirectional Communication with Another Process">
6170for examples of such things.
6171
6172On systems that support a close-on-exec flag on files, that flag is set
6173on all newly opened file descriptors whose
6174L<C<fileno>|/fileno FILEHANDLE>s are I<higher> than the current value of
6175L<C<$^F>|perlvar/$^F> (by default 2 for C<STDERR>).  See L<perlvar/$^F>.
6176
6177=item pop ARRAY
6178X<pop> X<stack>
6179
6180=item pop
6181
6182=for Pod::Functions remove the last element from an array and return it
6183
6184Removes and returns the B<last> element of the array, shortening the array by
6185one element.
6186
6187    my @arr  = ('cat', 'dog', 'mouse');
6188    my $item = pop(@arr); # 'mouse'
6189
6190    # @arr is now ('cat', 'dog')
6191
6192Returns C<undef> if the array is empty.
6193
6194B<Note:> C<pop> may also return C<undef> if the last element in the array
6195is C<undef>.
6196
6197    my @arr  = ('one', 'two', undef);
6198    my $item = pop(@arr); # undef
6199
6200If ARRAY is omitted, C<pop> operates on the L<C<@ARGV>|perlvar/@ARGV> array
6201in the main program, but the L<C<@_>|perlvar/@_> array in subroutines. C<pop>
6202will operate on the C<@ARGV> array in C<eval STRING>, C<BEGIN {}>, C<INIT {}>,
6203C<CHECK {}> blocks.
6204
6205Starting with Perl 5.14, an experimental feature allowed
6206L<C<pop>|/pop ARRAY> to take a
6207scalar expression. This experiment has been deemed unsuccessful, and was
6208removed as of Perl 5.24.
6209
6210=item pos SCALAR
6211X<pos> X<match, position>
6212
6213=item pos
6214
6215=for Pod::Functions find or set the offset for the last/next m//g search
6216
6217Returns the offset of where the last C<m//g> search left off for the
6218variable in question (L<C<$_>|perlvar/$_> is used when the variable is not
6219specified).  This offset is in characters unless the
6220(no-longer-recommended) L<C<use bytes>|bytes> pragma is in effect, in
6221which case the offset is in bytes.  Note that 0 is a valid match offset.
6222L<C<undef>|/undef EXPR> indicates
6223that the search position is reset (usually due to match failure, but
6224can also be because no match has yet been run on the scalar).
6225
6226L<C<pos>|/pos SCALAR> directly accesses the location used by the regexp
6227engine to store the offset, so assigning to L<C<pos>|/pos SCALAR> will
6228change that offset, and so will also influence the C<\G> zero-width
6229assertion in regular expressions.  Both of these effects take place for
6230the next match, so you can't affect the position with
6231L<C<pos>|/pos SCALAR> during the current match, such as in
6232C<(?{pos() = 5})> or C<s//pos() = 5/e>.
6233
6234Setting L<C<pos>|/pos SCALAR> also resets the I<matched with
6235zero-length> flag, described
6236under L<perlre/"Repeated Patterns Matching a Zero-length Substring">.
6237
6238Because a failed C<m//gc> match doesn't reset the offset, the return
6239from L<C<pos>|/pos SCALAR> won't change either in this case.  See
6240L<perlre> and L<perlop>.
6241
6242=item print FILEHANDLE LIST
6243X<print>
6244
6245=item print FILEHANDLE
6246
6247=item print LIST
6248
6249=item print
6250
6251=for Pod::Functions output a list to a filehandle
6252
6253Prints a string or a list of strings.  Returns true if successful.
6254FILEHANDLE may be a scalar variable containing the name of or a reference
6255to the filehandle, thus introducing one level of indirection.  (NOTE: If
6256FILEHANDLE is a variable and the next token is a term, it may be
6257misinterpreted as an operator unless you interpose a C<+> or put
6258parentheses around the arguments.)  If FILEHANDLE is omitted, prints to the
6259last selected (see L<C<select>|/select FILEHANDLE>) output handle.  If
6260LIST is omitted, prints L<C<$_>|perlvar/$_> to the currently selected
6261output handle.  To use FILEHANDLE alone to print the content of
6262L<C<$_>|perlvar/$_> to it, you must use a bareword filehandle like
6263C<FH>, not an indirect one like C<$fh>.  To set the default output handle
6264to something other than STDOUT, use the select operation.
6265
6266The current value of L<C<$,>|perlvar/$,> (if any) is printed between
6267each LIST item.  The current value of L<C<$\>|perlvar/$\> (if any) is
6268printed after the entire LIST has been printed.  Because print takes a
6269LIST, anything in the LIST is evaluated in list context, including any
6270subroutines whose return lists you pass to
6271L<C<print>|/print FILEHANDLE LIST>.  Be careful not to follow the print
6272keyword with a left
6273parenthesis unless you want the corresponding right parenthesis to
6274terminate the arguments to the print; put parentheses around all arguments
6275(or interpose a C<+>, but that doesn't look as good).
6276
6277If you're storing handles in an array or hash, or in general whenever
6278you're using any expression more complex than a bareword handle or a plain,
6279unsubscripted scalar variable to retrieve it, you will have to use a block
6280returning the filehandle value instead, in which case the LIST may not be
6281omitted:
6282
6283    print { $files[$i] } "stuff\n";
6284    print { $OK ? *STDOUT : *STDERR } "stuff\n";
6285
6286Printing to a closed pipe or socket will generate a SIGPIPE signal.  See
6287L<perlipc> for more on signal handling.
6288
6289=item printf FILEHANDLE FORMAT, LIST
6290X<printf>
6291
6292=item printf FILEHANDLE
6293
6294=item printf FORMAT, LIST
6295
6296=item printf
6297
6298=for Pod::Functions output a formatted list to a filehandle
6299
6300Equivalent to C<print FILEHANDLE sprintf(FORMAT, LIST)>, except that
6301L<C<$\>|perlvar/$\> (the output record separator) is not appended.  The
6302FORMAT and the LIST are actually parsed as a single list.  The first
6303argument of the list will be interpreted as the
6304L<C<printf>|/printf FILEHANDLE FORMAT, LIST> format.  This means that
6305C<printf(@_)> will use C<$_[0]> as the format.  See
6306L<sprintf|/sprintf FORMAT, LIST> for an explanation of the format
6307argument.  If C<use locale> (including C<use locale ':not_characters'>)
6308is in effect and L<C<POSIX::setlocale>|POSIX/C<setlocale>> has been
6309called, the character used for the decimal separator in formatted
6310floating-point numbers is affected by the C<LC_NUMERIC> locale setting.
6311See L<perllocale> and L<POSIX>.
6312
6313For historical reasons, if you omit the list, L<C<$_>|perlvar/$_> is
6314used as the format;
6315to use FILEHANDLE without a list, you must use a bareword filehandle like
6316C<FH>, not an indirect one like C<$fh>.  However, this will rarely do what
6317you want; if L<C<$_>|perlvar/$_> contains formatting codes, they will be
6318replaced with the empty string and a warning will be emitted if
6319L<warnings> are enabled.  Just use L<C<print>|/print FILEHANDLE LIST> if
6320you want to print the contents of L<C<$_>|perlvar/$_>.
6321
6322Don't fall into the trap of using a
6323L<C<printf>|/printf FILEHANDLE FORMAT, LIST> when a simple
6324L<C<print>|/print FILEHANDLE LIST> would do.  The
6325L<C<print>|/print FILEHANDLE LIST> is more efficient and less error
6326prone.
6327
6328=item prototype FUNCTION
6329X<prototype>
6330
6331=item prototype
6332
6333=for Pod::Functions +5.002 get the prototype (if any) of a subroutine
6334
6335Returns the prototype of a function as a string (or
6336L<C<undef>|/undef EXPR> if the
6337function has no prototype).  FUNCTION is a reference to, or the name of,
6338the function whose prototype you want to retrieve.  If FUNCTION is omitted,
6339L<C<$_>|perlvar/$_> is used.
6340
6341If FUNCTION is a string starting with C<CORE::>, the rest is taken as a
6342name for a Perl builtin.  If the builtin's arguments
6343cannot be adequately expressed by a prototype
6344(such as L<C<system>|/system LIST>), L<C<prototype>|/prototype FUNCTION>
6345returns L<C<undef>|/undef EXPR>, because the builtin
6346does not really behave like a Perl function.  Otherwise, the string
6347describing the equivalent prototype is returned.
6348
6349=item push ARRAY,LIST
6350X<push> X<stack>
6351
6352=for Pod::Functions append one or more elements to an array
6353
6354Adds one or more items to the B<end> of an array.
6355
6356	my @animals = ("cat");
6357	push(@animals, "mouse"); # ("cat", "mouse")
6358
6359	my @colors = ("red");
6360	push(@colors, ("blue", "green")); # ("red", "blue", "green")
6361
6362Returns the number of elements in the array following the completed
6363L<C<push>|/push ARRAY,LIST>.
6364
6365	my $color_count = push(@colors, ("yellow", "purple"));
6366
6367	say "There are $color_count colors in the updated array";
6368
6369Starting with Perl 5.14, an experimental feature allowed
6370L<C<push>|/push ARRAY,LIST> to take a
6371scalar expression. This experiment has been deemed unsuccessful, and was
6372removed as of Perl 5.24.
6373
6374=item q/STRING/
6375
6376=for Pod::Functions singly quote a string
6377
6378=item qq/STRING/
6379
6380=for Pod::Functions doubly quote a string
6381
6382=item qw/STRING/
6383
6384=for Pod::Functions quote a list of words
6385
6386=item qx/STRING/
6387
6388=for Pod::Functions backquote quote a string
6389
6390Generalized quotes.  See L<perlop/"Quote-Like Operators">.
6391
6392=item qr/STRING/
6393
6394=for Pod::Functions +5.005 compile pattern
6395
6396Regexp-like quote.  See L<perlop/"Regexp Quote-Like Operators">.
6397
6398=item quotemeta EXPR
6399X<quotemeta> X<metacharacter>
6400
6401=item quotemeta
6402
6403=for Pod::Functions quote regular expression magic characters
6404
6405Returns the value of EXPR with all the ASCII non-"word"
6406characters backslashed.  (That is, all ASCII characters not matching
6407C</[A-Za-z_0-9]/> will be preceded by a backslash in the
6408returned string, regardless of any locale settings.)
6409This is the internal function implementing
6410the C<\Q> escape in double-quoted strings.
6411(See below for the behavior on non-ASCII code points.)
6412
6413If EXPR is omitted, uses L<C<$_>|perlvar/$_>.
6414
6415quotemeta (and C<\Q> ... C<\E>) are useful when interpolating strings into
6416regular expressions, because by default an interpolated variable will be
6417considered a mini-regular expression.  For example:
6418
6419    my $sentence = 'The quick brown fox jumped over the lazy dog';
6420    my $substring = 'quick.*?fox';
6421    $sentence =~ s{$substring}{big bad wolf};
6422
6423Will cause C<$sentence> to become C<'The big bad wolf jumped over...'>.
6424
6425On the other hand:
6426
6427    my $sentence = 'The quick brown fox jumped over the lazy dog';
6428    my $substring = 'quick.*?fox';
6429    $sentence =~ s{\Q$substring\E}{big bad wolf};
6430
6431Or:
6432
6433    my $sentence = 'The quick brown fox jumped over the lazy dog';
6434    my $substring = 'quick.*?fox';
6435    my $quoted_substring = quotemeta($substring);
6436    $sentence =~ s{$quoted_substring}{big bad wolf};
6437
6438Will both leave the sentence as is.
6439Normally, when accepting literal string input from the user,
6440L<C<quotemeta>|/quotemeta EXPR> or C<\Q> must be used.
6441
6442Beware that if you put literal backslashes (those not inside
6443interpolated variables) between C<\Q> and C<\E>, double-quotish
6444backslash interpolation may lead to confusing results.  If you
6445I<need> to use literal backslashes within C<\Q...\E>,
6446consult L<perlop/"Gory details of parsing quoted constructs">.
6447
6448Because the result of S<C<"\Q I<STRING> \E">> has all metacharacters
6449quoted, there is no way to insert a literal C<$> or C<@> inside a
6450C<\Q\E> pair.  If protected by C<\>, C<$> will be quoted to become
6451C<"\\\$">; if not, it is interpreted as the start of an interpolated
6452scalar.
6453
6454In Perl v5.14, all non-ASCII characters are quoted in non-UTF-8-encoded
6455strings, but not quoted in UTF-8 strings.
6456
6457Starting in Perl v5.16, Perl adopted a Unicode-defined strategy for
6458quoting non-ASCII characters; the quoting of ASCII characters is
6459unchanged.
6460
6461Also unchanged is the quoting of non-UTF-8 strings when outside the
6462scope of a
6463L<C<use feature 'unicode_strings'>|feature/The 'unicode_strings' feature>,
6464which is to quote all
6465characters in the upper Latin1 range.  This provides complete backwards
6466compatibility for old programs which do not use Unicode.  (Note that
6467C<unicode_strings> is automatically enabled within the scope of a
6468S<C<use v5.12>> or greater.)
6469
6470Within the scope of L<C<use locale>|locale>, all non-ASCII Latin1 code
6471points
6472are quoted whether the string is encoded as UTF-8 or not.  As mentioned
6473above, locale does not affect the quoting of ASCII-range characters.
6474This protects against those locales where characters such as C<"|"> are
6475considered to be word characters.
6476
6477Otherwise, Perl quotes non-ASCII characters using an adaptation from
6478Unicode (see L<https://www.unicode.org/reports/tr31/>).
6479The only code points that are quoted are those that have any of the
6480Unicode properties:  Pattern_Syntax, Pattern_White_Space, White_Space,
6481Default_Ignorable_Code_Point, or General_Category=Control.
6482
6483Of these properties, the two important ones are Pattern_Syntax and
6484Pattern_White_Space.  They have been set up by Unicode for exactly this
6485purpose of deciding which characters in a regular expression pattern
6486should be quoted.  No character that can be in an identifier has these
6487properties.
6488
6489Perl promises, that if we ever add regular expression pattern
6490metacharacters to the dozen already defined
6491(C<\ E<verbar> ( ) [ { ^ $ * + ? .>), that we will only use ones that have the
6492Pattern_Syntax property.  Perl also promises, that if we ever add
6493characters that are considered to be white space in regular expressions
6494(currently mostly affected by C</x>), they will all have the
6495Pattern_White_Space property.
6496
6497Unicode promises that the set of code points that have these two
6498properties will never change, so something that is not quoted in v5.16
6499will never need to be quoted in any future Perl release.  (Not all the
6500code points that match Pattern_Syntax have actually had characters
6501assigned to them; so there is room to grow, but they are quoted
6502whether assigned or not.  Perl, of course, would never use an
6503unassigned code point as an actual metacharacter.)
6504
6505Quoting characters that have the other 3 properties is done to enhance
6506the readability of the regular expression and not because they actually
6507need to be quoted for regular expression purposes (characters with the
6508White_Space property are likely to be indistinguishable on the page or
6509screen from those with the Pattern_White_Space property; and the other
6510two properties contain non-printing characters).
6511
6512=item rand EXPR
6513X<rand> X<random>
6514
6515=item rand
6516
6517=for Pod::Functions retrieve the next pseudorandom number
6518
6519Returns a random fractional number greater than or equal to C<0> and less
6520than the value of EXPR.  (EXPR should be positive.)  If EXPR is
6521omitted, the value C<1> is used.  Currently EXPR with the value C<0> is
6522also special-cased as C<1> (this was undocumented before Perl 5.8.0
6523and is subject to change in future versions of Perl).  Automatically calls
6524L<C<srand>|/srand EXPR> unless L<C<srand>|/srand EXPR> has already been
6525called.  See also L<C<srand>|/srand EXPR>.
6526
6527Apply L<C<int>|/int EXPR> to the value returned by L<C<rand>|/rand EXPR>
6528if you want random integers instead of random fractional numbers.  For
6529example,
6530
6531    int(rand(10))
6532
6533returns a random integer between C<0> and C<9>, inclusive.
6534
6535(Note: If your rand function consistently returns numbers that are too
6536large or too small, then your version of Perl was probably compiled
6537with the wrong number of RANDBITS.)
6538
6539B<L<C<rand>|/rand EXPR> is not cryptographically secure.  You should not rely
6540on it in security-sensitive situations.>  As of this writing, a
6541number of third-party CPAN modules offer random number generators
6542intended by their authors to be cryptographically secure,
6543including: L<Data::Entropy>, L<Crypt::Random>, L<Math::Random::Secure>,
6544and L<Math::TrulyRandom>.
6545
6546=item read FILEHANDLE,SCALAR,LENGTH,OFFSET
6547X<read> X<file, read>
6548
6549=item read FILEHANDLE,SCALAR,LENGTH
6550
6551=for Pod::Functions fixed-length buffered input from a filehandle
6552
6553Attempts to read LENGTH I<characters> of data into variable SCALAR
6554from the specified FILEHANDLE.  Returns the number of characters
6555actually read, C<0> at end of file, or undef if there was an error (in
6556the latter case L<C<$!>|perlvar/$!> is also set).  SCALAR will be grown
6557or shrunk
6558so that the last character actually read is the last character of the
6559scalar after the read.
6560
6561An OFFSET may be specified to place the read data at some place in the
6562string other than the beginning.  A negative OFFSET specifies
6563placement at that many characters counting backwards from the end of
6564the string.  A positive OFFSET greater than the length of SCALAR
6565results in the string being padded to the required size with C<"\0">
6566bytes before the result of the read is appended.
6567
6568The call is implemented in terms of either Perl's or your system's native
6569L<fread(3)> library function, via the L<PerlIO> layers applied to the
6570handle.  To get a true L<read(2)> system call, see
6571L<sysread|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET>.
6572
6573Note the I<characters>: depending on the status of the filehandle,
6574either (8-bit) bytes or characters are read.  By default, all
6575filehandles operate on bytes, but for example if the filehandle has
6576been opened with the C<:utf8> I/O layer (see
6577L<C<open>|/open FILEHANDLE,MODE,EXPR>, and the L<open>
6578pragma), the I/O will operate on UTF8-encoded Unicode
6579characters, not bytes.  Similarly for the C<:encoding> layer:
6580in that case pretty much any characters can be read.
6581
6582=item readdir DIRHANDLE
6583X<readdir>
6584
6585=for Pod::Functions get a directory from a directory handle
6586
6587Returns the next directory entry for a directory opened by
6588L<C<opendir>|/opendir DIRHANDLE,EXPR>.
6589If used in list context, returns all the rest of the entries in the
6590directory.  If there are no more entries, returns the undefined value in
6591scalar context and the empty list in list context.
6592
6593If you're planning to filetest the return values out of a
6594L<C<readdir>|/readdir DIRHANDLE>, you'd better prepend the directory in
6595question.  Otherwise, because we didn't L<C<chdir>|/chdir EXPR> there,
6596it would have been testing the wrong file.
6597
6598    opendir(my $dh, $some_dir) || die "Can't opendir $some_dir: $!";
6599    my @dots = grep { /^\./ && -f "$some_dir/$_" } readdir($dh);
6600    closedir $dh;
6601
6602As of Perl 5.12 you can use a bare L<C<readdir>|/readdir DIRHANDLE> in a
6603C<while> loop, which will set L<C<$_>|perlvar/$_> on every iteration.
6604If either a C<readdir> expression or an explicit assignment of a
6605C<readdir> expression to a scalar is used as a C<while>/C<for> condition,
6606then the condition actually tests for definedness of the expression's
6607value, not for its regular truth value.
6608
6609    opendir(my $dh, $some_dir) || die "Can't open $some_dir: $!";
6610    while (readdir $dh) {
6611        print "$some_dir/$_\n";
6612    }
6613    closedir $dh;
6614
6615To avoid confusing would-be users of your code who are running earlier
6616versions of Perl with mysterious failures, put this sort of thing at the
6617top of your file to signal that your code will work I<only> on Perls of a
6618recent vintage:
6619
6620    use v5.12; # so readdir assigns to $_ in a lone while test
6621
6622=item readline EXPR
6623
6624=item readline
6625X<readline> X<gets> X<fgets>
6626
6627=for Pod::Functions fetch a record from a file
6628
6629Reads from the filehandle whose typeglob is contained in EXPR (or from
6630C<*ARGV> if EXPR is not provided).  In scalar context, each call reads and
6631returns the next line until end-of-file is reached, whereupon the
6632subsequent call returns L<C<undef>|/undef EXPR>.  In list context, reads
6633until end-of-file is reached and returns a list of lines.  Note that the
6634notion of "line" used here is whatever you may have defined with
6635L<C<$E<sol>>|perlvar/$E<sol>> (or C<$INPUT_RECORD_SEPARATOR> in
6636L<English>).  See L<perlvar/"$/">.
6637
6638When L<C<$E<sol>>|perlvar/$E<sol>> is set to L<C<undef>|/undef EXPR>,
6639when L<C<readline>|/readline EXPR> is in scalar context (i.e., file
6640slurp mode), and when an empty file is read, it returns C<''> the first
6641time, followed by L<C<undef>|/undef EXPR> subsequently.
6642
6643This is the internal function implementing the C<< <EXPR> >>
6644operator, but you can use it directly.  The C<< <EXPR> >>
6645operator is discussed in more detail in L<perlop/"I/O Operators">.
6646
6647    my $line = <STDIN>;
6648    my $line = readline(STDIN);    # same thing
6649
6650If L<C<readline>|/readline EXPR> encounters an operating system error,
6651L<C<$!>|perlvar/$!> will be set with the corresponding error message.
6652It can be helpful to check L<C<$!>|perlvar/$!> when you are reading from
6653filehandles you don't trust, such as a tty or a socket.  The following
6654example uses the operator form of L<C<readline>|/readline EXPR> and dies
6655if the result is not defined.
6656
6657    while ( ! eof($fh) ) {
6658        defined( $_ = readline $fh ) or die "readline failed: $!";
6659        ...
6660    }
6661
6662Note that you can't handle L<C<readline>|/readline EXPR> errors
6663that way with the C<ARGV> filehandle.  In that case, you have to open
6664each element of L<C<@ARGV>|perlvar/@ARGV> yourself since
6665L<C<eof>|/eof FILEHANDLE> handles C<ARGV> differently.
6666
6667    foreach my $arg (@ARGV) {
6668        open(my $fh, $arg) or warn "Can't open $arg: $!";
6669
6670        while ( ! eof($fh) ) {
6671            defined( $_ = readline $fh )
6672                or die "readline failed for $arg: $!";
6673            ...
6674        }
6675    }
6676
6677Like the C<< <EXPR> >> operator, if a C<readline> expression is
6678used as the condition of a C<while> or C<for> loop, then it will be
6679implicitly assigned to C<$_>.  If either a C<readline> expression or
6680an explicit assignment of a C<readline> expression to a scalar is used
6681as a C<while>/C<for> condition, then the condition actually tests for
6682definedness of the expression's value, not for its regular truth value.
6683
6684=item readlink EXPR
6685X<readlink>
6686
6687=item readlink
6688
6689=for Pod::Functions determine where a symbolic link is pointing
6690
6691Returns the value of a symbolic link, if symbolic links are
6692implemented.  If not, raises an exception.  If there is a system
6693error, returns the undefined value and sets L<C<$!>|perlvar/$!> (errno).
6694If EXPR is omitted, uses L<C<$_>|perlvar/$_>.
6695
6696Portability issues: L<perlport/readlink>.
6697
6698=item readpipe EXPR
6699
6700=item readpipe
6701X<readpipe>
6702
6703=for Pod::Functions execute a system command and collect standard output
6704
6705EXPR is executed as a system command.
6706The collected standard output of the command is returned.
6707In scalar context, it comes back as a single (potentially
6708multi-line) string.  In list context, returns a list of lines
6709(however you've defined lines with L<C<$E<sol>>|perlvar/$E<sol>> (or
6710C<$INPUT_RECORD_SEPARATOR> in L<English>)).
6711This is the internal function implementing the C<qx/EXPR/>
6712operator, but you can use it directly.  The C<qx/EXPR/>
6713operator is discussed in more detail in L<perlop/"C<qx/I<STRING>/>">.
6714If EXPR is omitted, uses L<C<$_>|perlvar/$_>.
6715
6716=item recv SOCKET,SCALAR,LENGTH,FLAGS
6717X<recv>
6718
6719=for Pod::Functions receive a message over a Socket
6720
6721Receives a message on a socket.  Attempts to receive LENGTH characters
6722of data into variable SCALAR from the specified SOCKET filehandle.
6723SCALAR will be grown or shrunk to the length actually read.  Takes the
6724same flags as the system call of the same name.  Returns the address
6725of the sender if SOCKET's protocol supports this; returns an empty
6726string otherwise.  If there's an error, returns the undefined value.
6727This call is actually implemented in terms of the L<recvfrom(2)> system call.
6728See L<perlipc/"UDP: Message Passing"> for examples.
6729
6730Note that if the socket has been marked as C<:utf8>, C<recv> will
6731throw an exception.  The C<:encoding(...)> layer implicitly introduces
6732the C<:utf8> layer.  See L<C<binmode>|/binmode FILEHANDLE, LAYER>.
6733
6734=item redo LABEL
6735X<redo>
6736
6737=item redo EXPR
6738
6739=item redo
6740
6741=for Pod::Functions start this loop iteration over again
6742
6743The L<C<redo>|/redo LABEL> command restarts the loop block without
6744evaluating the conditional again.  The L<C<continue>|/continue BLOCK>
6745block, if any, is not executed.  If
6746the LABEL is omitted, the command refers to the innermost enclosing
6747loop.  The C<redo EXPR> form, available starting in Perl 5.18.0, allows a
6748label name to be computed at run time, and is otherwise identical to C<redo
6749LABEL>.  Programs that want to lie to themselves about what was just input
6750normally use this command:
6751
6752    # a simpleminded Pascal comment stripper
6753    # (warning: assumes no { or } in strings)
6754    LINE: while (<STDIN>) {
6755        while (s|({.*}.*){.*}|$1 |) {}
6756        s|{.*}| |;
6757        if (s|{.*| |) {
6758            my $front = $_;
6759            while (<STDIN>) {
6760                if (/}/) {  # end of comment?
6761                    s|^|$front\{|;
6762                    redo LINE;
6763                }
6764            }
6765        }
6766        print;
6767    }
6768
6769L<C<redo>|/redo LABEL> cannot return a value from a block that typically
6770returns a value, such as C<eval {}>, C<sub {}>, or C<do {}>. It will perform
6771its flow control behavior, which precludes any return value. It should not be
6772used to exit a L<C<grep>|/grep BLOCK LIST> or L<C<map>|/map BLOCK LIST>
6773operation.
6774
6775Note that a block by itself is semantically identical to a loop
6776that executes once.  Thus L<C<redo>|/redo LABEL> inside such a block
6777will effectively turn it into a looping construct.
6778
6779See also L<C<continue>|/continue BLOCK> for an illustration of how
6780L<C<last>|/last LABEL>, L<C<next>|/next LABEL>, and
6781L<C<redo>|/redo LABEL> work.
6782
6783Unlike most named operators, this has the same precedence as assignment.
6784It is also exempt from the looks-like-a-function rule, so
6785C<redo ("foo")."bar"> will cause "bar" to be part of the argument to
6786L<C<redo>|/redo LABEL>.
6787
6788=item ref EXPR
6789X<ref> X<reference>
6790
6791=item ref
6792
6793=for Pod::Functions find out the type of thing being referenced
6794
6795Examines the value of EXPR, expecting it to be a reference, and returns
6796a string giving information about the reference and the type of referent.
6797If EXPR is not specified, L<C<$_>|perlvar/$_> will be used.
6798
6799If the operand is not a reference, then the empty string will be returned.
6800An empty string will only be returned in this situation.  C<ref> is often
6801useful to just test whether a value is a reference, which can be done
6802by comparing the result to the empty string.  It is a common mistake
6803to use the result of C<ref> directly as a truth value: this goes wrong
6804because C<0> (which is false) can be returned for a reference.
6805
6806If the operand is a reference to a blessed object, then the name of
6807the class into which the referent is blessed will be returned.  C<ref>
6808doesn't care what the physical type of the referent is; blessing takes
6809precedence over such concerns.  Beware that exact comparison of C<ref>
6810results against a class name doesn't perform a class membership test:
6811a class's members also include objects blessed into subclasses, for
6812which C<ref> will return the name of the subclass.  Also beware that
6813class names can clash with the built-in type names (described below).
6814
6815If the operand is a reference to an unblessed object, then the return
6816value indicates the type of object.  If the unblessed referent is not
6817a scalar, then the return value will be one of the strings C<ARRAY>,
6818C<HASH>, C<CODE>, C<FORMAT>, or C<IO>, indicating only which kind of
6819object it is.  If the unblessed referent is a scalar, then the return
6820value will be one of the strings C<SCALAR>, C<VSTRING>, C<REF>, C<GLOB>,
6821C<LVALUE>, or C<REGEXP>, depending on the kind of value the scalar
6822currently has.   But note that C<qr//> scalars are created already
6823blessed, so C<ref qr/.../> will likely return C<Regexp>.  Beware that
6824these built-in type names can also be used as
6825class names, so C<ref> returning one of these names doesn't unambiguously
6826indicate that the referent is of the kind to which the name refers.
6827
6828The ambiguity between built-in type names and class names significantly
6829limits the utility of C<ref>.  For unambiguous information, use
6830L<C<Scalar::Util::blessed()>|Scalar::Util/blessed> for information about
6831blessing, and L<C<Scalar::Util::reftype()>|Scalar::Util/reftype> for
6832information about physical types.  Use L<the C<isa> method|UNIVERSAL/C<<
6833$obj->isa( TYPE ) >>> for class membership tests, though one must be
6834sure of blessedness before attempting a method call.  Alternatively, the
6835L<C<isa> operator|perlop/"Class Instance Operator"> can test class
6836membership without checking blessedness first.
6837
6838See also L<perlref> and L<perlobj>.
6839
6840=item rename OLDNAME,NEWNAME
6841X<rename> X<move> X<mv> X<ren>
6842
6843=for Pod::Functions change a filename
6844
6845Changes the name of a file; an existing file NEWNAME will be
6846clobbered.  Returns true for success; on failure returns false and sets
6847L<C<$!>|perlvar/$!>.
6848
6849Behavior of this function varies wildly depending on your system
6850implementation.  For example, it will usually not work across file system
6851boundaries, even though the system I<mv> command sometimes compensates
6852for this.  Other restrictions include whether it works on directories,
6853open files, or pre-existing files.  Check L<perlport> and either the
6854L<rename(2)> manpage or equivalent system documentation for details.
6855
6856For a platform independent L<C<move>|File::Copy/move> function look at
6857the L<File::Copy> module.
6858
6859Portability issues: L<perlport/rename>.
6860
6861=item require VERSION
6862X<require>
6863
6864=item require EXPR
6865
6866=item require
6867
6868=for Pod::Functions load in external functions from a library at runtime
6869
6870Demands a version of Perl specified by VERSION, or demands some semantics
6871specified by EXPR or by L<C<$_>|perlvar/$_> if EXPR is not supplied.
6872
6873VERSION may be either a literal such as v5.24.1, which will be
6874compared to L<C<$^V>|perlvar/$^V> (or C<$PERL_VERSION> in L<English>),
6875or a numeric argument of the form 5.024001, which will be compared to
6876L<C<$]>|perlvar/$]>. An exception is raised if VERSION is greater than
6877the version of the current Perl interpreter.  Compare with
6878L<C<use>|/use Module VERSION LIST>, which can do a similar check at
6879compile time.
6880
6881Specifying VERSION as a numeric argument of the form 5.024001 should
6882generally be avoided as older less readable syntax compared to
6883v5.24.1. Before perl 5.8.0 (released in 2002), the more verbose numeric
6884form was the only supported syntax, which is why you might see it in
6885older code.
6886
6887    require v5.24.1;    # run time version check
6888    require 5.24.1;     # ditto
6889    require 5.024_001;  # ditto; older syntax compatible
6890                          with perl 5.6
6891
6892Otherwise, L<C<require>|/require VERSION> demands that a library file be
6893included if it hasn't already been included.  The file is included via
6894the do-FILE mechanism, which is essentially just a variety of
6895L<C<eval>|/eval EXPR> with the
6896caveat that lexical variables in the invoking script will be invisible
6897to the included code.  If it were implemented in pure Perl, it
6898would have semantics similar to the following:
6899
6900    use Carp 'croak';
6901    use version;
6902
6903    sub require {
6904        my ($filename) = @_;
6905        if ( my $version = eval { version->parse($filename) } ) {
6906            if ( $version > $^V ) {
6907               my $vn = $version->normal;
6908               croak "Perl $vn required--this is only $^V, stopped";
6909            }
6910            return 1;
6911        }
6912
6913        if (exists $INC{$filename}) {
6914            return 1 if $INC{$filename};
6915            croak "Compilation failed in require";
6916        }
6917
6918        local $INC;
6919        # this type of loop lets a hook overwrite $INC if they wish
6920        for($INC = 0; $INC < @INC; $INC++) {
6921            my $prefix = $INC[$INC];
6922            if (!defined $prefix) {
6923                next;
6924            }
6925            if (ref $prefix) {
6926                #... do other stuff - see text below ....
6927            }
6928            # (see text below about possible appending of .pmc
6929            # suffix to $filename)
6930            my $realfilename = "$prefix/$filename";
6931            next if ! -e $realfilename || -d _ || -b _;
6932            $INC{$filename} = $realfilename;
6933            my $result = do($realfilename);
6934                         # but run in caller's namespace
6935
6936            if (!defined $result) {
6937                $INC{$filename} = undef;
6938                croak $@ ? "$@Compilation failed in require"
6939                         : "Can't locate $filename: $!\n";
6940            }
6941            if (!$result) {
6942                delete $INC{$filename};
6943                croak "$filename did not return true value";
6944            }
6945            $! = 0;
6946            return $result;
6947        }
6948        croak "Can't locate $filename in \@INC ...";
6949    }
6950
6951Note that the file will not be included twice under the same specified
6952name.
6953
6954Historically the file must return true as the last statement to indicate
6955successful execution of any initialization code, so it's customary to
6956end such a file with C<1;> unless you're sure it'll return true
6957otherwise.  But it's better just to put the C<1;>, in case you add more
6958statements. As of 5.37.6 this requirement may be avoided by enabling
6959the 'module_true' feature, which is enabled by default in modern
6960version bundles. Thus code with C<use v5.37;> no longer needs to concern
6961itself with this issue. See L<feature> for more details. Note that this
6962affects the compilation unit within which the feature is used, and using
6963it before requiring a module will not change the behavior of existing
6964modules that do not themselves also use it.
6965
6966If EXPR is a bareword, L<C<require>|/require VERSION> assumes a F<.pm>
6967extension and replaces C<::> with C</> in the filename for you,
6968to make it easy to load standard modules.  This form of loading of
6969modules does not risk altering your namespace, however it will autovivify
6970the stash for the required module.
6971
6972In other words, if you try this:
6973
6974        require Foo::Bar;     # a splendid bareword
6975
6976The require function will actually look for the F<Foo/Bar.pm> file in the
6977directories specified in the L<C<@INC>|perlvar/@INC> array, and it will
6978autovivify the C<Foo::Bar::> stash at compile time.
6979
6980But if you try this:
6981
6982        my $class = 'Foo::Bar';
6983        require $class;       # $class is not a bareword
6984    #or
6985        require "Foo::Bar";   # not a bareword because of the ""
6986
6987The require function will look for the F<Foo::Bar> file in the
6988L<C<@INC>|perlvar/@INC>  array and
6989will complain about not finding F<Foo::Bar> there.  In this case you can do:
6990
6991        eval "require $class";
6992
6993or you could do
6994
6995        require "Foo/Bar.pm";
6996
6997Neither of these forms will autovivify any stashes at compile time and
6998only have run time effects.
6999
7000Now that you understand how L<C<require>|/require VERSION> looks for
7001files with a bareword argument, there is a little extra functionality
7002going on behind the scenes.  Before L<C<require>|/require VERSION> looks
7003for a F<.pm> extension, it will first look for a similar filename with a
7004F<.pmc> extension.  If this file is found, it will be loaded in place of
7005any file ending in a F<.pm> extension. This applies to both the explicit
7006C<require "Foo/Bar.pm";> form and the C<require Foo::Bar;> form.
7007
7008You can also insert hooks into the import facility by putting Perl
7009coderefs or objects directly into the L<C<@INC>|perlvar/@INC> array.
7010There are two types of hooks, INC filters, and INCDIR hooks, and there
7011are three forms of representing a hook: subroutine references, array
7012references, and blessed objects.
7013
7014Subroutine references are the simplest case.  When the inclusion system
7015walks through L<C<@INC>|perlvar/@INC> and encounters a subroutine, unless
7016this subroutine is blessed and supports an INCDIR hook this
7017subroutine will be assumed to be an INC hook will be called with two
7018parameters, the first a reference to itself, and the second the name of
7019the file to be included (e.g., F<Foo/Bar.pm>).  The subroutine should
7020return either nothing or else a list of up to four values in the
7021following order:
7022
7023=over
7024
7025=item 1
7026
7027A reference to a scalar, containing any initial source code to prepend to
7028the file or generator output.
7029
7030=item 2
7031
7032A filehandle, from which the file will be read.
7033
7034=item 3
7035
7036A reference to a subroutine.  If there is no filehandle (previous item),
7037then this subroutine is expected to generate one line of source code per
7038call, writing the line into L<C<$_>|perlvar/$_> and returning 1, then
7039finally at end of file returning 0.  If there is a filehandle, then the
7040subroutine will be called to act as a simple source filter, with the
7041line as read in L<C<$_>|perlvar/$_>.
7042Again, return 1 for each valid line, and 0 after all lines have been
7043returned.
7044For historical reasons the subroutine will receive a meaningless argument
7045(in fact always the numeric value zero) as C<$_[0]>.
7046
7047=item 4
7048
7049Optional state for the subroutine.  The state is passed in as C<$_[1]>.
7050
7051=back
7052
7053C<AUTOLOAD> cannot be used to resolve the C<INCDIR> method, C<INC> is
7054checked first, and C<AUTOLOAD> would resolve that.
7055
7056If an empty list, L<C<undef>|/undef EXPR>, or nothing that matches the
7057first 3 values above is returned, then L<C<require>|/require VERSION>
7058looks at the remaining elements of L<C<@INC>|perlvar/@INC>.
7059Note that this filehandle must be a real filehandle (strictly a typeglob
7060or reference to a typeglob, whether blessed or unblessed); tied filehandles
7061will be ignored and processing will stop there.
7062
7063If the hook is an object, it should provide an C<INC> or C<INCDIR>
7064method that will be called as above, the first parameter being the
7065object itself. If it does not provide either method, and the object is
7066not CODE ref then an exception will be thrown, otherwise it will simply
7067be executed like an unblessed CODE ref would. Note that you must fully
7068qualify the method name when you declare an C<INC> sub (unlike the
7069C<INCDIR> sub), as the unqualified symbol C<INC> is always forced into
7070package C<main>.  Here is a typical code layout for an C<INC> hook:
7071
7072    # In Foo.pm
7073    package Foo;
7074    sub new { ... }
7075    sub Foo::INC {
7076        my ($self, $filename) = @_;
7077        ...
7078    }
7079
7080    # In the main program
7081    push @INC, Foo->new(...);
7082
7083If the hook is an array reference, its first element must be a
7084subroutine reference or an object as described above. When the first
7085element is an object that supports an C<INC> or C<INCDIR> method then
7086the method will be called with the object as the first argument, the
7087filename requested as the second, and the hook array reference as the
7088the third. When the first element is a subroutine then it will be
7089called with the array as the first argument, and the filename as the
7090second, no third parameter will be passed in. In both forms you can
7091modify the contents of the array to provide state between calls, or
7092whatever you like.
7093
7094
7095In other words, you can write:
7096
7097    push @INC, \&my_sub;
7098    sub my_sub {
7099        my ($coderef, $filename) = @_;  # $coderef is \&my_sub
7100        ...
7101    }
7102
7103or:
7104
7105    push @INC, [ \&my_sub, $x, $y, ... ];
7106    sub my_sub {
7107        my ($arrayref, $filename) = @_;
7108        # Retrieve $x, $y, ...
7109        my (undef, @parameters) = @$arrayref;
7110        ...
7111    }
7112
7113or:
7114
7115    push @INC, [ HookObj->new(), $x, $y, ... ];
7116    sub HookObj::INC {
7117        my ($self, $filename, $arrayref)= @_;
7118        my (undef, @parameters) = @$arrayref;
7119        ...
7120    }
7121
7122These hooks are also permitted to set the L<C<%INC>|perlvar/%INC> entry
7123corresponding to the files they have loaded.  See L<perlvar/%INC>.
7124Should an C<INC> hook not do this then perl will set the C<%INC> entry
7125to be the hook reference itself.
7126
7127A hook may also be used to rewrite the C<@INC> array. While this might
7128sound strange, there are situations where it can be very useful to do
7129this. Such hooks usually just return undef and do not mix filtering and
7130C<@INC> modifications. While in older versions of perl having a hook
7131modify C<@INC> was fraught with issues and could even result in
7132segfaults or assert failures, as of 5.37.7 the logic has been made much
7133more robust and the hook now has control over the loop iteration if it
7134wishes to do so.
7135
7136There is a now a facility to control the iterator for the C<@INC> array
7137traversal that is performed during require. The C<$INC> variable will be
7138initialized with the index of the currently executing hook. Once the
7139hook returns the next slot in C<@INC> that will be checked will be the
7140integer successor of value in C<$INC> (or -1 if it is undef). For example
7141the following code
7142
7143    push @INC, sub {
7144        splice @INC, $INC, 1; # remove this hook from @INC
7145        unshift @INC, sub { warn "A" };
7146        undef $INC; # reset the $INC iterator so we
7147                    # execute the newly installed sub
7148                    # immediately.
7149    };
7150
7151would install a sub into C<@INC> that when executed as a hook (by for
7152instance a require of a file that does not exist), the hook will splice
7153itself out of C<@INC>, and add a new sub to the front that will warn
7154whenever someone does a require operation that requires an C<@INC>
7155search, and then immediately execute that hook.
7156
7157Prior to 5.37.7, there was no way to cause perl to use the newly
7158installed hook immediately, or to inspect any changed items in C<@INC> to
7159the left of the iterator, and so the warning would only be generated on
7160the second call to require. In more recent perl the presence of the last
7161statement which undefines C<$INC> will cause perl to restart the
7162traversal of the C<@INC> array at the beginning and execute the newly
7163installed sub immediately.
7164
7165Whatever value C<$INC> held, if any, will be restored at the end of the
7166require. Any changes made to C<$INC> during the lifetime of the hook
7167will be unrolled after the hook exits, and its value only has meaning
7168immediately after execution of the hook, thus setting C<$INC> to some
7169value prior to executing a C<require> will have no effect on how the
7170require executes at all.
7171
7172As of 5.37.7 C<@INC> values of undef will be silently ignored.
7173
7174The function C<require()> is difficult to wrap properly. Many modules
7175consult the stack to find information about their caller, and injecting
7176a new stack frame by wrapping C<require()> often breaks things.
7177Nevertheless it can be very helpful to have the ability to perform
7178actions before and after a C<require>, for instance for trace utilities
7179like C<Devel::TraceUse> or to measure time to load and the memory
7180consumption of the require graph. Because of the difficulties in safely
7181creating a C<require()> wrapper in 5.37.10 we introduced a new mechanism.
7182
7183As of 5.37.10, prior to any other actions it performs, C<require> will
7184check if C<${^HOOK}{require__before}> contains a coderef, and if it does
7185it will be called with the filename form of the item being loaded. The hook
7186may modify C<$_[0]> to load a different filename, or it may throw a fatal
7187exception to cause the require to fail, which will be treated as though the
7188required code itself had thrown an exception.
7189
7190The C<${^HOOK}{require__before}> hook may return a code reference, in
7191which case the code reference will be executed (in an eval with the
7192filname as a parameter) after the require completes. It will be executed
7193regardless of how the compilation completed, and even if the require
7194throws a fatal exception.  The function may consult C<%INC> to determine
7195if the require failed or not.  For instance the following code will print
7196some diagnostics before and after every C<require> statement.  The
7197example also includes logic to chain the signal, so that multiple
7198signals can cooperate. Well behaved C<${^HOOK}{require__before}>
7199handlers should always take this into account.
7200
7201    {
7202        use Scalar::Util qw(reftype);
7203        my $old_hook = ${^HOOK}{require__before};
7204        local ${^HOOK}{require__before} = sub {
7205            my ($name) = @_;
7206            my $old_hook_ret;
7207            $old_hook_ret = $old_hook->($name) if $old_hook;
7208            warn "Requiring: $name\n";
7209            return sub {
7210                $old_hook_ret->() if ref($old_hook_ret)
7211                                  && reftype($old_hook_ret) eq "CODE";
7212                warn sprintf "Finished requiring %s: %s\n",
7213                        $name, $INC{$name} ? "loaded" :"failed";
7214            };
7215        };
7216        require Whatever;
7217    }
7218
7219This hook executes for ALL C<require> statements, unlike C<INC> and
7220C<INCDIR> hooks, which are only executed for relative file names, and it
7221executes first before any other special behaviour inside of require.
7222Note that the initial hook in C<${^HOOK}{require__before}> is *not*
7223executed inside of an eval, and throwing an exception will stop further
7224processing, but the after hook it may return is executed inside of an
7225eval, and any exceptions it throws will be silently ignored.  This is
7226because it executes inside of the scope cleanup logic that is triggered
7227after the require completes, and an exception at this time would not
7228stop the module from being loaded, etc.
7229
7230There is a similar hook that fires after require completes,
7231C<${^HOOK}{require__after}>, which will be called after each require statement
7232completes, either via an exception or successfully. It will be called with
7233the filename of the most recently executed require statement. It is executed
7234in an eval, and will not in any way affect execution.
7235
7236For a yet-more-powerful import facility built around C<require>, see
7237L<C<use>|/use Module VERSION LIST> and L<perlmod>.
7238
7239=item reset EXPR
7240X<reset>
7241
7242=item reset
7243
7244=for Pod::Functions clear all variables of a given name
7245
7246Generally used in a L<C<continue>|/continue BLOCK> block at the end of a
7247loop to clear variables and reset C<m?pattern?> searches so that they
7248work again.  The
7249expression is interpreted as a list of single characters (hyphens
7250allowed for ranges).  All variables (scalars, arrays, and hashes)
7251in the current package beginning with one of
7252those letters are reset to their pristine state.  If the expression is
7253omitted, one-match searches (C<m?pattern?>) are reset to match again.
7254Only resets variables or searches in the current package.  Always returns
72551.  Examples:
7256
7257    reset 'X';      # reset all X variables
7258    reset 'a-z';    # reset lower case variables
7259    reset;          # just reset m?one-time? searches
7260
7261Resetting C<"A-Z"> is not recommended because you'll wipe out your
7262L<C<@ARGV>|perlvar/@ARGV> and L<C<@INC>|perlvar/@INC> arrays and your
7263L<C<%ENV>|perlvar/%ENV> hash.
7264
7265Resets only package variables; lexical variables are unaffected, but
7266they clean themselves up on scope exit anyway, so you'll probably want
7267to use them instead.  See L<C<my>|/my VARLIST>.
7268
7269=item return EXPR
7270X<return>
7271
7272=item return
7273
7274=for Pod::Functions get out of a function early
7275
7276Returns from a subroutine, L<C<eval>|/eval EXPR>,
7277L<C<do FILE>|/do EXPR>, L<C<sort>|/sort SUBNAME LIST> block or regex
7278eval block (but not a L<C<grep>|/grep BLOCK LIST>,
7279L<C<map>|/map BLOCK LIST>, or L<C<do BLOCK>|/do BLOCK> block) with the value
7280given in EXPR.  Evaluation of EXPR may be in list, scalar, or void
7281context, depending on how the return value will be used, and the context
7282may vary from one execution to the next (see
7283L<C<wantarray>|/wantarray>).  If no EXPR
7284is given, returns an empty list in list context, the undefined value in
7285scalar context, and (of course) nothing at all in void context.
7286
7287(In the absence of an explicit L<C<return>|/return EXPR>, a subroutine,
7288L<C<eval>|/eval EXPR>,
7289or L<C<do FILE>|/do EXPR> automatically returns the value of the last expression
7290evaluated.)
7291
7292Unlike most named operators, this is also exempt from the
7293looks-like-a-function rule, so C<return ("foo")."bar"> will
7294cause C<"bar"> to be part of the argument to L<C<return>|/return EXPR>.
7295
7296=item reverse LIST
7297X<reverse> X<rev> X<invert>
7298
7299=for Pod::Functions flip a string or a list
7300
7301In list context, returns a list value consisting of the elements
7302of LIST in the opposite order.  In scalar context, concatenates the
7303elements of LIST and returns a string value with all characters
7304in the opposite order.
7305
7306    print join(", ", reverse "world", "Hello"); # Hello, world
7307
7308    print scalar reverse "dlrow ,", "olleH";    # Hello, world
7309
7310Used without arguments in scalar context, L<C<reverse>|/reverse LIST>
7311reverses L<C<$_>|perlvar/$_>.
7312
7313    $_ = "dlrow ,olleH";
7314    print reverse;                         # No output, list context
7315    print scalar reverse;                  # Hello, world
7316
7317Note that reversing an array to itself (as in C<@a = reverse @a>) will
7318preserve non-existent elements whenever possible; i.e., for non-magical
7319arrays or for tied arrays with C<EXISTS> and C<DELETE> methods.
7320
7321This operator is also handy for inverting a hash, although there are some
7322caveats.  If a value is duplicated in the original hash, only one of those
7323can be represented as a key in the inverted hash.  Also, this has to
7324unwind one hash and build a whole new one, which may take some time
7325on a large hash, such as from a DBM file.
7326
7327    my %by_name = reverse %by_address;  # Invert the hash
7328
7329=item rewinddir DIRHANDLE
7330X<rewinddir>
7331
7332=for Pod::Functions reset directory handle
7333
7334Sets the current position to the beginning of the directory for the
7335L<C<readdir>|/readdir DIRHANDLE> routine on DIRHANDLE.
7336
7337Portability issues: L<perlport/rewinddir>.
7338
7339=item rindex STR,SUBSTR,POSITION
7340X<rindex>
7341
7342=item rindex STR,SUBSTR
7343
7344=for Pod::Functions right-to-left substring search
7345
7346Works just like L<C<index>|/index STR,SUBSTR,POSITION> except that it
7347returns the position of the I<last>
7348occurrence of SUBSTR in STR.  If POSITION is specified, returns the
7349last occurrence beginning at or before that position.
7350
7351=item rmdir FILENAME
7352X<rmdir> X<rd> X<directory, remove>
7353
7354=item rmdir
7355
7356=for Pod::Functions remove a directory
7357
7358Deletes the directory specified by FILENAME if that directory is
7359empty.  If it succeeds it returns true; otherwise it returns false and
7360sets L<C<$!>|perlvar/$!> (errno).  If FILENAME is omitted, uses
7361L<C<$_>|perlvar/$_>.
7362
7363To remove a directory tree recursively (C<rm -rf> on Unix) look at
7364the L<C<rmtree>|File::Path/rmtree( $dir )> function of the L<File::Path>
7365module.
7366
7367=item s///
7368
7369=for Pod::Functions replace a pattern with a string
7370
7371The substitution operator.  See L<perlop/"Regexp Quote-Like Operators">.
7372
7373=item say FILEHANDLE LIST
7374X<say>
7375
7376=item say FILEHANDLE
7377
7378=item say LIST
7379
7380=item say
7381
7382=for Pod::Functions +say output a list to a filehandle, appending a newline
7383
7384Just like L<C<print>|/print FILEHANDLE LIST>, but implicitly appends a
7385newline at the end of the LIST instead of any value L<C<$\>|perlvar/$\>
7386might have.  To use FILEHANDLE without a LIST to
7387print the contents of L<C<$_>|perlvar/$_> to it, you must use a bareword
7388filehandle like C<FH>, not an indirect one like C<$fh>.
7389
7390L<C<say>|/say FILEHANDLE LIST> is available only if the
7391L<C<"say"> feature|feature/The 'say' feature> is enabled or if it is
7392prefixed with C<CORE::>.  The
7393L<C<"say"> feature|feature/The 'say' feature> is enabled automatically
7394with a C<use v5.10> (or higher) declaration in the current scope.
7395
7396=item scalar EXPR
7397X<scalar> X<context>
7398
7399=for Pod::Functions force a scalar context
7400
7401Forces EXPR to be interpreted in scalar context and returns the value
7402of EXPR.
7403
7404    my @counts = ( scalar @a, scalar @b, scalar @c );
7405
7406There is no equivalent operator to force an expression to
7407be interpolated in list context because in practice, this is never
7408needed.  If you really wanted to do so, however, you could use
7409the construction C<@{[ (some expression) ]}>, but usually a simple
7410C<(some expression)> suffices.
7411
7412Because L<C<scalar>|/scalar EXPR> is a unary operator, if you
7413accidentally use a
7414parenthesized list for the EXPR, this behaves as a scalar comma expression,
7415evaluating all but the last element in void context and returning the final
7416element evaluated in scalar context.  This is seldom what you want.
7417
7418The following single statement:
7419
7420    print uc(scalar(foo(), $bar)), $baz;
7421
7422is the moral equivalent of these two:
7423
7424    foo();
7425    print(uc($bar), $baz);
7426
7427See L<perlop> for more details on unary operators and the comma operator,
7428and L<perldata> for details on evaluating a hash in scalar context.
7429
7430=item seek FILEHANDLE,POSITION,WHENCE
7431X<seek> X<fseek> X<filehandle, position>
7432
7433=for Pod::Functions reposition file pointer for random-access I/O
7434
7435Sets FILEHANDLE's position, just like the L<fseek(3)> call of C C<stdio>.
7436FILEHANDLE may be an expression whose value gives the name of the
7437filehandle.  The values for WHENCE are C<0> to set the new position
7438I<in bytes> to POSITION; C<1> to set it to the current position plus
7439POSITION; and C<2> to set it to EOF plus POSITION, typically
7440negative.  For WHENCE you may use the constants C<SEEK_SET>,
7441C<SEEK_CUR>, and C<SEEK_END> (start of the file, current position, end
7442of the file) from the L<Fcntl> module.  Returns C<1> on success, false
7443otherwise.
7444
7445Note the emphasis on bytes: even if the filehandle has been set to operate
7446on characters (for example using the C<:encoding(UTF-8)> I/O layer), the
7447L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
7448L<C<tell>|/tell FILEHANDLE>, and
7449L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>
7450family of functions use byte offsets, not character offsets,
7451because seeking to a character offset would be very slow in a UTF-8 file.
7452
7453If you want to position the file for
7454L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET> or
7455L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET>, don't use
7456L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>, because buffering makes its
7457effect on the file's read-write position unpredictable and non-portable.
7458Use L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE> instead.
7459
7460Due to the rules and rigors of ANSI C, on some systems you have to do a
7461seek whenever you switch between reading and writing.  Amongst other
7462things, this may have the effect of calling stdio's L<clearerr(3)>.
7463A WHENCE of C<1> (C<SEEK_CUR>) is useful for not moving the file position:
7464
7465    seek($fh, 0, 1);
7466
7467This is also useful for applications emulating C<tail -f>.  Once you hit
7468EOF on your read and then sleep for a while, you (probably) have to stick in a
7469dummy L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE> to reset things.  The
7470L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE> doesn't change the position,
7471but it I<does> clear the end-of-file condition on the handle, so that the
7472next C<readline FILE> makes Perl try again to read something.  (We hope.)
7473
7474If that doesn't work (some I/O implementations are particularly
7475cantankerous), you might need something like this:
7476
7477    for (;;) {
7478        for ($curpos = tell($fh); $_ = readline($fh);
7479             $curpos = tell($fh)) {
7480            # search for some stuff and put it into files
7481        }
7482        sleep($for_a_while);
7483        seek($fh, $curpos, 0);
7484    }
7485
7486=item seekdir DIRHANDLE,POS
7487X<seekdir>
7488
7489=for Pod::Functions reposition directory pointer
7490
7491Sets the current position for the L<C<readdir>|/readdir DIRHANDLE>
7492routine on DIRHANDLE.  POS must be a value returned by
7493L<C<telldir>|/telldir DIRHANDLE>.  L<C<seekdir>|/seekdir DIRHANDLE,POS>
7494also has the same caveats about possible directory compaction as the
7495corresponding system library routine.
7496
7497=item select FILEHANDLE
7498X<select> X<filehandle, default>
7499
7500=item select
7501
7502=for Pod::Functions reset default output or do I/O multiplexing
7503
7504Returns the currently selected filehandle.  If FILEHANDLE is supplied,
7505sets the new current default filehandle for output.  This has two
7506effects: first, a L<C<write>|/write FILEHANDLE>,  L<C<print>|/print
7507FILEHANDLE LIST>, or L<C<say>|/say FILEHANDLE LIST> without a
7508filehandle will default to this FILEHANDLE.  Second, references to variables
7509related to output will refer to this output channel.
7510
7511For example, to set the top-of-form format for more than one
7512output channel, you might do the following:
7513
7514    select(REPORT1);
7515    $^ = 'report1_top';
7516    select(REPORT2);
7517    $^ = 'report2_top';
7518
7519FILEHANDLE may be an expression whose value gives the name of the
7520actual filehandle.  Thus:
7521
7522    my $oldfh = select(STDERR); $| = 1; select($oldfh);
7523
7524Some programmers may prefer to think of filehandles as objects with
7525methods, preferring to write the last example as:
7526
7527    STDERR->autoflush(1);
7528
7529(Prior to Perl version 5.14, you have to C<use IO::Handle;> explicitly
7530first.)
7531
7532Whilst you can use C<select> to temporarily "capture" the output of
7533C<print> like this:
7534
7535    {
7536        my $old_handle = select $new_handle;
7537
7538        # This goes to $new_handle:
7539        print "ok 1\n";
7540        ...
7541
7542        select $old_handle;
7543    }
7544
7545you might find it easier to localize the typeglob instead:
7546
7547    {
7548        local *STDOUT = $new_handle;
7549
7550        print "ok 1\n";
7551        ...
7552    }
7553
7554The two are not exactly equivalent, but the latter might be clearer and will
7555restore STDOUT if the wrapped code dies.  The difference is that in the
7556former, the original STDOUT can still be accessed by explicitly using it in a
7557C<print> statement (as C<print STDOUT ...>), whereas in the latter the meaning
7558of the STDOUT handle itself has temporarily been changed.
7559
7560Portability issues: L<perlport/select>.
7561
7562=item select RBITS,WBITS,EBITS,TIMEOUT
7563X<select>
7564
7565This calls the L<select(2)> syscall with the bit masks specified, which
7566can be constructed using L<C<fileno>|/fileno FILEHANDLE> and
7567L<C<vec>|/vec EXPR,OFFSET,BITS>, along these lines:
7568
7569    my $rin = my $win = my $ein = '';
7570    vec($rin, fileno(STDIN),  1) = 1;
7571    vec($win, fileno(STDOUT), 1) = 1;
7572    $ein = $rin | $win;
7573
7574If you want to select on many filehandles, you may wish to write a
7575subroutine like this:
7576
7577    sub fhbits {
7578        my @fhlist = @_;
7579        my $bits = "";
7580        for my $fh (@fhlist) {
7581            vec($bits, fileno($fh), 1) = 1;
7582        }
7583        return $bits;
7584    }
7585    my $rin = fhbits(\*STDIN, $tty, $mysock);
7586
7587The usual idiom is:
7588
7589 my ($nfound, $timeleft) =
7590   select(my $rout = $rin, my $wout = $win, my $eout = $ein,
7591                                                          $timeout);
7592
7593or to block until something becomes ready just do this
7594
7595 my $nfound =
7596   select(my $rout = $rin, my $wout = $win, my $eout = $ein, undef);
7597
7598Most systems do not bother to return anything useful in C<$timeleft>, so
7599calling L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT> in scalar context
7600just returns C<$nfound>.
7601
7602Any of the bit masks can also be L<C<undef>|/undef EXPR>.  The timeout,
7603if specified, is
7604in seconds, which may be fractional.  Note: not all implementations are
7605capable of returning the C<$timeleft>.  If not, they always return
7606C<$timeleft> equal to the supplied C<$timeout>.
7607
7608You can effect a sleep of 250 milliseconds this way:
7609
7610    select(undef, undef, undef, 0.25);
7611
7612Note that whether L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT> gets
7613restarted after signals (say, SIGALRM) is implementation-dependent.  See
7614also L<perlport> for notes on the portability of
7615L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT>.
7616
7617On error, L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT> behaves just
7618like L<select(2)>: it returns C<-1> and sets L<C<$!>|perlvar/$!>.
7619
7620On some Unixes, L<select(2)> may report a socket file descriptor as
7621"ready for reading" even when no data is available, and thus any
7622subsequent L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET> would block.
7623This can be avoided if you always use C<O_NONBLOCK> on the socket.  See
7624L<select(2)> and L<fcntl(2)> for further details.
7625
7626The standard L<C<IO::Select>|IO::Select> module provides a
7627user-friendlier interface to
7628L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT>, mostly because it does
7629all the bit-mask work for you.
7630
7631B<WARNING>: One should not attempt to mix buffered I/O (like
7632L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET> or
7633L<C<readline>|/readline EXPR>) with
7634L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT>, except as permitted by
7635POSIX, and even then only on POSIX systems.  You have to use
7636L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET> instead.
7637
7638Portability issues: L<perlport/select>.
7639
7640=item semctl ID,SEMNUM,CMD,ARG
7641X<semctl>
7642
7643=for Pod::Functions SysV semaphore control operations
7644
7645Calls the System V IPC function L<semctl(2)>.  You'll probably have to say
7646
7647    use IPC::SysV;
7648
7649first to get the correct constant definitions.  If CMD is IPC_STAT or
7650GETALL, then ARG must be a variable that will hold the returned
7651semid_ds structure or semaphore value array.  Returns like
7652L<C<ioctl>|/ioctl FILEHANDLE,FUNCTION,SCALAR>:
7653the undefined value for error, "C<0 but true>" for zero, or the actual
7654return value otherwise.  The ARG must consist of a vector of native
7655short integers, which may be created with C<pack("s!",(0)x$nsem)>.
7656See also L<perlipc/"SysV IPC"> and the documentation for
7657L<C<IPC::SysV>|IPC::SysV> and L<C<IPC::Semaphore>|IPC::Semaphore>.
7658
7659Portability issues: L<perlport/semctl>.
7660
7661=item semget KEY,NSEMS,FLAGS
7662X<semget>
7663
7664=for Pod::Functions get set of SysV semaphores
7665
7666Calls the System V IPC function L<semget(2)>.  Returns the semaphore id, or
7667the undefined value on error.  See also
7668L<perlipc/"SysV IPC"> and the documentation for
7669L<C<IPC::SysV>|IPC::SysV> and L<C<IPC::Semaphore>|IPC::Semaphore>.
7670
7671Portability issues: L<perlport/semget>.
7672
7673=item semop KEY,OPSTRING
7674X<semop>
7675
7676=for Pod::Functions SysV semaphore operations
7677
7678Calls the System V IPC function L<semop(2)> for semaphore operations
7679such as signalling and waiting.  OPSTRING must be a packed array of
7680semop structures.  Each semop structure can be generated with
7681C<pack("s!3", $semnum, $semop, $semflag)>.  The length of OPSTRING
7682implies the number of semaphore operations.  Returns true if
7683successful, false on error.  As an example, the
7684following code waits on semaphore $semnum of semaphore id $semid:
7685
7686    my $semop = pack("s!3", $semnum, -1, 0);
7687    die "Semaphore trouble: $!\n" unless semop($semid, $semop);
7688
7689To signal the semaphore, replace C<-1> with C<1>.  See also
7690L<perlipc/"SysV IPC"> and the documentation for
7691L<C<IPC::SysV>|IPC::SysV> and L<C<IPC::Semaphore>|IPC::Semaphore>.
7692
7693Portability issues: L<perlport/semop>.
7694
7695=item send SOCKET,MSG,FLAGS,TO
7696X<send>
7697
7698=item send SOCKET,MSG,FLAGS
7699
7700=for Pod::Functions send a message over a socket
7701
7702Sends a message on a socket.  Attempts to send the scalar MSG to the SOCKET
7703filehandle.  Takes the same flags as the system call of the same name.  On
7704unconnected sockets, you must specify a destination to I<send to>, in which
7705case it does a L<sendto(2)> syscall.  Returns the number of characters sent,
7706or the undefined value on error.  The L<sendmsg(2)> syscall is currently
7707unimplemented.  See L<perlipc/"UDP: Message Passing"> for examples.
7708
7709Note that if the socket has been marked as C<:utf8>, C<send> will
7710throw an exception.  The C<:encoding(...)> layer implicitly introduces
7711the C<:utf8> layer.  See L<C<binmode>|/binmode FILEHANDLE, LAYER>.
7712
7713=item setpgrp PID,PGRP
7714X<setpgrp> X<group>
7715
7716=for Pod::Functions set the process group of a process
7717
7718Sets the current process group for the specified PID, C<0> for the current
7719process.  Raises an exception when used on a machine that doesn't
7720implement POSIX L<setpgid(2)> or BSD L<setpgrp(2)>.  If the arguments
7721are omitted, it defaults to C<0,0>.  Note that the BSD 4.2 version of
7722L<C<setpgrp>|/setpgrp PID,PGRP> does not accept any arguments, so only
7723C<setpgrp(0,0)> is portable.  See also
7724L<C<POSIX::setsid()>|POSIX/C<setsid>>.
7725
7726Portability issues: L<perlport/setpgrp>.
7727
7728=item setpriority WHICH,WHO,PRIORITY
7729X<setpriority> X<priority> X<nice> X<renice>
7730
7731=for Pod::Functions set a process's nice value
7732
7733Sets the current priority for a process, a process group, or a user.
7734(See L<setpriority(2)>.)  Raises an exception when used on a machine
7735that doesn't implement L<setpriority(2)>.
7736
7737C<WHICH> can be any of C<PRIO_PROCESS>, C<PRIO_PGRP> or C<PRIO_USER>
7738imported from L<POSIX/RESOURCE CONSTANTS>.
7739
7740Portability issues: L<perlport/setpriority>.
7741
7742=item setsockopt SOCKET,LEVEL,OPTNAME,OPTVAL
7743X<setsockopt>
7744
7745=for Pod::Functions set some socket options
7746
7747Sets the socket option requested.  Returns L<C<undef>|/undef EXPR> on
7748error.  Use integer constants provided by the L<C<Socket>|Socket> module
7749for
7750LEVEL and OPNAME.  Values for LEVEL can also be obtained from
7751getprotobyname.  OPTVAL might either be a packed string or an integer.
7752An integer OPTVAL is shorthand for pack("i", OPTVAL).
7753
7754An example disabling Nagle's algorithm on a socket:
7755
7756    use Socket qw(IPPROTO_TCP TCP_NODELAY);
7757    setsockopt($socket, IPPROTO_TCP, TCP_NODELAY, 1);
7758
7759Portability issues: L<perlport/setsockopt>.
7760
7761=item shift ARRAY
7762X<shift>
7763
7764=item shift
7765
7766=for Pod::Functions remove the first element of an array, and return it
7767
7768Removes and returns the B<first> element of an array. This shortens the
7769array by one and moves everything down.
7770
7771    my @arr  = ('cat', 'dog');
7772    my $item = shift(@arr); # 'cat'
7773
7774    # @arr is now ('dog');
7775
7776Returns C<undef> if the array is empty.
7777
7778B<Note:> C<shift> may also return C<undef> if the first element in the array
7779is C<undef>.
7780
7781    my @arr  = (undef, 'two', 'three');
7782    my $item = shift(@arr); # undef
7783
7784If ARRAY is omitted, C<shift> operates on the C<@ARGV> array in the main
7785program, and the C<@_> array in subroutines. C<shift> will operate on the
7786C<@ARGV> array in C<eval STRING>, C<BEGIN {}>, C<INIT {}>, C<CHECK {}> blocks.
7787
7788Starting with Perl 5.14, an experimental feature allowed
7789L<C<shift>|/shift ARRAY> to take a
7790scalar expression. This experiment has been deemed unsuccessful, and was
7791removed as of Perl 5.24.
7792
7793See also L<C<unshift>|/unshift ARRAY,LIST>, L<C<push>|/push ARRAY,LIST>,
7794and L<C<pop>|/pop ARRAY>.  L<C<shift>|/shift ARRAY> and
7795L<C<unshift>|/unshift ARRAY,LIST> do the same thing to the left end of
7796an array that L<C<pop>|/pop ARRAY> and L<C<push>|/push ARRAY,LIST> do to
7797the right end.
7798
7799=item shmctl ID,CMD,ARG
7800X<shmctl>
7801
7802=for Pod::Functions SysV shared memory operations
7803
7804Calls the System V IPC function shmctl.  You'll probably have to say
7805
7806    use IPC::SysV;
7807
7808first to get the correct constant definitions.  If CMD is C<IPC_STAT>,
7809then ARG must be a variable that will hold the returned C<shmid_ds>
7810structure.  Returns like ioctl: L<C<undef>|/undef EXPR> for error; "C<0>
7811but true" for zero; and the actual return value otherwise.
7812See also L<perlipc/"SysV IPC"> and the documentation for
7813L<C<IPC::SysV>|IPC::SysV>.
7814
7815Portability issues: L<perlport/shmctl>.
7816
7817=item shmget KEY,SIZE,FLAGS
7818X<shmget>
7819
7820=for Pod::Functions get SysV shared memory segment identifier
7821
7822Calls the System V IPC function shmget.  Returns the shared memory
7823segment id, or L<C<undef>|/undef EXPR> on error.
7824See also L<perlipc/"SysV IPC"> and the documentation for
7825L<C<IPC::SysV>|IPC::SysV>.
7826
7827Portability issues: L<perlport/shmget>.
7828
7829=item shmread ID,VAR,POS,SIZE
7830X<shmread>
7831X<shmwrite>
7832
7833=for Pod::Functions read SysV shared memory
7834
7835=item shmwrite ID,STRING,POS,SIZE
7836
7837=for Pod::Functions write SysV shared memory
7838
7839Reads or writes the System V shared memory segment ID starting at
7840position POS for size SIZE by attaching to it, copying in/out, and
7841detaching from it.  When reading, VAR must be a variable that will
7842hold the data read.  When writing, if STRING is too long, only SIZE
7843bytes are used; if STRING is too short, nulls are written to fill out
7844SIZE bytes.  Return true if successful, false on error.
7845L<C<shmread>|/shmread ID,VAR,POS,SIZE> taints the variable.  See also
7846L<perlipc/"SysV IPC"> and the documentation for
7847L<C<IPC::SysV>|IPC::SysV> and the L<C<IPC::Shareable>|IPC::Shareable>
7848module from CPAN.
7849
7850Portability issues: L<perlport/shmread> and L<perlport/shmwrite>.
7851
7852=item shutdown SOCKET,HOW
7853X<shutdown>
7854
7855=for Pod::Functions close down just half of a socket connection
7856
7857Shuts down a socket connection in the manner indicated by HOW, which
7858has the same interpretation as in the syscall of the same name.
7859
7860    shutdown($socket, 0);    # I/we have stopped reading data
7861    shutdown($socket, 1);    # I/we have stopped writing data
7862    shutdown($socket, 2);    # I/we have stopped using this socket
7863
7864This is useful with sockets when you want to tell the other
7865side you're done writing but not done reading, or vice versa.
7866It's also a more insistent form of close because it also
7867disables the file descriptor in any forked copies in other
7868processes.
7869
7870Returns C<1> for success; on error, returns L<C<undef>|/undef EXPR> if
7871the first argument is not a valid filehandle, or returns C<0> and sets
7872L<C<$!>|perlvar/$!> for any other failure.
7873
7874=item sin EXPR
7875X<sin> X<sine> X<asin> X<arcsine>
7876
7877=item sin
7878
7879=for Pod::Functions return the sine of a number
7880
7881Returns the sine of EXPR (expressed in radians).  If EXPR is omitted,
7882returns sine of L<C<$_>|perlvar/$_>.
7883
7884For the inverse sine operation, you may use the C<Math::Trig::asin>
7885function, or use this relation:
7886
7887    sub asin { atan2($_[0], sqrt(1 - $_[0] * $_[0])) }
7888
7889=item sleep EXPR
7890X<sleep> X<pause>
7891
7892=item sleep
7893
7894=for Pod::Functions block for some number of seconds
7895
7896Causes the script to sleep for (integer) EXPR seconds, or forever if no
7897argument is given.  Returns the integer number of seconds actually slept.
7898
7899EXPR should be a positive integer. If called with a negative integer,
7900L<C<sleep>|/sleep EXPR> does not sleep but instead emits a warning, sets
7901C<$!> (C<errno>), and returns zero.
7902
7903If called with a non-integer, the fractional part is ignored.
7904
7905
7906C<sleep 0> is permitted, but a function call to the underlying platform
7907implementation still occurs, with any side effects that may have.
7908C<sleep 0> is therefore not exactly identical to not sleeping at all.
7909
7910May be interrupted if the process receives a signal such as C<SIGALRM>.
7911
7912    eval {
7913        local $SIG{ALRM} = sub { die "Alarm!\n" };
7914        sleep;
7915    };
7916    die $@ unless $@ eq "Alarm!\n";
7917
7918You probably cannot mix L<C<alarm>|/alarm SECONDS> and
7919L<C<sleep>|/sleep EXPR> calls, because L<C<sleep>|/sleep EXPR> is often
7920implemented using L<C<alarm>|/alarm SECONDS>.
7921
7922On some older systems, it may sleep up to a full second less than what
7923you requested, depending on how it counts seconds.  Most modern systems
7924always sleep the full amount.  They may appear to sleep longer than that,
7925however, because your process might not be scheduled right away in a
7926busy multitasking system.
7927
7928For delays of finer granularity than one second, the L<Time::HiRes>
7929module (from CPAN, and starting from Perl 5.8 part of the standard
7930distribution) provides L<C<usleep>|Time::HiRes/usleep ( $useconds )>.
7931You may also use Perl's four-argument
7932version of L<C<select>|/select RBITS,WBITS,EBITS,TIMEOUT> leaving the
7933first three arguments undefined, or you might be able to use the
7934L<C<syscall>|/syscall NUMBER, LIST> interface to access L<setitimer(2)>
7935if your system supports it.  See L<perlfaq8> for details.
7936
7937See also the L<POSIX> module's L<C<pause>|POSIX/C<pause>> function.
7938
7939=item socket SOCKET,DOMAIN,TYPE,PROTOCOL
7940X<socket>
7941
7942=for Pod::Functions create a socket
7943
7944Opens a socket of the specified kind and attaches it to filehandle
7945SOCKET.  DOMAIN, TYPE, and PROTOCOL are specified the same as for
7946the syscall of the same name.  You should C<use Socket> first
7947to get the proper definitions imported.  See the examples in
7948L<perlipc/"Sockets: Client/Server Communication">.
7949
7950On systems that support a close-on-exec flag on files, the flag will
7951be set for the newly opened file descriptor, as determined by the
7952value of L<C<$^F>|perlvar/$^F>.  See L<perlvar/$^F>.
7953
7954=item socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL
7955X<socketpair>
7956
7957=for Pod::Functions create a pair of sockets
7958
7959Creates an unnamed pair of sockets in the specified domain, of the
7960specified type.  DOMAIN, TYPE, and PROTOCOL are specified the same as
7961for the syscall of the same name.  If unimplemented, raises an exception.
7962Returns true if successful.
7963
7964On systems that support a close-on-exec flag on files, the flag will
7965be set for the newly opened file descriptors, as determined by the value
7966of L<C<$^F>|perlvar/$^F>.  See L<perlvar/$^F>.
7967
7968Some systems define L<C<pipe>|/pipe READHANDLE,WRITEHANDLE> in terms of
7969L<C<socketpair>|/socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL>, in
7970which a call to C<pipe($rdr, $wtr)> is essentially:
7971
7972    use Socket;
7973    socketpair(my $rdr, my $wtr, AF_UNIX, SOCK_STREAM, PF_UNSPEC);
7974    shutdown($rdr, 1);        # no more writing for reader
7975    shutdown($wtr, 0);        # no more reading for writer
7976
7977See L<perlipc> for an example of socketpair use.  Perl 5.8 and later will
7978emulate socketpair using IP sockets to localhost if your system implements
7979sockets but not socketpair.
7980
7981Portability issues: L<perlport/socketpair>.
7982
7983=item sort SUBNAME LIST
7984X<sort>
7985
7986=item sort BLOCK LIST
7987
7988=item sort LIST
7989
7990=for Pod::Functions sort a list of values
7991
7992In list context, this sorts the LIST and returns the sorted list value.
7993In scalar context, the behaviour of L<C<sort>|/sort SUBNAME LIST> is
7994undefined.
7995
7996If SUBNAME or BLOCK is omitted, L<C<sort>|/sort SUBNAME LIST>s in
7997standard string comparison
7998order.  If SUBNAME is specified, it gives the name of a subroutine
7999that returns a numeric value less than, equal to, or greater than C<0>,
8000depending on how the elements of the list are to be ordered.  (The
8001C<< <=> >> and C<cmp> operators are extremely useful in such routines.)
8002SUBNAME may be a scalar variable name (unsubscripted), in which case
8003the value provides the name of (or a reference to) the actual
8004subroutine to use.  In place of a SUBNAME, you can provide a BLOCK as
8005an anonymous, in-line sort subroutine.
8006
8007If the subroutine's prototype is C<($$)>, the elements to be compared are
8008passed by reference in L<C<@_>|perlvar/@_>, as for a normal subroutine.
8009This is slower than unprototyped subroutines, where the elements to be
8010compared are passed into the subroutine as the package global variables
8011C<$a> and C<$b> (see example below).
8012
8013If the subroutine is an XSUB, the elements to be compared are pushed on
8014to the stack, the way arguments are usually passed to XSUBs.  C<$a> and
8015C<$b> are not set.
8016
8017The values to be compared are always passed by reference and should not
8018be modified.
8019
8020You also cannot exit out of the sort block or subroutine using any of the
8021loop control operators described in L<perlsyn> or with
8022L<C<goto>|/goto LABEL>.
8023
8024When L<C<use locale>|locale> (but not C<use locale ':not_characters'>)
8025is in effect, C<sort LIST> sorts LIST according to the
8026current collation locale.  See L<perllocale>.
8027
8028L<C<sort>|/sort SUBNAME LIST> returns aliases into the original list,
8029much as a for loop's index variable aliases the list elements.  That is,
8030modifying an element of a list returned by L<C<sort>|/sort SUBNAME LIST>
8031(for example, in a C<foreach>, L<C<map>|/map BLOCK LIST> or
8032L<C<grep>|/grep BLOCK LIST>)
8033actually modifies the element in the original list.  This is usually
8034something to be avoided when writing clear code.
8035
8036Historically Perl has varied in whether sorting is stable by default.
8037If stability matters, it can be controlled explicitly by using the
8038L<sort> pragma.
8039
8040Examples:
8041
8042    # sort lexically
8043    my @articles = sort @files;
8044
8045    # same thing, but with explicit sort routine
8046    my @articles = sort {$a cmp $b} @files;
8047
8048    # now case-insensitively
8049    my @articles = sort {fc($a) cmp fc($b)} @files;
8050
8051    # same thing in reversed order
8052    my @articles = sort {$b cmp $a} @files;
8053
8054    # sort numerically ascending
8055    my @articles = sort {$a <=> $b} @files;
8056
8057    # sort numerically descending
8058    my @articles = sort {$b <=> $a} @files;
8059
8060    # this sorts the %age hash by value instead of key
8061    # using an in-line function
8062    my @eldest = sort { $age{$b} <=> $age{$a} } keys %age;
8063
8064    # sort using explicit subroutine name
8065    sub byage {
8066        $age{$a} <=> $age{$b};  # presuming numeric
8067    }
8068    my @sortedclass = sort byage @class;
8069
8070    sub backwards { $b cmp $a }
8071    my @harry  = qw(dog cat x Cain Abel);
8072    my @george = qw(gone chased yz Punished Axed);
8073    print sort @harry;
8074        # prints AbelCaincatdogx
8075    print sort backwards @harry;
8076        # prints xdogcatCainAbel
8077    print sort @george, 'to', @harry;
8078        # prints AbelAxedCainPunishedcatchaseddoggonetoxyz
8079
8080    # inefficiently sort by descending numeric compare using
8081    # the first integer after the first = sign, or the
8082    # whole record case-insensitively otherwise
8083
8084    my @new = sort {
8085        ($b =~ /=(\d+)/)[0] <=> ($a =~ /=(\d+)/)[0]
8086                            ||
8087                    fc($a)  cmp  fc($b)
8088    } @old;
8089
8090    # same thing, but much more efficiently;
8091    # we'll build auxiliary indices instead
8092    # for speed
8093    my (@nums, @caps);
8094    for (@old) {
8095        push @nums, ( /=(\d+)/ ? $1 : undef );
8096        push @caps, fc($_);
8097    }
8098
8099    my @new = @old[ sort {
8100                           $nums[$b] <=> $nums[$a]
8101                                    ||
8102                           $caps[$a] cmp $caps[$b]
8103                         } 0..$#old
8104                  ];
8105
8106    # same thing, but without any temps
8107    my @new = map { $_->[0] }
8108           sort { $b->[1] <=> $a->[1]
8109                           ||
8110                  $a->[2] cmp $b->[2]
8111           } map { [$_, /=(\d+)/, fc($_)] } @old;
8112
8113    # using a prototype allows you to use any comparison subroutine
8114    # as a sort subroutine (including other package's subroutines)
8115    package Other;
8116    sub backwards ($$) { $_[1] cmp $_[0]; }  # $a and $b are
8117                                             # not set here
8118    package main;
8119    my @new = sort Other::backwards @old;
8120
8121    ## using a prototype with function signature
8122    use feature 'signatures';
8123    sub function_with_signature :prototype($$) ($one, $two) {
8124        return $one <=> $two
8125    }
8126
8127    my @new = sort function_with_signature @old;
8128
8129    # guarantee stability
8130    use sort 'stable';
8131    my @new = sort { substr($a, 3, 5) cmp substr($b, 3, 5) } @old;
8132
8133Warning: syntactical care is required when sorting the list returned from
8134a function.  If you want to sort the list returned by the function call
8135C<find_records(@key)>, you can use:
8136
8137    my @contact = sort { $a cmp $b } find_records @key;
8138    my @contact = sort +find_records(@key);
8139    my @contact = sort &find_records(@key);
8140    my @contact = sort(find_records(@key));
8141
8142If instead you want to sort the array C<@key> with the comparison routine
8143C<find_records()> then you can use:
8144
8145    my @contact = sort { find_records() } @key;
8146    my @contact = sort find_records(@key);
8147    my @contact = sort(find_records @key);
8148    my @contact = sort(find_records (@key));
8149
8150C<$a> and C<$b> are set as package globals in the package the sort() is
8151called from.  That means C<$main::a> and C<$main::b> (or C<$::a> and
8152C<$::b>) in the C<main> package, C<$FooPack::a> and C<$FooPack::b> in the
8153C<FooPack> package, etc.  If the sort block is in scope of a C<my> or
8154C<state> declaration of C<$a> and/or C<$b>, you I<must> spell out the full
8155name of the variables in the sort block :
8156
8157   package main;
8158   my $a = "C"; # DANGER, Will Robinson, DANGER !!!
8159
8160   print sort { $a cmp $b }               qw(A C E G B D F H);
8161                                          # WRONG
8162   sub badlexi { $a cmp $b }
8163   print sort badlexi                     qw(A C E G B D F H);
8164                                          # WRONG
8165   # the above prints BACFEDGH or some other incorrect ordering
8166
8167   print sort { $::a cmp $::b }           qw(A C E G B D F H);
8168                                          # OK
8169   print sort { our $a cmp our $b }       qw(A C E G B D F H);
8170                                          # also OK
8171   print sort { our ($a, $b); $a cmp $b } qw(A C E G B D F H);
8172                                          # also OK
8173   sub lexi { our $a cmp our $b }
8174   print sort lexi                        qw(A C E G B D F H);
8175                                          # also OK
8176   # the above print ABCDEFGH
8177
8178With proper care you may mix package and my (or state) C<$a> and/or C<$b>:
8179
8180   my $a = {
8181      tiny   => -2,
8182      small  => -1,
8183      normal => 0,
8184      big    => 1,
8185      huge   => 2
8186   };
8187
8188   say sort { $a->{our $a} <=> $a->{our $b} }
8189       qw{ huge normal tiny small big};
8190
8191   # prints tinysmallnormalbighuge
8192
8193C<$a> and C<$b> are implicitly local to the sort() execution and regain their
8194former values upon completing the sort.
8195
8196Sort subroutines written using C<$a> and C<$b> are bound to their calling
8197package. It is possible, but of limited interest, to define them in a
8198different package, since the subroutine must still refer to the calling
8199package's C<$a> and C<$b> :
8200
8201   package Foo;
8202   sub lexi { $Bar::a cmp $Bar::b }
8203   package Bar;
8204   ... sort Foo::lexi ...
8205
8206Use the prototyped versions (see above) for a more generic alternative.
8207
8208The comparison function is required to behave.  If it returns
8209inconsistent results (sometimes saying C<$x[1]> is less than C<$x[2]> and
8210sometimes saying the opposite, for example) the results are not
8211well-defined.
8212
8213Because C<< <=> >> returns L<C<undef>|/undef EXPR> when either operand
8214is C<NaN> (not-a-number), be careful when sorting with a
8215comparison function like C<< $a <=> $b >> any lists that might contain a
8216C<NaN>.  The following example takes advantage that C<NaN != NaN> to
8217eliminate any C<NaN>s from the input list.
8218
8219    my @result = sort { $a <=> $b } grep { $_ == $_ } @input;
8220
8221In this version of F<perl>, the C<sort> function is implemented via the
8222mergesort algorithm.
8223
8224=item splice ARRAY,OFFSET,LENGTH,LIST
8225X<splice>
8226
8227=item splice ARRAY,OFFSET,LENGTH
8228
8229=item splice ARRAY,OFFSET
8230
8231=item splice ARRAY
8232
8233=for Pod::Functions add or remove elements anywhere in an array
8234
8235Removes the elements designated by OFFSET and LENGTH from an array, and
8236replaces them with the elements of LIST, if any.  In list context,
8237returns the elements removed from the array.  In scalar context,
8238returns the last element removed, or L<C<undef>|/undef EXPR> if no
8239elements are
8240removed.  The array grows or shrinks as necessary.
8241If OFFSET is negative then it starts that far from the end of the array.
8242If LENGTH is omitted, removes everything from OFFSET onward.
8243If LENGTH is negative, removes the elements from OFFSET onward
8244except for -LENGTH elements at the end of the array.
8245If both OFFSET and LENGTH are omitted, removes everything.  If OFFSET is
8246past the end of the array and a LENGTH was provided, Perl issues a warning,
8247and splices at the end of the array.
8248
8249The following equivalences hold (assuming C<< $#a >= $i >> )
8250
8251    push(@a,$x,$y)      splice(@a,@a,0,$x,$y)
8252    pop(@a)             splice(@a,-1)
8253    shift(@a)           splice(@a,0,1)
8254    unshift(@a,$x,$y)   splice(@a,0,0,$x,$y)
8255    $a[$i] = $y         splice(@a,$i,1,$y)
8256
8257L<C<splice>|/splice ARRAY,OFFSET,LENGTH,LIST> can be used, for example,
8258to implement n-ary queue processing:
8259
8260    sub nary_print {
8261      my $n = shift;
8262      while (my @next_n = splice @_, 0, $n) {
8263        say join q{ -- }, @next_n;
8264      }
8265    }
8266
8267    nary_print(3, qw(a b c d e f g h));
8268    # prints:
8269    #   a -- b -- c
8270    #   d -- e -- f
8271    #   g -- h
8272
8273Starting with Perl 5.14, an experimental feature allowed
8274L<C<splice>|/splice ARRAY,OFFSET,LENGTH,LIST> to take a
8275scalar expression. This experiment has been deemed unsuccessful, and was
8276removed as of Perl 5.24.
8277
8278=item split /PATTERN/,EXPR,LIMIT
8279X<split>
8280
8281=item split /PATTERN/,EXPR
8282
8283=item split /PATTERN/
8284
8285=item split
8286
8287=for Pod::Functions split up a string using a regexp delimiter
8288
8289Splits the string EXPR into a list of strings and returns the
8290list in list context, or the size of the list in scalar context.
8291(Prior to Perl 5.11, it also overwrote C<@_> with the list in
8292void and scalar context. If you target old perls, beware.)
8293
8294If only PATTERN is given, EXPR defaults to L<C<$_>|perlvar/$_>.
8295
8296Anything in EXPR that matches PATTERN is taken to be a separator
8297that separates the EXPR into substrings (called "I<fields>") that
8298do B<not> include the separator.  Note that a separator may be
8299longer than one character or even have no characters at all (the
8300empty string, which is a zero-width match).
8301
8302The PATTERN need not be constant; an expression may be used
8303to specify a pattern that varies at runtime.
8304
8305If PATTERN matches the empty string, the EXPR is split at the match
8306position (between characters).  As an example, the following:
8307
8308    my @x = split(/b/, "abc"); # ("a", "c")
8309
8310uses the C<b> in C<'abc'> as a separator to produce the list ("a", "c").
8311However, this:
8312
8313    my @x = split(//, "abc"); # ("a", "b", "c")
8314
8315uses empty string matches as separators; thus, the empty string
8316may be used to split EXPR into a list of its component characters.
8317
8318As a special case for L<C<split>|/split E<sol>PATTERNE<sol>,EXPR,LIMIT>,
8319the empty pattern given in
8320L<match operator|perlop/"m/PATTERN/msixpodualngc"> syntax (C<//>)
8321specifically matches the empty string, which is contrary to its usual
8322interpretation as the last successful match.
8323
8324If PATTERN is C</^/>, then it is treated as if it used the
8325L<multiline modifier|perlreref/OPERATORS> (C</^/m>), since it
8326isn't much use otherwise.
8327
8328C<E<sol>m> and any of the other pattern modifiers valid for C<qr>
8329(summarized in L<perlop/qrE<sol>STRINGE<sol>msixpodualn>) may be
8330specified explicitly.
8331
8332As another special case,
8333L<C<split>|/split E<sol>PATTERNE<sol>,EXPR,LIMIT> emulates the default
8334behavior of the
8335command line tool B<awk> when the PATTERN is either omitted or a
8336string composed of a single space character (such as S<C<' '>> or
8337S<C<"\x20">>, but not e.g. S<C</ />>).  In this case, any leading
8338whitespace in EXPR is removed before splitting occurs, and the PATTERN is
8339instead treated as if it were C</\s+/>; in particular, this means that
8340I<any> contiguous whitespace (not just a single space character) is used as
8341a separator.
8342
8343    my @x = split(" ", "  Quick brown fox\n");
8344    # ("Quick", "brown", "fox")
8345
8346    my @x = split(" ", "RED\tGREEN\tBLUE");
8347    # ("RED", "GREEN", "BLUE")
8348
8349Using split in this fashion is very similar to how
8350L<C<qwE<sol>E<sol>>|/qwE<sol>STRINGE<sol>> works.
8351
8352However, this special treatment can be avoided by specifying
8353the pattern S<C</ />> instead of the string S<C<" ">>, thereby allowing
8354only a single space character to be a separator.  In earlier Perls this
8355special case was restricted to the use of a plain S<C<" ">> as the
8356pattern argument to split; in Perl 5.18.0 and later this special case is
8357triggered by any expression which evaluates to the simple string S<C<" ">>.
8358
8359As of Perl 5.28, this special-cased whitespace splitting works as expected in
8360the scope of L<< S<C<"use feature 'unicode_strings'">>|feature/The
8361'unicode_strings' feature >>. In previous versions, and outside the scope of
8362that feature, it exhibits L<perlunicode/The "Unicode Bug">: characters that are
8363whitespace according to Unicode rules but not according to ASCII rules can be
8364treated as part of fields rather than as field separators, depending on the
8365string's internal encoding.
8366
8367As of Perl 5.39.9 the C</x> default modifier does NOT affect
8368C<split STRING> but does affect C<split PATTERN>, this means that
8369C<split " "> will produces the expected I<awk> emulation regardless as
8370to whether it is used in the scope of a C<use re "/x"> statement. If you
8371want to split by spaces under C<use re "/x"> you must do something like
8372C<split /(?-x: )/> or C<split /\x{20}/> instead of C<split / />.
8373
8374If omitted, PATTERN defaults to a single space, S<C<" ">>, triggering
8375the previously described I<awk> emulation.
8376
8377If LIMIT is specified and positive, it represents the maximum number
8378of fields into which the EXPR may be split; in other words, LIMIT is
8379one greater than the maximum number of times EXPR may be split.  Thus,
8380the LIMIT value C<1> means that EXPR may be split a maximum of zero
8381times, producing a maximum of one field (namely, the entire value of
8382EXPR).  For instance:
8383
8384    my @x = split(/,/, "a,b,c", 1); # ("a,b,c")
8385    my @x = split(/,/, "a,b,c", 2); # ("a", "b,c")
8386    my @x = split(/,/, "a,b,c", 3); # ("a", "b", "c")
8387    my @x = split(/,/, "a,b,c", 4); # ("a", "b", "c")
8388
8389If LIMIT is negative, it is treated as if it were instead arbitrarily
8390large; as many fields as possible are produced.
8391
8392If LIMIT is omitted (or, equivalently, zero), then it is usually
8393treated as if it were instead negative but with the exception that
8394trailing empty fields are stripped (empty leading fields are always
8395preserved); if all fields are empty, then all fields are considered to
8396be trailing (and are thus stripped in this case).  Thus, the following:
8397
8398    my @x = split(/,/, "a,b,c,,,"); # ("a", "b", "c")
8399
8400produces only a three element list.
8401
8402    my @x = split(/,/, "a,b,c,,,", -1); # ("a", "b", "c", "", "", "")
8403
8404produces a six element list.
8405
8406In time-critical applications, it is worthwhile to avoid splitting
8407into more fields than necessary.  Thus, when assigning to a list,
8408if LIMIT is omitted (or zero), then LIMIT is treated as though it
8409were one larger than the number of variables in the list; for the
8410following, LIMIT is implicitly 3:
8411
8412    my ($login, $passwd) = split(/:/);
8413
8414Note that splitting an EXPR that evaluates to the empty string always
8415produces zero fields, regardless of the LIMIT specified.
8416
8417An empty leading field is produced when there is a positive-width
8418match at the beginning of EXPR.  For instance:
8419
8420    my @x = split(/ /, " abc"); # ("", "abc")
8421
8422splits into two elements.  However, a zero-width match at the
8423beginning of EXPR never produces an empty field, so that:
8424
8425    my @x = split(//, " abc"); # (" ", "a", "b", "c")
8426
8427splits into four elements instead of five.
8428
8429An empty trailing field, on the other hand, is produced when there is a
8430match at the end of EXPR, regardless of the length of the match
8431(of course, unless a non-zero LIMIT is given explicitly, such fields are
8432removed, as in the last example).  Thus:
8433
8434    my @x = split(//, " abc", -1); # (" ", "a", "b", "c", "")
8435
8436If the PATTERN contains
8437L<capturing groups|perlretut/Grouping things and hierarchical matching>,
8438then for each separator, an additional field is produced for each substring
8439captured by a group (in the order in which the groups are specified,
8440as per L<backreferences|perlretut/Backreferences>); if any group does not
8441match, then it captures the L<C<undef>|/undef EXPR> value instead of a
8442substring.  Also,
8443note that any such additional field is produced whenever there is a
8444separator (that is, whenever a split occurs), and such an additional field
8445does B<not> count towards the LIMIT.  Consider the following expressions
8446evaluated in list context (each returned list is provided in the associated
8447comment):
8448
8449    my @x = split(/-|,/    , "1-10,20", 3);
8450    # ("1", "10", "20")
8451
8452    my @x = split(/(-|,)/  , "1-10,20", 3);
8453    # ("1", "-", "10", ",", "20")
8454
8455    my @x = split(/-|(,)/  , "1-10,20", 3);
8456    # ("1", undef, "10", ",", "20")
8457
8458    my @x = split(/(-)|,/  , "1-10,20", 3);
8459    # ("1", "-", "10", undef, "20")
8460
8461    my @x = split(/(-)|(,)/, "1-10,20", 3);
8462    # ("1", "-", undef, "10", undef, ",", "20")
8463
8464=item sprintf FORMAT, LIST
8465X<sprintf>
8466
8467=for Pod::Functions formatted print into a string
8468
8469Returns a string formatted by the usual
8470L<C<printf>|/printf FILEHANDLE FORMAT, LIST> conventions of the C
8471library function L<C<sprintf>|/sprintf FORMAT, LIST>.  See below for
8472more details and see L<sprintf(3)> or L<printf(3)> on your system for an
8473explanation of the general principles.
8474
8475For example:
8476
8477        # Format number with up to 8 leading zeroes
8478        my $result = sprintf("%08d", $number);
8479
8480        # Round number to 3 digits after decimal point
8481        my $rounded = sprintf("%.3f", $number);
8482
8483Perl does its own L<C<sprintf>|/sprintf FORMAT, LIST> formatting: it
8484emulates the C
8485function L<sprintf(3)>, but doesn't use it except for floating-point
8486numbers, and even then only standard modifiers are allowed.
8487Non-standard extensions in your local L<sprintf(3)> are
8488therefore unavailable from Perl.
8489
8490Unlike L<C<printf>|/printf FILEHANDLE FORMAT, LIST>,
8491L<C<sprintf>|/sprintf FORMAT, LIST> does not do what you probably mean
8492when you pass it an array as your first argument.
8493The array is given scalar context,
8494and instead of using the 0th element of the array as the format, Perl will
8495use the count of elements in the array as the format, which is almost never
8496useful.
8497
8498Perl's L<C<sprintf>|/sprintf FORMAT, LIST> permits the following
8499universally-known conversions:
8500
8501   %%    a percent sign
8502   %c    a character with the given number
8503   %s    a string
8504   %d    a signed integer, in decimal
8505   %u    an unsigned integer, in decimal
8506   %o    an unsigned integer, in octal
8507   %x    an unsigned integer, in hexadecimal
8508   %e    a floating-point number, in scientific notation
8509   %f    a floating-point number, in fixed decimal notation
8510   %g    a floating-point number, in %e or %f notation
8511
8512In addition, Perl permits the following widely-supported conversions:
8513
8514   %X    like %x, but using upper-case letters
8515   %E    like %e, but using an upper-case "E"
8516   %G    like %g, but with an upper-case "E" (if applicable)
8517   %b    an unsigned integer, in binary
8518   %B    like %b, but using an upper-case "B" with the # flag
8519   %p    a pointer (outputs the Perl value's address in hexadecimal)
8520   %n    special: *stores* the number of characters output so far
8521         into the next argument in the parameter list
8522   %a    hexadecimal floating point
8523   %A    like %a, but using upper-case letters
8524   %i    a synonym for %d
8525   %D    a synonym for %ld
8526   %U    a synonym for %lu
8527   %O    a synonym for %lo
8528   %F    a synonym for %f
8529
8530Note that the number of exponent digits in the scientific notation produced
8531by C<%e>, C<%E>, C<%g> and C<%G> for numbers with the modulus of the
8532exponent less than 100 is system-dependent: it may be three or less
8533(zero-padded as necessary).  In other words, 1.23 times ten to the
853499th may be either "1.23e99" or "1.23e099".  Similarly for C<%a> and C<%A>:
8535the exponent or the hexadecimal digits may float: especially the
8536"long doubles" Perl configuration option may cause surprises.
8537
8538Between the C<%> and the format letter, you may specify several
8539additional attributes controlling the interpretation of the format.
8540In order, these are:
8541
8542=over 4
8543
8544=item format parameter index
8545
8546An explicit format parameter index, such as C<2$>.  By default sprintf
8547will format the next unused argument in the list, but this allows you
8548to take the arguments out of order:
8549
8550  printf '%2$d %1$d', 12, 34;      # prints "34 12"
8551  printf '%3$d %d %1$d', 1, 2, 3;  # prints "3 1 1"
8552
8553=item flags
8554
8555one or more of:
8556
8557   space   prefix non-negative number with a space
8558   +       prefix non-negative number with a plus sign
8559   -       left-justify within the field
8560   0       use zeros, not spaces, to right-justify
8561   #       ensure the leading "0" for any octal,
8562           prefix non-zero hexadecimal with "0x" or "0X",
8563           prefix non-zero binary with "0b" or "0B"
8564
8565For example:
8566
8567  printf '<% d>',  12;   # prints "< 12>"
8568  printf '<% d>',   0;   # prints "< 0>"
8569  printf '<% d>', -12;   # prints "<-12>"
8570  printf '<%+d>',  12;   # prints "<+12>"
8571  printf '<%+d>',   0;   # prints "<+0>"
8572  printf '<%+d>', -12;   # prints "<-12>"
8573  printf '<%6s>',  12;   # prints "<    12>"
8574  printf '<%-6s>', 12;   # prints "<12    >"
8575  printf '<%06s>', 12;   # prints "<000012>"
8576  printf '<%#o>',  12;   # prints "<014>"
8577  printf '<%#x>',  12;   # prints "<0xc>"
8578  printf '<%#X>',  12;   # prints "<0XC>"
8579  printf '<%#b>',  12;   # prints "<0b1100>"
8580  printf '<%#B>',  12;   # prints "<0B1100>"
8581
8582When a space and a plus sign are given as the flags at once,
8583the space is ignored.
8584
8585  printf '<%+ d>', 12;   # prints "<+12>"
8586  printf '<% +d>', 12;   # prints "<+12>"
8587
8588When the # flag and a precision are given in the %o conversion,
8589the precision is incremented if it's necessary for the leading "0".
8590
8591  printf '<%#.5o>', 012;      # prints "<00012>"
8592  printf '<%#.5o>', 012345;   # prints "<012345>"
8593  printf '<%#.0o>', 0;        # prints "<0>"
8594
8595=item vector flag
8596
8597This flag tells Perl to interpret the supplied string as a vector of
8598integers, one for each character in the string.  Perl applies the format to
8599each integer in turn, then joins the resulting strings with a separator (a
8600dot C<.> by default).  This can be useful for displaying ordinal values of
8601characters in arbitrary strings:
8602
8603  printf "%vd", "AB\x{100}";           # prints "65.66.256"
8604  printf "version is v%vd\n", $^V;     # Perl's version
8605
8606Put an asterisk C<*> before the C<v> to override the string to
8607use to separate the numbers:
8608
8609  printf "address is %*vX\n", ":", $addr;   # IPv6 address
8610  printf "bits are %0*v8b\n", " ", $bits;   # random bitstring
8611
8612You can also explicitly specify the argument number to use for
8613the join string using something like C<*2$v>; for example:
8614
8615  printf '%*4$vX %*4$vX %*4$vX',       # 3 IPv6 addresses
8616          @addr[1..3], ":";
8617
8618=item (minimum) width
8619
8620Arguments are usually formatted to be only as wide as required to
8621display the given value.  You can override the width by putting
8622a number here, or get the width from the next argument (with C<*>)
8623or from a specified argument (e.g., with C<*2$>):
8624
8625 printf "<%s>", "a";       # prints "<a>"
8626 printf "<%6s>", "a";      # prints "<     a>"
8627 printf "<%*s>", 6, "a";   # prints "<     a>"
8628 printf '<%*2$s>', "a", 6; # prints "<     a>"
8629 printf "<%2s>", "long";   # prints "<long>" (does not truncate)
8630
8631If a field width obtained through C<*> is negative, it has the same
8632effect as the C<-> flag: left-justification.
8633
8634=item precision, or maximum width
8635X<precision>
8636
8637You can specify a precision (for numeric conversions) or a maximum
8638width (for string conversions) by specifying a C<.> followed by a number.
8639For floating-point formats except C<g> and C<G>, this specifies
8640how many places right of the decimal point to show (the default being 6).
8641For example:
8642
8643  # these examples are subject to system-specific variation
8644  printf '<%f>', 1;       # prints "<1.000000>"
8645  printf '<%.1f>', 1;     # prints "<1.0>"
8646  printf '<%.0f>', 1;     # prints "<1>"
8647  printf '<%07.2f>', 1.3; # prints "<0001.30>"
8648  printf '<%e>', 10;      # prints "<1.000000e+01>"
8649  printf '<%.1e>', 10;    # prints "<1.0e+01>"
8650
8651For "g" and "G", this specifies the maximum number of significant digits to
8652show; for example:
8653
8654  # These examples are subject to system-specific variation.
8655  printf '<%g>', 1;        # prints "<1>"
8656  printf '<%.10g>', 1;     # prints "<1>"
8657  printf '<%g>', 100;      # prints "<100>"
8658  printf '<%.1g>', 100;    # prints "<1e+02>"
8659  printf '<%.2g>', 100.01; # prints "<1e+02>"
8660  printf '<%.5g>', 100.01; # prints "<100.01>"
8661  printf '<%.4g>', 100.01; # prints "<100>"
8662  printf '<%.1g>', 0.0111; # prints "<0.01>"
8663  printf '<%.2g>', 0.0111; # prints "<0.011>"
8664  printf '<%.3g>', 0.0111; # prints "<0.0111>"
8665
8666For integer conversions, specifying a precision implies that the
8667output of the number itself should be zero-padded to this width,
8668where the 0 flag is ignored:
8669
8670  printf '<%.6d>', 1;      # prints "<000001>"
8671  printf '<%+.6d>', 1;     # prints "<+000001>"
8672  printf '<%-10.6d>', 1;   # prints "<000001    >"
8673  printf '<%10.6d>', 1;    # prints "<    000001>"
8674  printf '<%010.6d>', 1;   # prints "<    000001>"
8675  printf '<%+10.6d>', 1;   # prints "<   +000001>"
8676
8677  printf '<%.6x>', 1;      # prints "<000001>"
8678  printf '<%#.6x>', 1;     # prints "<0x000001>"
8679  printf '<%-10.6x>', 1;   # prints "<000001    >"
8680  printf '<%10.6x>', 1;    # prints "<    000001>"
8681  printf '<%010.6x>', 1;   # prints "<    000001>"
8682  printf '<%#10.6x>', 1;   # prints "<  0x000001>"
8683
8684For string conversions, specifying a precision truncates the string
8685to fit the specified width:
8686
8687  printf '<%.5s>', "truncated";   # prints "<trunc>"
8688  printf '<%10.5s>', "truncated"; # prints "<     trunc>"
8689
8690You can also get the precision from the next argument using C<.*>, or from a
8691specified argument (e.g., with C<.*2$>):
8692
8693  printf '<%.6x>', 1;       # prints "<000001>"
8694  printf '<%.*x>', 6, 1;    # prints "<000001>"
8695
8696  printf '<%.*2$x>', 1, 6;  # prints "<000001>"
8697
8698  printf '<%6.*2$x>', 1, 4; # prints "<  0001>"
8699
8700If a precision obtained through C<*> is negative, it counts
8701as having no precision at all.
8702
8703  printf '<%.*s>',  7, "string";   # prints "<string>"
8704  printf '<%.*s>',  3, "string";   # prints "<str>"
8705  printf '<%.*s>',  0, "string";   # prints "<>"
8706  printf '<%.*s>', -1, "string";   # prints "<string>"
8707
8708  printf '<%.*d>',  1, 0;   # prints "<0>"
8709  printf '<%.*d>',  0, 0;   # prints "<>"
8710  printf '<%.*d>', -1, 0;   # prints "<0>"
8711
8712=item size
8713
8714For numeric conversions, you can specify the size to interpret the
8715number as using C<l>, C<h>, C<V>, C<q>, C<L>, or C<ll>.  For integer
8716conversions (C<d u o x X b i D U O>), numbers are usually assumed to be
8717whatever the default integer size is on your platform (usually 32 or 64
8718bits), but you can override this to use instead one of the standard C types,
8719as supported by the compiler used to build Perl:
8720
8721   hh          interpret integer as C type "char" or "unsigned
8722               char" on Perl 5.14 or later
8723   h           interpret integer as C type "short" or
8724               "unsigned short"
8725   j           interpret integer as C type "intmax_t" on Perl
8726               5.14 or later.  Prior to Perl 5.30 this worked only
8727               with a C99 compiler, hence was unportable before
8728               that release.
8729   l           interpret integer as C type "long" or
8730               "unsigned long"
8731   q, L, or ll interpret integer as C type "long long",
8732               "unsigned long long", or "quad" (typically
8733               64-bit integers)
8734   t           interpret integer as C type "ptrdiff_t" on Perl
8735               5.14 or later
8736   z           interpret integer as C types "size_t" or
8737               "ssize_t" on Perl 5.14 or later
8738
8739Note that, in general, using the C<l> modifier (for example, when writing
8740C<"%ld"> or C<"%lu"> instead of C<"%d"> and C<"%u">) is unnecessary
8741when used from Perl code.  Moreover, it may be harmful, for example on
8742Windows 64-bit where a long is 32-bits.
8743
8744As of 5.14, none of these raises an exception if they are not supported on
8745your platform.  However, if warnings are enabled, a warning of the
8746L<C<printf>|warnings> warning class is issued on an unsupported
8747conversion flag.  Should you instead prefer an exception, do this:
8748
8749    use warnings FATAL => "printf";
8750
8751If you would like to know about a version dependency before you
8752start running the program, put something like this at its top:
8753
8754    use v5.14;  # for hh/j/t/z/ printf modifiers
8755
8756You can find out whether your Perl supports quads via L<Config>:
8757
8758    use Config;
8759    if ($Config{use64bitint} eq "define"
8760        || $Config{longsize} >= 8) {
8761        print "Nice quads!\n";
8762    }
8763
8764For floating-point conversions (C<e f g E F G>), numbers are usually assumed
8765to be the default floating-point size on your platform (double or long double),
8766but you can force "long double" with C<q>, C<L>, or C<ll> if your
8767platform supports them.  You can find out whether your Perl supports long
8768doubles via L<Config>:
8769
8770    use Config;
8771    print "long doubles\n" if $Config{d_longdbl} eq "define";
8772
8773You can find out whether Perl considers "long double" to be the default
8774floating-point size to use on your platform via L<Config>:
8775
8776    use Config;
8777    if ($Config{uselongdouble} eq "define") {
8778        print "long doubles by default\n";
8779    }
8780
8781It can also be that long doubles and doubles are the same thing:
8782
8783        use Config;
8784        ($Config{doublesize} == $Config{longdblsize}) &&
8785                print "doubles are long doubles\n";
8786
8787The size specifier C<V> has no effect for Perl code, but is supported for
8788compatibility with XS code.  It means "use the standard size for a Perl
8789integer or floating-point number", which is the default.
8790
8791=item order of arguments
8792
8793Normally, L<C<sprintf>|/sprintf FORMAT, LIST> takes the next unused
8794argument as the value to
8795format for each format specification.  If the format specification
8796uses C<*> to require additional arguments, these are consumed from
8797the argument list in the order they appear in the format
8798specification I<before> the value to format.  Where an argument is
8799specified by an explicit index, this does not affect the normal
8800order for the arguments, even when the explicitly specified index
8801would have been the next argument.
8802
8803So:
8804
8805    printf "<%*.*s>", $x, $y, $z;
8806
8807uses C<$x> for the width, C<$y> for the precision, and C<$z>
8808as the value to format; while:
8809
8810  printf '<%*1$.*s>', $x, $y;
8811
8812would use C<$x> for the width and precision, and C<$y> as the
8813value to format.
8814
8815Here are some more examples; be aware that when using an explicit
8816index, the C<$> may need escaping:
8817
8818 printf "%2\$d %d\n",      12, 34;     # will print "34 12\n"
8819 printf "%2\$d %d %d\n",   12, 34;     # will print "34 12 34\n"
8820 printf "%3\$d %d %d\n",   12, 34, 56; # will print "56 12 34\n"
8821 printf "%2\$*3\$d %d\n",  12, 34,  3; # will print " 34 12\n"
8822 printf "%*1\$.*f\n",       4,  5, 10; # will print "5.0000\n"
8823
8824=back
8825
8826If L<C<use locale>|locale> (including C<use locale ':not_characters'>)
8827is in effect and L<C<POSIX::setlocale>|POSIX/C<setlocale>> has been
8828called,
8829the character used for the decimal separator in formatted floating-point
8830numbers is affected by the C<LC_NUMERIC> locale.  See L<perllocale>
8831and L<POSIX>.
8832
8833=item sqrt EXPR
8834X<sqrt> X<root> X<square root>
8835
8836=item sqrt
8837
8838=for Pod::Functions square root function
8839
8840Return the positive square root of EXPR.  If EXPR is omitted, uses
8841L<C<$_>|perlvar/$_>.  Works only for non-negative operands unless you've
8842loaded the L<C<Math::Complex>|Math::Complex> module.
8843
8844    use Math::Complex;
8845    print sqrt(-4);    # prints 2i
8846
8847=item srand EXPR
8848X<srand> X<seed> X<randseed>
8849
8850=item srand
8851
8852=for Pod::Functions seed the random number generator
8853
8854Sets and returns the random number seed for the L<C<rand>|/rand EXPR>
8855operator.
8856
8857The point of the function is to "seed" the L<C<rand>|/rand EXPR>
8858function so that L<C<rand>|/rand EXPR> can produce a different sequence
8859each time you run your program.  When called with a parameter,
8860L<C<srand>|/srand EXPR> uses that for the seed; otherwise it
8861(semi-)randomly chooses a seed (see below).  In either case, starting with Perl 5.14,
8862it returns the seed.  To signal that your code will work I<only> on Perls
8863of a recent vintage:
8864
8865    use v5.14;	# so srand returns the seed
8866
8867If L<C<srand>|/srand EXPR> is not called explicitly, it is called
8868implicitly without a parameter at the first use of the
8869L<C<rand>|/rand EXPR> operator.  However, there are a few situations
8870where programs are likely to want to call L<C<srand>|/srand EXPR>.  One
8871is for generating predictable results, generally for testing or
8872debugging.  There, you use C<srand($seed)>, with the same C<$seed> each
8873time.  Another case is that you may want to call L<C<srand>|/srand EXPR>
8874after a L<C<fork>|/fork> to avoid child processes sharing the same seed
8875value as the parent (and consequently each other).
8876
8877Do B<not> call C<srand()> (i.e., without an argument) more than once per
8878process.  The internal state of the random number generator should
8879contain more entropy than can be provided by any seed, so calling
8880L<C<srand>|/srand EXPR> again actually I<loses> randomness.
8881
8882Most implementations of L<C<srand>|/srand EXPR> take an integer and will
8883silently
8884truncate decimal numbers.  This means C<srand(42)> will usually
8885produce the same results as C<srand(42.1)>.  To be safe, always pass
8886L<C<srand>|/srand EXPR> an integer.
8887
8888A typical use of the returned seed is for a test program which has too many
8889combinations to test comprehensively in the time available to it each run.  It
8890can test a random subset each time, and should there be a failure, log the seed
8891used for that run so that it can later be used to reproduce the same results.
8892
8893If the C<PERL_RAND_SEED> environment variable is set to a non-negative
8894integer during process startup then calls to C<srand()> with no
8895arguments will initialize the perl random number generator with a
8896consistent seed each time it is called, whether called explicitly with
8897no arguments or implicitly via use of C<rand()>. The exact seeding that
8898a given C<PERL_RAND_SEED> will produce is deliberately unspecified, but
8899using different values for C<PERL_RAND_SEED> should produce different
8900results. This is intended for debugging and performance analysis and is
8901only guaranteed to produce consistent results between invocations of the
8902same perl executable running the same code when all other factors are
8903equal. The environment variable is read only once during process
8904startup, and changing it during the program flow will not affect the
8905currently running process. See L<perlrun> for more details.
8906
8907B<L<C<rand>|/rand EXPR> is not cryptographically secure.  You should not rely
8908on it in security-sensitive situations.>  As of this writing, a
8909number of third-party CPAN modules offer random number generators
8910intended by their authors to be cryptographically secure,
8911including: L<Data::Entropy>, L<Crypt::Random>, L<Math::Random::Secure>,
8912and L<Math::TrulyRandom>.
8913
8914=item stat FILEHANDLE
8915X<stat> X<file, status> X<ctime>
8916
8917=item stat EXPR
8918
8919=item stat DIRHANDLE
8920
8921=item stat
8922
8923=for Pod::Functions get a file's status information
8924
8925Returns a 13-element list giving the status info for a file, either
8926the file opened via FILEHANDLE or DIRHANDLE, or named by EXPR.  If EXPR is
8927omitted, it stats L<C<$_>|perlvar/$_> (not C<_>!).  Returns the empty
8928list if L<C<stat>|/stat FILEHANDLE> fails.  Typically
8929used as follows:
8930
8931    my ($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size,
8932        $atime,$mtime,$ctime,$blksize,$blocks)
8933           = stat($filename);
8934
8935Not all fields are supported on all filesystem types.  Here are the
8936meanings of the fields:
8937
8938  0 dev      device number of filesystem
8939  1 ino      inode number
8940  2 mode     file mode  (type and permissions)
8941  3 nlink    number of (hard) links to the file
8942  4 uid      numeric user ID of file's owner
8943  5 gid      numeric group ID of file's owner
8944  6 rdev     the device identifier (special files only)
8945  7 size     total size of file, in bytes
8946  8 atime    last access time in seconds since the epoch
8947  9 mtime    last modify time in seconds since the epoch
8948 10 ctime    inode change time in seconds since the epoch (*)
8949 11 blksize  preferred I/O size in bytes for interacting with the
8950             file (may vary from file to file)
8951 12 blocks   actual number of system-specific blocks allocated
8952             on disk (often, but not always, 512 bytes each)
8953
8954(The epoch was at 00:00 January 1, 1970 GMT.)
8955
8956(*) Not all fields are supported on all filesystem types.  Notably, the
8957ctime field is non-portable.  In particular, you cannot expect it to be a
8958"creation time"; see L<perlport/"Files and Filesystems"> for details.
8959
8960If L<C<stat>|/stat FILEHANDLE> is passed the special filehandle
8961consisting of an underline, no stat is done, but the current contents of
8962the stat structure from the last L<C<stat>|/stat FILEHANDLE>,
8963L<C<lstat>|/lstat FILEHANDLE>, or filetest are returned.  Example:
8964
8965    if (-x $file && (($d) = stat(_)) && $d < 0) {
8966        print "$file is executable NFS file\n";
8967    }
8968
8969(This works on machines only for which the device number is negative
8970under NFS.)
8971
8972On some platforms inode numbers are of a type larger than perl knows how
8973to handle as integer numerical values.  If necessary, an inode number will
8974be returned as a decimal string in order to preserve the entire value.
8975If used in a numeric context, this will be converted to a floating-point
8976numerical value, with rounding, a fate that is best avoided.  Therefore,
8977you should prefer to compare inode numbers using C<eq> rather than C<==>.
8978C<eq> will work fine on inode numbers that are represented numerically,
8979as well as those represented as strings.
8980
8981Because the mode contains both the file type and its permissions, you
8982should mask off the file type portion and (s)printf using a C<"%o">
8983if you want to see the real permissions.
8984
8985    my $mode = (stat($filename))[2];
8986    printf "Permissions are %04o\n", $mode & 07777;
8987
8988In scalar context, L<C<stat>|/stat FILEHANDLE> returns a boolean value
8989indicating success
8990or failure, and, if successful, sets the information associated with
8991the special filehandle C<_>.
8992
8993The L<File::stat> module provides a convenient, by-name access mechanism:
8994
8995    use File::stat;
8996    my $sb = stat($filename);
8997    printf "File is %s, size is %s, perm %04o, mtime %s\n",
8998           $filename, $sb->size, $sb->mode & 07777,
8999           scalar localtime $sb->mtime;
9000
9001You can import symbolic mode constants and functions
9002(C<S_I*>) from the L<Fcntl> module:
9003
9004    use Fcntl ':mode';
9005
9006    my $mode = (stat($filename))[2];
9007
9008    my $user_rwx      = ($mode & S_IRWXU) >> 6;
9009    my $group_read    = ($mode & S_IRGRP) >> 3;
9010    my $other_execute =  $mode & S_IXOTH;
9011
9012    printf "Permissions are %04o\n", S_IMODE($mode), "\n";
9013
9014    my $is_setuid     =  $mode & S_ISUID;
9015    my $is_directory  =  S_ISDIR($mode);
9016
9017You could write the last two using the C<-u> and C<-d> operators.
9018Commonly available C<S_I*> constants are:
9019
9020    # Permissions: read, write, execute, for user, group, others.
9021
9022    S_IRWXU S_IRUSR S_IWUSR S_IXUSR
9023    S_IRWXG S_IRGRP S_IWGRP S_IXGRP
9024    S_IRWXO S_IROTH S_IWOTH S_IXOTH
9025
9026    # Setuid/Setgid/Stickiness/SaveText/EnforcedLocks.
9027    # Note that the exact meaning of these is system-dependent.
9028
9029    S_ISUID S_ISGID S_ISVTX S_ISTXT S_ENFMT
9030
9031    # File types.  Not all are necessarily available on
9032    # your system.
9033
9034    S_IFREG S_IFDIR S_IFLNK S_IFBLK S_IFCHR
9035    S_IFIFO S_IFSOCK S_IFWHT
9036
9037    # The following are compatibility aliases for S_IRUSR,
9038    # S_IWUSR, and S_IXUSR.
9039
9040    S_IREAD S_IWRITE S_IEXEC
9041
9042and the C<S_I*> functions are
9043
9044    S_IMODE($mode)    the part of $mode containing the permission
9045                      bits and the setuid/setgid/sticky bits
9046
9047    S_IFMT($mode)     the part of $mode containing the file type,
9048                      which will match one of the S_IF* constants
9049                      (e.g. S_IFMT($mode) == S_IFDIR for directories),
9050                      but see the following helper functions
9051
9052    # The operators -f, -d, -l, -b, -c, -p, and -S.
9053
9054    S_ISREG($mode) S_ISDIR($mode) S_ISLNK($mode)
9055    S_ISBLK($mode) S_ISCHR($mode) S_ISFIFO($mode) S_ISSOCK($mode)
9056
9057    # No direct -X operator counterpart, but for the first one
9058    # the -g operator is often equivalent.  The ENFMT stands for
9059    # record flocking enforcement, a platform-dependent feature.
9060
9061    S_ISENFMT($mode) S_ISWHT($mode)
9062
9063See your native L<chmod(2)> and L<stat(2)> documentation for more details
9064about the C<S_I*> constants.  To get status info for a symbolic link
9065instead of the target file behind the link, use the
9066L<C<lstat>|/lstat FILEHANDLE> function.
9067
9068Portability issues: L<perlport/stat>.
9069
9070=item state VARLIST
9071X<state>
9072
9073=item state TYPE VARLIST
9074
9075=item state VARLIST : ATTRS
9076
9077=item state TYPE VARLIST : ATTRS
9078
9079=for Pod::Functions +state declare and assign a persistent lexical variable
9080
9081L<C<state>|/state VARLIST> declares a lexically scoped variable, just
9082like L<C<my>|/my VARLIST>.
9083However, those variables will never be reinitialized, contrary to
9084lexical variables that are reinitialized each time their enclosing block
9085is entered.
9086See L<perlsub/"Persistent Private Variables"> for details.
9087
9088If more than one variable is listed, the list must be placed in
9089parentheses.  With a parenthesised list, L<C<undef>|/undef EXPR> can be
9090used as a
9091dummy placeholder.  However, since initialization of state variables in
9092such lists is currently not possible this would serve no purpose.
9093
9094Like L<C<my>|/my VARLIST>, L<C<local>|/local EXPR>, and
9095L<C<our>|/our VARLIST>, L<C<state>|/state VARLIST> can operate on a variable
9096anywhere it appears in an expression (aside from interpolation in strings).
9097The declaration will not apply to additional uses of the same variable until
9098the next statement. This means additional uses of that variable within the
9099same statement will act as they would have before that declaration occurred,
9100or result in a strict 'vars' error, as appropriate.
9101
9102    package main;
9103    use feature 'state';
9104    our $x = 2;
9105    foo($x, state $x = $x + 1, $x); # foo() receives (2, 3, 2)
9106    foo($x, $main::x);              # foo() receives (3, 2)
9107
9108Redeclaring a variable in the same scope or statement will "shadow" the
9109previous declaration, creating a new instance and preventing access to
9110the previous one. This is usually undesired and, if warnings are enabled,
9111will result in a warning in the C<shadow> category.
9112
9113L<C<state>|/state VARLIST> is available only if the
9114L<C<"state"> feature|feature/The 'state' feature> is enabled or if it is
9115prefixed with C<CORE::>.  The
9116L<C<"state"> feature|feature/The 'state' feature> is enabled
9117automatically with a C<use v5.10> (or higher) declaration in the current
9118scope.
9119
9120
9121=item study SCALAR
9122X<study>
9123
9124=item study
9125
9126=for Pod::Functions no-op, formerly optimized input data for repeated searches
9127
9128At this time, C<study> does nothing. This may change in the future.
9129
9130Prior to Perl version 5.16, it would create an inverted index of all characters
9131that occurred in the given SCALAR (or L<C<$_>|perlvar/$_> if unspecified). When
9132matching a pattern, the rarest character from the pattern would be looked up in
9133this index. Rarity was based on some static frequency tables constructed from
9134some C programs and English text.
9135
9136
9137=item sub NAME BLOCK
9138X<sub>
9139
9140=item sub NAME (PROTO) BLOCK
9141
9142=item sub NAME : ATTRS BLOCK
9143
9144=item sub NAME (PROTO) : ATTRS BLOCK
9145
9146=for Pod::Functions declare a subroutine, possibly anonymously
9147
9148This is subroutine definition, not a real function I<per se>.  Without a
9149BLOCK it's just a forward declaration.  Without a NAME, it's an anonymous
9150function declaration, so does return a value: the CODE ref of the closure
9151just created.
9152
9153See L<perlsub> and L<perlref> for details about subroutines and
9154references; see L<attributes> and L<Attribute::Handlers> for more
9155information about attributes.
9156
9157=item __SUB__
9158X<__SUB__>
9159
9160=for Pod::Functions +current_sub the current subroutine, or C<undef> if not in a subroutine
9161
9162A special token that returns a reference to the current subroutine, or
9163L<C<undef>|/undef EXPR> outside of a subroutine.
9164
9165The behaviour of L<C<__SUB__>|/__SUB__> within a regex code block (such
9166as C</(?{...})/>) is subject to change.
9167
9168This token is only available under C<use v5.16> or the
9169L<C<"current_sub"> feature|feature/The 'current_sub' feature>.
9170See L<feature>.
9171
9172=item substr EXPR,OFFSET,LENGTH,REPLACEMENT
9173X<substr> X<substring> X<mid> X<left> X<right>
9174
9175=item substr EXPR,OFFSET,LENGTH
9176
9177=item substr EXPR,OFFSET
9178
9179=for Pod::Functions get or alter a portion of a string
9180
9181Extracts a substring out of EXPR and returns it.  First character is at
9182offset zero.  If OFFSET is negative, starts
9183that far back from the end of the string.  If LENGTH is omitted, returns
9184everything through the end of the string.  If LENGTH is negative, leaves that
9185many characters off the end of the string.
9186
9187    my $s = "The black cat climbed the green tree";
9188    my $color  = substr $s, 4, 5;      # black
9189    my $middle = substr $s, 4, -11;    # black cat climbed the
9190    my $end    = substr $s, 14;        # climbed the green tree
9191    my $tail   = substr $s, -4;        # tree
9192    my $z      = substr $s, -4, 2;     # tr
9193
9194You can use the L<C<substr>|/substr EXPR,OFFSET,LENGTH,REPLACEMENT>
9195function as an lvalue, in which case EXPR
9196must itself be an lvalue.  If you assign something shorter than LENGTH,
9197the string will shrink, and if you assign something longer than LENGTH,
9198the string will grow to accommodate it.  To keep the string the same
9199length, you may need to pad or chop your value using
9200L<C<sprintf>|/sprintf FORMAT, LIST>.
9201
9202If OFFSET and LENGTH specify a substring that is partly outside the
9203string, only the part within the string is returned.  If the substring
9204is beyond either end of the string,
9205L<C<substr>|/substr EXPR,OFFSET,LENGTH,REPLACEMENT> returns the undefined
9206value and produces a warning.  When used as an lvalue, specifying a
9207substring that is entirely outside the string raises an exception.
9208Here's an example showing the behavior for boundary cases:
9209
9210    my $name = 'fred';
9211    substr($name, 4) = 'dy';         # $name is now 'freddy'
9212    my $null = substr $name, 6, 2;   # returns "" (no warning)
9213    my $oops = substr $name, 7;      # returns undef, with warning
9214    substr($name, 7) = 'gap';        # raises an exception
9215
9216An alternative to using
9217L<C<substr>|/substr EXPR,OFFSET,LENGTH,REPLACEMENT> as an lvalue is to
9218specify the
9219REPLACEMENT string as the 4th argument.  This allows you to replace
9220parts of the EXPR and return what was there before in one operation,
9221just as you can with
9222L<C<splice>|/splice ARRAY,OFFSET,LENGTH,LIST>.
9223
9224    my $s = "The black cat climbed the green tree";
9225    my $z = substr $s, 14, 7, "jumped from";    # climbed
9226    # $s is now "The black cat jumped from the green tree"
9227
9228Note that the lvalue returned by the three-argument version of
9229L<C<substr>|/substr EXPR,OFFSET,LENGTH,REPLACEMENT> acts as
9230a 'magic bullet'; each time it is assigned to, it remembers which part
9231of the original string is being modified; for example:
9232
9233    my $x = '1234';
9234    for (substr($x,1,2)) {
9235        $_ = 'a';   print $x,"\n";    # prints 1a4
9236        $_ = 'xyz'; print $x,"\n";    # prints 1xyz4
9237        $x = '56789';
9238        $_ = 'pq';  print $x,"\n";    # prints 5pq9
9239    }
9240
9241With negative offsets, it remembers its position from the end of the string
9242when the target string is modified:
9243
9244    my $x = '1234';
9245    for (substr($x, -3, 2)) {
9246        $_ = 'a';   print $x,"\n";    # prints 1a4, as above
9247        $x = 'abcdefg';
9248        print $_,"\n";                # prints f
9249    }
9250
9251Prior to Perl version 5.10, the result of using an lvalue multiple times was
9252unspecified.  Prior to 5.16, the result with negative offsets was
9253unspecified.
9254
9255=item symlink OLDFILE,NEWFILE
9256X<symlink> X<link> X<symbolic link> X<link, symbolic>
9257
9258=for Pod::Functions create a symbolic link to a file
9259
9260Creates a new filename symbolically linked to the old filename.
9261Returns C<1> for success, C<0> otherwise.  On systems that don't support
9262symbolic links, raises an exception.  To check for that,
9263use eval:
9264
9265    my $symlink_exists = eval { symlink("",""); 1 };
9266
9267Portability issues: L<perlport/symlink>.
9268
9269=item syscall NUMBER, LIST
9270X<syscall> X<system call>
9271
9272=for Pod::Functions execute an arbitrary system call
9273
9274Calls the system call specified as the first element of the list,
9275passing the remaining elements as arguments to the system call.  If
9276unimplemented, raises an exception.  The arguments are interpreted
9277as follows: if a given argument is numeric, the argument is passed as
9278an int.  If not, the pointer to the string value is passed.  You are
9279responsible to make sure a string is pre-extended long enough to
9280receive any result that might be written into a string.  You can't use a
9281string literal (or other read-only string) as an argument to
9282L<C<syscall>|/syscall NUMBER, LIST> because Perl has to assume that any
9283string pointer might be written through.  If your
9284integer arguments are not literals and have never been interpreted in a
9285numeric context, you may need to add C<0> to them to force them to look
9286like numbers.  This emulates the
9287L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET> function (or
9288vice versa):
9289
9290    require 'syscall.ph';        # may need to run h2ph
9291    my $s = "hi there\n";
9292    syscall(SYS_write(), fileno(STDOUT), $s, length $s);
9293
9294Note that Perl supports passing of up to only 14 arguments to your syscall,
9295which in practice should (usually) suffice.
9296
9297Syscall returns whatever value returned by the system call it calls.
9298If the system call fails, L<C<syscall>|/syscall NUMBER, LIST> returns
9299C<-1> and sets L<C<$!>|perlvar/$!> (errno).
9300Note that some system calls I<can> legitimately return C<-1>.  The proper
9301way to handle such calls is to assign C<$! = 0> before the call, then
9302check the value of L<C<$!>|perlvar/$!> if
9303L<C<syscall>|/syscall NUMBER, LIST> returns C<-1>.
9304
9305There's a problem with C<syscall(SYS_pipe())>: it returns the file
9306number of the read end of the pipe it creates, but there is no way
9307to retrieve the file number of the other end.  You can avoid this
9308problem by using L<C<pipe>|/pipe READHANDLE,WRITEHANDLE> instead.
9309
9310Portability issues: L<perlport/syscall>.
9311
9312=item sysopen FILEHANDLE,FILENAME,MODE
9313X<sysopen>
9314
9315=item sysopen FILEHANDLE,FILENAME,MODE,PERMS
9316
9317=for Pod::Functions +5.002 open a file, pipe, or descriptor
9318
9319Opens the file whose filename is given by FILENAME, and associates it with
9320FILEHANDLE.  If FILEHANDLE is an expression, its value is used as the real
9321filehandle wanted; an undefined scalar will be suitably autovivified.  This
9322function calls the underlying operating system's L<open(2)> function with the
9323parameters FILENAME, MODE, and PERMS.
9324
9325Returns true on success and L<C<undef>|/undef EXPR> otherwise.
9326
9327L<PerlIO> layers will be applied to the handle the same way they would in an
9328L<C<open>|/open FILEHANDLE,MODE,EXPR> call that does not specify layers. That is,
9329the current value of L<C<${^OPEN}>|perlvar/${^OPEN}> as set by the L<open>
9330pragma in a lexical scope, or the C<-C> command-line option or C<PERL_UNICODE>
9331environment variable in the main program scope, falling back to the platform
9332defaults as described in L<PerlIO/Defaults and how to override them>. If you
9333want to remove any layers that may transform the byte stream, use
9334L<C<binmode>|/binmode FILEHANDLE, LAYER> after opening it.
9335
9336The possible values and flag bits of the MODE parameter are
9337system-dependent; they are available via the standard module
9338L<C<Fcntl>|Fcntl>.  See the documentation of your operating system's
9339L<open(2)> syscall to see
9340which values and flag bits are available.  You may combine several flags
9341using the C<|>-operator.
9342
9343Some of the most common values are C<O_RDONLY> for opening the file in
9344read-only mode, C<O_WRONLY> for opening the file in write-only mode,
9345and C<O_RDWR> for opening the file in read-write mode.
9346X<O_RDONLY> X<O_RDWR> X<O_WRONLY>
9347
9348For historical reasons, some values work on almost every system
9349supported by Perl: 0 means read-only, 1 means write-only, and 2
9350means read/write.  We know that these values do I<not> work under
9351OS/390; you probably don't want to use them in new code.
9352
9353If the file named by FILENAME does not exist and the
9354L<C<open>|/open FILEHANDLE,MODE,EXPR> call creates
9355it (typically because MODE includes the C<O_CREAT> flag), then the value of
9356PERMS specifies the permissions of the newly created file.  If you omit
9357the PERMS argument to L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE>,
9358Perl uses the octal value C<0666>.
9359These permission values need to be in octal, and are modified by your
9360process's current L<C<umask>|/umask EXPR>.
9361X<O_CREAT>
9362
9363In many systems the C<O_EXCL> flag is available for opening files in
9364exclusive mode.  This is B<not> locking: exclusiveness means here that
9365if the file already exists,
9366L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE> fails.  C<O_EXCL> may
9367not work
9368on network filesystems, and has no effect unless the C<O_CREAT> flag
9369is set as well.  Setting C<O_CREAT|O_EXCL> prevents the file from
9370being opened if it is a symbolic link.  It does not protect against
9371symbolic links in the file's path.
9372X<O_EXCL>
9373
9374Sometimes you may want to truncate an already-existing file.  This
9375can be done using the C<O_TRUNC> flag.  The behavior of
9376C<O_TRUNC> with C<O_RDONLY> is undefined.
9377X<O_TRUNC>
9378
9379You should seldom if ever use C<0644> as argument to
9380L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE>, because
9381that takes away the user's option to have a more permissive umask.
9382Better to omit it.  See L<C<umask>|/umask EXPR> for more on this.
9383
9384This function has no direct relation to the usage of
9385L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET>,
9386L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET>,
9387or L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>.  A handle opened with
9388this function can be used with buffered IO just as one opened with
9389L<C<open>|/open FILEHANDLE,MODE,EXPR> can be used with unbuffered IO.
9390
9391Note that under Perls older than 5.8.0,
9392L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE> depends on the
9393L<fdopen(3)> C library function.  On many Unix systems, L<fdopen(3)> is known
9394to fail when file descriptors exceed a certain value, typically 255.  If
9395you need more file descriptors than that, consider using the
9396L<C<POSIX::open>|POSIX/C<open>> function.  For Perls 5.8.0 and later,
9397PerlIO is (most often) the default.
9398
9399See L<perlopentut> for a kinder, gentler explanation of opening files.
9400
9401Portability issues: L<perlport/sysopen>.
9402
9403=item sysread FILEHANDLE,SCALAR,LENGTH,OFFSET
9404X<sysread>
9405
9406=item sysread FILEHANDLE,SCALAR,LENGTH
9407
9408=for Pod::Functions fixed-length unbuffered input from a filehandle
9409
9410Attempts to read LENGTH bytes of data into variable SCALAR from the
9411specified FILEHANDLE, using L<read(2)>.  It bypasses any L<PerlIO> layers
9412including buffered IO (but is affected by the presence of the C<:utf8>
9413layer as described later), so mixing this with other kinds of reads,
9414L<C<print>|/print FILEHANDLE LIST>, L<C<write>|/write FILEHANDLE>,
9415L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
9416L<C<tell>|/tell FILEHANDLE>, or L<C<eof>|/eof FILEHANDLE> can cause
9417confusion because the
9418C<:perlio> or C<:crlf> layers usually buffer data.  Returns the number of
9419bytes actually read, C<0> at end of file, or undef if there was an
9420error (in the latter case L<C<$!>|perlvar/$!> is also set).  SCALAR will
9421be grown or
9422shrunk so that the last byte actually read is the last byte of the
9423scalar after the read.
9424
9425An OFFSET may be specified to place the read data at some place in the
9426string other than the beginning.  A negative OFFSET specifies
9427placement at that many characters counting backwards from the end of
9428the string.  A positive OFFSET greater than the length of SCALAR
9429results in the string being padded to the required size with C<"\0">
9430bytes before the result of the read is appended.
9431
9432There is no syseof() function, which is ok, since
9433L<C<eof>|/eof FILEHANDLE> doesn't work well on device files (like ttys)
9434anyway.  Use L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET> and
9435check for a return value of 0 to decide whether you're done.
9436
9437Note that if the filehandle has been marked as C<:utf8>, C<sysread> will
9438throw an exception.  The C<:encoding(...)> layer implicitly
9439introduces the C<:utf8> layer.  See
9440L<C<binmode>|/binmode FILEHANDLE, LAYER>,
9441L<C<open>|/open FILEHANDLE,MODE,EXPR>, and the L<open> pragma.
9442
9443=item sysseek FILEHANDLE,POSITION,WHENCE
9444X<sysseek> X<lseek>
9445
9446=for Pod::Functions +5.004 position I/O pointer on handle used with sysread and syswrite
9447
9448Sets FILEHANDLE's system position I<in bytes> using L<lseek(2)>.  FILEHANDLE may
9449be an expression whose value gives the name of the filehandle.  The values
9450for WHENCE are C<0> to set the new position to POSITION; C<1> to set it
9451to the current position plus POSITION; and C<2> to set it to EOF plus
9452POSITION, typically negative.
9453
9454Note the emphasis on bytes: even if the filehandle has been set to operate
9455on characters (for example using the C<:encoding(UTF-8)> I/O layer), the
9456L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
9457L<C<tell>|/tell FILEHANDLE>, and
9458L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>
9459family of functions use byte offsets, not character offsets,
9460because seeking to a character offset would be very slow in a UTF-8 file.
9461
9462L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE> bypasses normal
9463buffered IO, so mixing it with reads other than
9464L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET> (for example
9465L<C<readline>|/readline EXPR> or
9466L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET>),
9467L<C<print>|/print FILEHANDLE LIST>, L<C<write>|/write FILEHANDLE>,
9468L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
9469L<C<tell>|/tell FILEHANDLE>, or L<C<eof>|/eof FILEHANDLE> may cause
9470confusion.
9471
9472For WHENCE, you may also use the constants C<SEEK_SET>, C<SEEK_CUR>,
9473and C<SEEK_END> (start of the file, current position, end of the file)
9474from the L<Fcntl> module.  Use of the constants is also more portable
9475than relying on 0, 1, and 2.  For example to define a "systell" function:
9476
9477    use Fcntl 'SEEK_CUR';
9478    sub systell { sysseek($_[0], 0, SEEK_CUR) }
9479
9480Returns the new position, or the undefined value on failure.  A position
9481of zero is returned as the string C<"0 but true">; thus
9482L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE> returns
9483true on success and false on failure, yet you can still easily determine
9484the new position.
9485
9486=item system LIST
9487X<system> X<shell>
9488
9489=item system PROGRAM LIST
9490
9491=for Pod::Functions run a separate program
9492
9493Does exactly the same thing as L<C<exec>|/exec LIST>, except that a fork is
9494done first and the parent process waits for the child process to
9495exit.  Note that argument processing varies depending on the
9496number of arguments.  If there is more than one argument in LIST,
9497or if LIST is an array with more than one value, starts the program
9498given by the first element of the list with arguments given by the
9499rest of the list.  If there is only one scalar argument, the argument
9500is checked for shell metacharacters, and if there are any, the
9501entire argument is passed to the system's command shell for parsing
9502(this is C</bin/sh -c> on Unix platforms, but varies on other
9503platforms).  If there are no shell metacharacters in the argument,
9504it is split into words and passed directly to C<execvp>, which is
9505more efficient.  On Windows, only the C<system PROGRAM LIST> syntax will
9506reliably avoid using the shell; C<system LIST>, even with more than one
9507element, will fall back to the shell if the first spawn fails.
9508
9509Perl will attempt to flush all files opened for
9510output before any operation that may do a fork, but this may not be
9511supported on some platforms (see L<perlport>).  To be safe, you may need
9512to set L<C<$E<verbar>>|perlvar/$E<verbar>> (C<$AUTOFLUSH> in L<English>)
9513or call the C<autoflush> method of L<C<IO::Handle>|IO::Handle/METHODS>
9514on any open handles.
9515
9516The return value is the exit status of the program as returned by the
9517L<C<wait>|/wait> call.  To get the actual exit value, shift right by
9518eight (see below).  See also L<C<exec>|/exec LIST>.  This is I<not> what
9519you want to use to capture the output from a command; for that you
9520should use merely backticks or
9521L<C<qxE<sol>E<sol>>|/qxE<sol>STRINGE<sol>>, as described in
9522L<perlop/"`STRING`">.  Return value of -1 indicates a failure to start
9523the program or an error of the L<wait(2)> system call (inspect
9524L<C<$!>|perlvar/$!> for the reason).
9525
9526If you'd like to make L<C<system>|/system LIST> (and many other bits of
9527Perl) die on error, have a look at the L<autodie> pragma.
9528
9529Like L<C<exec>|/exec LIST>, L<C<system>|/system LIST> allows you to lie
9530to a program about its name if you use the C<system PROGRAM LIST>
9531syntax.  Again, see L<C<exec>|/exec LIST>.
9532
9533Since C<SIGINT> and C<SIGQUIT> are ignored during the execution of
9534L<C<system>|/system LIST>, if you expect your program to terminate on
9535receipt of these signals you will need to arrange to do so yourself
9536based on the return value.
9537
9538    my @args = ("command", "arg1", "arg2");
9539    system(@args) == 0
9540        or die "system @args failed: $?";
9541
9542If you'd like to manually inspect L<C<system>|/system LIST>'s failure,
9543you can check all possible failure modes by inspecting
9544L<C<$?>|perlvar/$?> like this:
9545
9546    if ($? == -1) {
9547        print "failed to execute: $!\n";
9548    }
9549    elsif ($? & 127) {
9550        printf "child died with signal %d, %s coredump\n",
9551            ($? & 127),  ($? & 128) ? 'with' : 'without';
9552    }
9553    else {
9554        printf "child exited with value %d\n", $? >> 8;
9555    }
9556
9557Alternatively, you may inspect the value of
9558L<C<${^CHILD_ERROR_NATIVE}>|perlvar/${^CHILD_ERROR_NATIVE}> with the
9559L<C<W*()>|POSIX/C<WIFEXITED>> calls from the L<POSIX> module.
9560
9561When L<C<system>|/system LIST>'s arguments are executed indirectly by
9562the shell, results and return codes are subject to its quirks.
9563See L<perlop/"`STRING`"> and L<C<exec>|/exec LIST> for details.
9564
9565Since L<C<system>|/system LIST> does a L<C<fork>|/fork> and
9566L<C<wait>|/wait> it may affect a C<SIGCHLD> handler.  See L<perlipc> for
9567details.
9568
9569Portability issues: L<perlport/system>.
9570
9571=item syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET
9572X<syswrite>
9573
9574=item syswrite FILEHANDLE,SCALAR,LENGTH
9575
9576=item syswrite FILEHANDLE,SCALAR
9577
9578=for Pod::Functions fixed-length unbuffered output to a filehandle
9579
9580Attempts to write LENGTH bytes of data from variable SCALAR to the
9581specified FILEHANDLE, using L<write(2)>.  If LENGTH is
9582not specified, writes whole SCALAR.  It bypasses any L<PerlIO> layers
9583including buffered IO (but is affected by the presence of the C<:utf8>
9584layer as described later), so
9585mixing this with reads (other than C<sysread)>),
9586L<C<print>|/print FILEHANDLE LIST>, L<C<write>|/write FILEHANDLE>,
9587L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
9588L<C<tell>|/tell FILEHANDLE>, or L<C<eof>|/eof FILEHANDLE> may cause
9589confusion because the C<:perlio> and C<:crlf> layers usually buffer data.
9590Returns the number of bytes actually written, or L<C<undef>|/undef EXPR>
9591if there was an error (in this case the errno variable
9592L<C<$!>|perlvar/$!> is also set).  If the LENGTH is greater than the
9593data available in the SCALAR after the OFFSET, only as much data as is
9594available will be written.
9595
9596An OFFSET may be specified to write the data from some part of the
9597string other than the beginning.  A negative OFFSET specifies writing
9598that many characters counting backwards from the end of the string.
9599If SCALAR is of length zero, you can only use an OFFSET of 0.
9600
9601B<WARNING>: If the filehandle is marked C<:utf8>, C<syswrite> will raise an exception.
9602The C<:encoding(...)> layer implicitly introduces the C<:utf8> layer.
9603Alternately, if the handle is not marked with an encoding but you
9604attempt to write characters with code points over 255, raises an exception.
9605See L<C<binmode>|/binmode FILEHANDLE, LAYER>,
9606L<C<open>|/open FILEHANDLE,MODE,EXPR>, and the L<open> pragma.
9607
9608=item tell FILEHANDLE
9609X<tell>
9610
9611=item tell
9612
9613=for Pod::Functions get current seekpointer on a filehandle
9614
9615Returns the current position I<in bytes> for FILEHANDLE, or -1 on
9616error.  FILEHANDLE may be an expression whose value gives the name of
9617the actual filehandle.  If FILEHANDLE is omitted, assumes the file
9618last read.
9619
9620Note the emphasis on bytes: even if the filehandle has been set to operate
9621on characters (for example using the C<:encoding(UTF-8)> I/O layer), the
9622L<C<seek>|/seek FILEHANDLE,POSITION,WHENCE>,
9623L<C<tell>|/tell FILEHANDLE>, and
9624L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>
9625family of functions use byte offsets, not character offsets,
9626because seeking to a character offset would be very slow in a UTF-8 file.
9627
9628The return value of L<C<tell>|/tell FILEHANDLE> for the standard streams
9629like the STDIN depends on the operating system: it may return -1 or
9630something else.  L<C<tell>|/tell FILEHANDLE> on pipes, fifos, and
9631sockets usually returns -1.
9632
9633There is no C<systell> function.  Use
9634L<C<sysseek($fh, 0, 1)>|/sysseek FILEHANDLE,POSITION,WHENCE> for that.
9635
9636Do not use L<C<tell>|/tell FILEHANDLE> (or other buffered I/O
9637operations) on a filehandle that has been manipulated by
9638L<C<sysread>|/sysread FILEHANDLE,SCALAR,LENGTH,OFFSET>,
9639L<C<syswrite>|/syswrite FILEHANDLE,SCALAR,LENGTH,OFFSET>, or
9640L<C<sysseek>|/sysseek FILEHANDLE,POSITION,WHENCE>.  Those functions
9641ignore the buffering, while L<C<tell>|/tell FILEHANDLE> does not.
9642
9643=item telldir DIRHANDLE
9644X<telldir>
9645
9646=for Pod::Functions get current seekpointer on a directory handle
9647
9648Returns the current position of the L<C<readdir>|/readdir DIRHANDLE>
9649routines on DIRHANDLE.  Value may be given to
9650L<C<seekdir>|/seekdir DIRHANDLE,POS> to access a particular location in
9651a directory.  L<C<telldir>|/telldir DIRHANDLE> has the same caveats
9652about possible directory compaction as the corresponding system library
9653routine.
9654
9655=item tie VARIABLE,CLASSNAME,LIST
9656X<tie>
9657
9658=for Pod::Functions +5.002 bind a variable to an object class
9659
9660This function binds a variable to a package class that will provide the
9661implementation for the variable.  VARIABLE is the name of the variable
9662to be enchanted.  CLASSNAME is the name of a class implementing objects
9663of correct type.  Any additional arguments are passed to the
9664appropriate constructor
9665method of the class (meaning C<TIESCALAR>, C<TIEHANDLE>, C<TIEARRAY>,
9666or C<TIEHASH>).  Typically these are arguments such as might be passed
9667to the L<dbm_open(3)> function of C.  The object returned by the
9668constructor is also returned by the
9669L<C<tie>|/tie VARIABLE,CLASSNAME,LIST> function, which would be useful
9670if you want to access other methods in CLASSNAME.
9671
9672Note that functions such as L<C<keys>|/keys HASH> and
9673L<C<values>|/values HASH> may return huge lists when used on large
9674objects, like DBM files.  You may prefer to use the L<C<each>|/each
9675HASH> function to iterate over such.  Example:
9676
9677    # print out history file offsets
9678    use NDBM_File;
9679    tie(my %HIST, 'NDBM_File', '/usr/lib/news/history', 1, 0);
9680    while (my ($key,$val) = each %HIST) {
9681        print $key, ' = ', unpack('L', $val), "\n";
9682    }
9683
9684A class implementing a hash should have the following methods:
9685
9686    TIEHASH classname, LIST
9687    FETCH this, key
9688    STORE this, key, value
9689    DELETE this, key
9690    CLEAR this
9691    EXISTS this, key
9692    FIRSTKEY this
9693    NEXTKEY this, lastkey
9694    SCALAR this
9695    DESTROY this
9696    UNTIE this
9697
9698A class implementing an ordinary array should have the following methods:
9699
9700    TIEARRAY classname, LIST
9701    FETCH this, key
9702    STORE this, key, value
9703    FETCHSIZE this
9704    STORESIZE this, count
9705    CLEAR this
9706    PUSH this, LIST
9707    POP this
9708    SHIFT this
9709    UNSHIFT this, LIST
9710    SPLICE this, offset, length, LIST
9711    EXTEND this, count
9712    DELETE this, key
9713    EXISTS this, key
9714    DESTROY this
9715    UNTIE this
9716
9717A class implementing a filehandle should have the following methods:
9718
9719    TIEHANDLE classname, LIST
9720    READ this, scalar, length, offset
9721    READLINE this
9722    GETC this
9723    WRITE this, scalar, length, offset
9724    PRINT this, LIST
9725    PRINTF this, format, LIST
9726    BINMODE this
9727    EOF this
9728    FILENO this
9729    SEEK this, position, whence
9730    TELL this
9731    OPEN this, mode, LIST
9732    CLOSE this
9733    DESTROY this
9734    UNTIE this
9735
9736A class implementing a scalar should have the following methods:
9737
9738    TIESCALAR classname, LIST
9739    FETCH this,
9740    STORE this, value
9741    DESTROY this
9742    UNTIE this
9743
9744Not all methods indicated above need be implemented.  See L<perltie>,
9745L<Tie::Hash>, L<Tie::Array>, L<Tie::Scalar>, and L<Tie::Handle>.
9746
9747Unlike L<C<dbmopen>|/dbmopen HASH,DBNAME,MASK>, the
9748L<C<tie>|/tie VARIABLE,CLASSNAME,LIST> function will not
9749L<C<use>|/use Module VERSION LIST> or L<C<require>|/require VERSION> a
9750module for you; you need to do that explicitly yourself.  See L<DB_File>
9751or the L<Config> module for interesting
9752L<C<tie>|/tie VARIABLE,CLASSNAME,LIST> implementations.
9753
9754For further details see L<perltie>, L<C<tied>|/tied VARIABLE>.
9755
9756=item tied VARIABLE
9757X<tied>
9758
9759=for Pod::Functions get a reference to the object underlying a tied variable
9760
9761Returns a reference to the object underlying VARIABLE (the same value
9762that was originally returned by the
9763L<C<tie>|/tie VARIABLE,CLASSNAME,LIST> call that bound the variable
9764to a package.)  Returns the undefined value if VARIABLE isn't tied to a
9765package.
9766
9767=item time
9768X<time> X<epoch>
9769
9770=for Pod::Functions return number of seconds since 1970
9771
9772Returns the number of non-leap seconds since whatever time the system
9773considers to be the epoch, suitable for feeding to
9774L<C<gmtime>|/gmtime EXPR> and L<C<localtime>|/localtime EXPR>.  On most
9775systems the epoch is 00:00:00 UTC, January 1, 1970;
9776a prominent exception being Mac OS Classic which uses 00:00:00, January 1,
97771904 in the current local time zone for its epoch.
9778
9779For measuring time in better granularity than one second, use the
9780L<Time::HiRes> module from Perl 5.8 onwards (or from CPAN before then), or,
9781if you have L<gettimeofday(2)>, you may be able to use the
9782L<C<syscall>|/syscall NUMBER, LIST> interface of Perl.  See L<perlfaq8>
9783for details.
9784
9785For date and time processing look at the many related modules on CPAN.
9786For a comprehensive date and time representation look at the
9787L<DateTime> module.
9788
9789=item times
9790X<times>
9791
9792=for Pod::Functions return elapsed time for self and child processes
9793
9794Returns a four-element list giving the user and system times in
9795seconds for this process and any exited children of this process.
9796
9797    my ($user,$system,$cuser,$csystem) = times;
9798
9799In scalar context, L<C<times>|/times> returns C<$user>.
9800
9801Children's times are only included for terminated children.
9802
9803Portability issues: L<perlport/times>.
9804
9805=item tr///
9806
9807=for Pod::Functions transliterate a string
9808
9809The transliteration operator.  Same as
9810L<C<yE<sol>E<sol>E<sol>>|/yE<sol>E<sol>E<sol>>.  See
9811L<perlop/"Quote-Like Operators">.
9812
9813=item truncate FILEHANDLE,LENGTH
9814X<truncate>
9815
9816=item truncate EXPR,LENGTH
9817
9818=for Pod::Functions shorten a file
9819
9820Truncates the file opened on FILEHANDLE, or named by EXPR, to the
9821specified length.  Raises an exception if truncate isn't implemented
9822on your system.  Returns true if successful, L<C<undef>|/undef EXPR> on
9823error.
9824
9825The behavior is undefined if LENGTH is greater than the length of the
9826file.
9827
9828The position in the file of FILEHANDLE is left unchanged.  You may want to
9829call L<seek|/"seek FILEHANDLE,POSITION,WHENCE"> before writing to the
9830file.
9831
9832Portability issues: L<perlport/truncate>.
9833
9834=item uc EXPR
9835X<uc> X<uppercase> X<toupper>
9836
9837=item uc
9838
9839=for Pod::Functions return upper-case version of a string
9840
9841Returns an uppercased version of EXPR.  If EXPR is omitted, uses
9842L<C<$_>|perlvar/$_>.
9843
9844    my $str = uc("Perl is GREAT"); # "PERL IS GREAT"
9845
9846This function behaves the same way under various pragmas, such as in a locale,
9847as L<C<lc>|/lc EXPR> does.
9848
9849If you want titlecase mapping on initial letters see
9850L<C<ucfirst>|/ucfirst EXPR> instead.
9851
9852B<Note:> This is the internal function implementing the
9853L<C<\U>|perlop/"Quote and Quote-like Operators"> escape in double-quoted
9854strings.
9855
9856    my $str = "Perl is \Ugreat\E"; # "Perl is GREAT"
9857
9858=item ucfirst EXPR
9859X<ucfirst> X<uppercase>
9860
9861=item ucfirst
9862
9863=for Pod::Functions return a string with the first letter in upper case
9864
9865Returns the value of EXPR with the B<first> character in uppercase
9866(Unicode calls this titlecase). If EXPR is omitted, C<ucfirst> uses L<C<$_>|perlvar/$_>.
9867
9868    my $str = ucfirst("hello world!"); # "Hello world!"
9869
9870This function behaves the same way under various pragmas, such as in a locale,
9871as L<C<lc>|/lc EXPR> does.
9872
9873B<Note:> This is the internal function implementing the C<\u> escape in
9874double-quoted strings.
9875
9876    my $str = "\uperl\E is great"; # "Perl is great"
9877
9878=item umask EXPR
9879X<umask>
9880
9881=item umask
9882
9883=for Pod::Functions set file creation mode mask
9884
9885Sets the umask for the process to EXPR and returns the previous value.
9886If EXPR is omitted, merely returns the current umask.
9887
9888The Unix permission C<rwxr-x---> is represented as three sets of three
9889bits, or three octal digits: C<0750> (the leading 0 indicates octal
9890and isn't one of the digits).  The L<C<umask>|/umask EXPR> value is such
9891a number representing disabled permissions bits.  The permission (or
9892"mode") values you pass L<C<mkdir>|/mkdir FILENAME,MODE> or
9893L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE> are modified by your
9894umask, so even if you tell
9895L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE> to create a file with
9896permissions C<0777>, if your umask is C<0022>, then the file will
9897actually be created with permissions C<0755>.  If your
9898L<C<umask>|/umask EXPR> were C<0027> (group can't write; others can't
9899read, write, or execute), then passing
9900L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE> C<0666> would create a
9901file with mode C<0640> (because C<0666 &~ 027> is C<0640>).
9902
9903Here's some advice: supply a creation mode of C<0666> for regular
9904files (in L<C<sysopen>|/sysopen FILEHANDLE,FILENAME,MODE>) and one of
9905C<0777> for directories (in L<C<mkdir>|/mkdir FILENAME,MODE>) and
9906executable files.  This gives users the freedom of
9907choice: if they want protected files, they might choose process umasks
9908of C<022>, C<027>, or even the particularly antisocial mask of C<077>.
9909Programs should rarely if ever make policy decisions better left to
9910the user.  The exception to this is when writing files that should be
9911kept private: mail files, web browser cookies, F<.rhosts> files, and
9912so on.
9913
9914If L<umask(2)> is not implemented on your system and you are trying to
9915restrict access for I<yourself> (i.e., C<< (EXPR & 0700) > 0 >>),
9916raises an exception.  If L<umask(2)> is not implemented and you are
9917not trying to restrict access for yourself, returns
9918L<C<undef>|/undef EXPR>.
9919
9920Remember that a umask is a number, usually given in octal; it is I<not> a
9921string of octal digits.  See also L<C<oct>|/oct EXPR>, if all you have
9922is a string.
9923
9924Portability issues: L<perlport/umask>.
9925
9926=item undef EXPR
9927X<undef> X<undefine>
9928
9929=item undef
9930
9931=for Pod::Functions remove a variable or function definition
9932
9933Undefines the value of EXPR, which must be an lvalue.  Use only on a
9934scalar value, an array (using C<@>), a hash (using C<%>), a subroutine
9935(using C<&>), or a typeglob (using C<*>).  Saying C<undef $hash{$key}>
9936will probably not do what you expect on most predefined variables or
9937DBM list values, so don't do that; see L<C<delete>|/delete EXPR>.
9938Always returns the undefined value.
9939You can omit the EXPR, in which case nothing is
9940undefined, but you still get an undefined value that you could, for
9941instance, return from a subroutine, assign to a variable, or pass as a
9942parameter.  Examples:
9943
9944    undef $foo;
9945    undef $bar{'blurfl'};      # Compare to: delete $bar{'blurfl'};
9946    undef @ary;
9947    undef %hash;
9948    undef &mysub;
9949    undef *xyz;       # destroys $xyz, @xyz, %xyz, &xyz, etc.
9950    return (wantarray ? (undef, $errmsg) : undef) if $they_blew_it;
9951    select undef, undef, undef, 0.25;
9952    my ($x, $y, undef, $z) = foo();    # Ignore third value returned
9953
9954Note that this is a unary operator, not a list operator.
9955
9956=item unlink LIST
9957X<unlink> X<delete> X<remove> X<rm> X<del>
9958
9959=item unlink
9960
9961=for Pod::Functions remove one link to a file
9962
9963Deletes a list of files.  On success, it returns the number of files
9964it successfully deleted.  On failure, it returns false and sets
9965L<C<$!>|perlvar/$!> (errno):
9966
9967    my $unlinked = unlink 'a', 'b', 'c';
9968    unlink @goners;
9969    unlink glob "*.bak";
9970
9971On error, L<C<unlink>|/unlink LIST> will not tell you which files it
9972could not remove.
9973If you want to know which files you could not remove, try them one
9974at a time:
9975
9976     foreach my $file ( @goners ) {
9977         unlink $file or warn "Could not unlink $file: $!";
9978     }
9979
9980Note: L<C<unlink>|/unlink LIST> will not attempt to delete directories
9981unless you are
9982superuser and the B<-U> flag is supplied to Perl.  Even if these
9983conditions are met, be warned that unlinking a directory can inflict
9984damage on your filesystem.  Finally, using L<C<unlink>|/unlink LIST> on
9985directories is not supported on many operating systems.  Use
9986L<C<rmdir>|/rmdir FILENAME> instead.
9987
9988If LIST is omitted, L<C<unlink>|/unlink LIST> uses L<C<$_>|perlvar/$_>.
9989
9990=item unpack TEMPLATE,EXPR
9991X<unpack>
9992
9993=item unpack TEMPLATE
9994
9995=for Pod::Functions convert binary structure into normal perl variables
9996
9997L<C<unpack>|/unpack TEMPLATE,EXPR> does the reverse of
9998L<C<pack>|/pack TEMPLATE,LIST>: it takes a string
9999and expands it out into a list of values.
10000(In scalar context, it returns merely the first value produced.)
10001
10002If EXPR is omitted, unpacks the L<C<$_>|perlvar/$_> string.
10003See L<perlpacktut> for an introduction to this function.
10004
10005The string is broken into chunks described by the TEMPLATE.  Each chunk
10006is converted separately to a value.  Typically, either the string is a result
10007of L<C<pack>|/pack TEMPLATE,LIST>, or the characters of the string
10008represent a C structure of some kind.
10009
10010The TEMPLATE has the same format as in the
10011L<C<pack>|/pack TEMPLATE,LIST> function.
10012Here's a subroutine that does substring:
10013
10014    sub substr {
10015        my ($what, $where, $howmuch) = @_;
10016        unpack("x$where a$howmuch", $what);
10017    }
10018
10019and then there's
10020
10021    sub ordinal { unpack("W",$_[0]); } # same as ord()
10022
10023In addition to fields allowed in L<C<pack>|/pack TEMPLATE,LIST>, you may
10024prefix a field with a %<number> to indicate that
10025you want a <number>-bit checksum of the items instead of the items
10026themselves.  Default is a 16-bit checksum.  The checksum is calculated by
10027summing numeric values of expanded values (for string fields the sum of
10028C<ord($char)> is taken; for bit fields the sum of zeroes and ones).
10029
10030For example, the following
10031computes the same number as the System V sum program:
10032
10033    my $checksum = do {
10034        local $/;  # slurp!
10035        unpack("%32W*", readline) % 65535;
10036    };
10037
10038The following efficiently counts the number of set bits in a bit vector:
10039
10040    my $setbits = unpack("%32b*", $selectmask);
10041
10042The C<p> and C<P> formats should be used with care.  Since Perl
10043has no way of checking whether the value passed to
10044L<C<unpack>|/unpack TEMPLATE,EXPR>
10045corresponds to a valid memory location, passing a pointer value that's
10046not known to be valid is likely to have disastrous consequences.
10047
10048If there are more pack codes or if the repeat count of a field or a group
10049is larger than what the remainder of the input string allows, the result
10050is not well defined: the repeat count may be decreased, or
10051L<C<unpack>|/unpack TEMPLATE,EXPR> may produce empty strings or zeros,
10052or it may raise an exception.
10053If the input string is longer than one described by the TEMPLATE,
10054the remainder of that input string is ignored.
10055
10056See L<C<pack>|/pack TEMPLATE,LIST> for more examples and notes.
10057
10058=item unshift ARRAY,LIST
10059X<unshift>
10060
10061=for Pod::Functions prepend more elements to the beginning of a list
10062
10063Add one or more elements to the B<beginning> of an array. This is the
10064opposite of a L<C<shift>|/shift ARRAY>.
10065
10066    my @animals = ("cat");
10067    unshift(@animals, "mouse"); # ("mouse", "cat")
10068
10069    my @colors = ("red");
10070    unshift(@colors, ("blue", "green")); # ("blue", "green", "red")
10071
10072Returns the new number of elements in the updated array.
10073
10074    # Return value is the number of items in the updated array
10075    my $color_count = unshift(@colors, ("yellow", "purple"));
10076
10077    say "There are $color_count colors in the updated array";
10078
10079Note the LIST is prepended whole, not one element at a time, so the
10080prepended elements stay in the same order.  Use
10081L<C<reverse>|/reverse LIST> to do the reverse.
10082
10083Starting with Perl 5.14, an experimental feature allowed
10084L<C<unshift>|/unshift ARRAY,LIST> to take
10085a scalar expression. This experiment has been deemed unsuccessful, and was
10086removed as of Perl 5.24.
10087
10088=item untie VARIABLE
10089X<untie>
10090
10091=for Pod::Functions break a tie binding to a variable
10092
10093Breaks the binding between a variable and a package.
10094(See L<tie|/tie VARIABLE,CLASSNAME,LIST>.)
10095Has no effect if the variable is not tied.
10096
10097=item use Module VERSION LIST
10098X<use> X<module> X<import>
10099
10100=item use Module VERSION
10101
10102=item use Module LIST
10103
10104=item use Module
10105
10106=for Pod::Functions load in a module at compile time and import its namespace
10107
10108Imports some semantics into the current package from the named module,
10109generally by aliasing certain subroutine or variable names into your
10110package.  It is exactly equivalent to
10111
10112    BEGIN { require Module; Module->import( LIST ); }
10113
10114except that Module I<must> be a bareword.
10115The importation can be made conditional by using the L<if> module.
10116
10117The C<BEGIN> forces the L<C<require>|/require VERSION> and
10118L<C<import>|/import LIST> to happen at compile time.  The
10119L<C<require>|/require VERSION> makes sure the module is loaded into
10120memory if it hasn't been yet.  The L<C<import>|/import LIST> is not a
10121builtin; it's just an ordinary static method
10122call into the C<Module> package to tell the module to import the list of
10123features back into the current package.  The module can implement its
10124L<C<import>|/import LIST> method any way it likes, though most modules
10125just choose to derive their L<C<import>|/import LIST> method via
10126inheritance from the C<Exporter> class that is defined in the
10127L<C<Exporter>|Exporter> module.  See L<Exporter>.  If no
10128L<C<import>|/import LIST> method can be found, then the call is skipped,
10129even if there is an AUTOLOAD method.
10130
10131If you do not want to call the package's L<C<import>|/import LIST>
10132method (for instance,
10133to stop your namespace from being altered), explicitly supply the empty list:
10134
10135    use Module ();
10136
10137That is exactly equivalent to
10138
10139    BEGIN { require Module }
10140
10141If the VERSION argument is present between Module and LIST, then the
10142L<C<use>|/use Module VERSION LIST> will call the C<VERSION> method in
10143class Module with the given version as an argument:
10144
10145    use Module 12.34;
10146
10147is equivalent to:
10148
10149    BEGIN { require Module; Module->VERSION(12.34) }
10150
10151The L<default C<VERSION> method|UNIVERSAL/C<VERSION ( [ REQUIRE ] )>>,
10152inherited from the L<C<UNIVERSAL>|UNIVERSAL> class, croaks if the given
10153version is larger than the value of the variable C<$Module::VERSION>.
10154
10155The VERSION argument cannot be an arbitrary expression.  It only counts
10156as a VERSION argument if it is a version number literal, starting with
10157either a digit or C<v> followed by a digit.  Anything that doesn't
10158look like a version literal will be parsed as the start of the LIST.
10159Nevertheless, many attempts to use an arbitrary expression as a VERSION
10160argument will appear to work, because L<Exporter>'s C<import> method
10161handles numeric arguments specially, performing version checks rather
10162than treating them as things to export.
10163
10164Again, there is a distinction between omitting LIST (L<C<import>|/import
10165LIST> called with no arguments) and an explicit empty LIST C<()>
10166(L<C<import>|/import LIST> not called).  Note that there is no comma
10167after VERSION!
10168
10169Because this is a wide-open interface, pragmas (compiler directives)
10170are also implemented this way.  Some of the currently implemented
10171pragmas are:
10172
10173    use constant;
10174    use diagnostics;
10175    use integer;
10176    use sigtrap  qw(SEGV BUS);
10177    use strict   qw(subs vars refs);
10178    use subs     qw(afunc blurfl);
10179    use warnings qw(all);
10180    use sort     qw(stable);
10181
10182Some of these pseudo-modules import semantics into the current
10183block scope (like L<C<strict>|strict> or L<C<integer>|integer>, unlike
10184ordinary modules, which import symbols into the current package (which
10185are effective through the end of the file).
10186
10187Because L<C<use>|/use Module VERSION LIST> takes effect at compile time,
10188it doesn't respect the ordinary flow control of the code being compiled.
10189In particular, putting a L<C<use>|/use Module VERSION LIST> inside the
10190false branch of a conditional doesn't prevent it
10191from being processed.  If a module or pragma only needs to be loaded
10192conditionally, this can be done using the L<if> pragma:
10193
10194    use if $] < 5.008, "utf8";
10195    use if WANT_WARNINGS, warnings => qw(all);
10196
10197There's a corresponding L<C<no>|/no MODULE VERSION LIST> declaration
10198that unimports meanings imported by L<C<use>|/use Module VERSION LIST>,
10199i.e., it calls C<< Module->unimport(LIST) >> instead of
10200L<C<import>|/import LIST>.  It behaves just as L<C<import>|/import LIST>
10201does with VERSION, an omitted or empty LIST,
10202or no unimport method being found.
10203
10204    no integer;
10205    no strict 'refs';
10206    no warnings;
10207
10208See L<perlmodlib> for a list of standard modules and pragmas.  See
10209L<perlrun|perlrun/-m[-]module> for the C<-M> and C<-m> command-line
10210options to Perl that give L<C<use>|/use Module VERSION LIST>
10211functionality from the command-line.
10212
10213=item use VERSION
10214
10215=for Pod::Functions enable Perl language features and declare required version
10216
10217Lexically enables all features available in the requested version as
10218defined by the L<feature> pragma, disabling any features not in the
10219requested version's feature bundle.  See L<feature>.
10220
10221VERSION may be either a v-string such as v5.24.1, which will be compared
10222to L<C<$^V>|perlvar/$^V> (aka $PERL_VERSION), or a numeric argument of the
10223form 5.024001, which will be compared to L<C<$]>|perlvar/$]>.  An
10224exception is raised if VERSION is greater than the version of the current
10225Perl interpreter; Perl will not attempt to parse the rest of the file.
10226Compare with L<C<require>|/require VERSION>, which can do a similar check
10227at run time.
10228
10229If the specified Perl version is 5.12 or higher, strictures are enabled
10230lexically as with L<C<use strict>|strict>.
10231
10232If the specified Perl version is 5.35.0 or higher, L<warnings> are enabled.
10233
10234If the specified Perl version is 5.39.0 or higher, builtin functions are
10235imported lexically as with L<C<use builtin>|builtin> with a corresponding
10236version bundle.
10237
10238Use of C<use VERSION> while another is in effect is not allowed with a
10239C<use v5.39;> or greater version.  For lower versions, C<use VERSION> will
10240override most behavior of a previous C<use VERSION>, possibly removing
10241C<warnings> and C<feature> effects added by it. This behavior is deprecated,
10242and a future release of perl will disallow changing the version once one has
10243been declared.  Additionally, a C<use VERSION> with a version less than 5.11
10244is not allowed after a C<use VERSION> with a version greater than 5.11.
10245
10246C<use VERSION> does not load the F<feature.pm>, F<strict.pm>, F<warnings.pm>
10247or F<builtin.pm> files, but instead implements the equivalent functionality
10248directly.
10249
10250In the current implementation, any explicit use of C<no strict> overrides
10251C<use VERSION>, even if it comes before it. However, this may be subject to
10252change in a future release of Perl, so new code should not rely on this fact.
10253It is recommended that a C<use VERSION> declaration be the first significant
10254statement within a file (possibly after a C<package> statement or any amount
10255of whitespace or comment), so that its effects happen first, and other pragmata
10256are applied after it.
10257
10258Specifying VERSION as a numeric argument of the form 5.024001 should
10259generally be avoided as older less readable syntax compared to
10260v5.24.1. Before perl 5.8.0 released in 2002 the more verbose numeric
10261form was the only supported syntax, which is why you might see it in
10262older code.
10263
10264    use v5.24.1;    # compile time version check
10265    use 5.24.1;     # ditto
10266    use 5.024_001;  # ditto; older syntax compatible with perl 5.6
10267
10268This is often useful if you need to check the current Perl version before
10269L<C<use>|/use Module VERSION LIST>ing library modules that won't work
10270with older versions of Perl.
10271(We try not to do this more than we have to.)
10272
10273Symmetrically, C<no VERSION> allows you to specify that you want a version
10274of Perl older than the specified one.  Historically this was added during
10275early designs of the Raku language (formerly "Perl 6"), so that a Perl 5
10276program could begin
10277
10278    no 6;
10279
10280to declare that it is not a Perl 6 program.  As the two languages have
10281different implementations, file naming conventions, and other
10282infrastructure, this feature is now little used in practice and should be
10283avoided in newly-written code.
10284
10285Care should be taken when using the C<no VERSION> form, as it is I<only>
10286meant to be used to assert that the running Perl is of a earlier version
10287than its argument and I<not> to undo the feature-enabling side effects
10288of C<use VERSION>.
10289
10290=item utime LIST
10291X<utime>
10292
10293=for Pod::Functions set a file's last access and modify times
10294
10295Changes the access and modification times on each file of a list of
10296files.  The first two elements of the list must be the NUMERIC access
10297and modification times, in that order.  Returns the number of files
10298successfully changed.  The inode change time of each file is set
10299to the current time.  For example, this code has the same effect as the
10300Unix L<touch(1)> command when the files I<already exist> and belong to
10301the user running the program:
10302
10303    #!/usr/bin/perl
10304    my $atime = my $mtime = time;
10305    utime $atime, $mtime, @ARGV;
10306
10307Since Perl 5.8.0, if the first two elements of the list are
10308L<C<undef>|/undef EXPR>,
10309the L<utime(2)> syscall from your C library is called with a null second
10310argument.  On most systems, this will set the file's access and
10311modification times to the current time (i.e., equivalent to the example
10312above) and will work even on files you don't own provided you have write
10313permission:
10314
10315    for my $file (@ARGV) {
10316	utime(undef, undef, $file)
10317	    || warn "Couldn't touch $file: $!";
10318    }
10319
10320Under NFS this will use the time of the NFS server, not the time of
10321the local machine.  If there is a time synchronization problem, the
10322NFS server and local machine will have different times.  The Unix
10323L<touch(1)> command will in fact normally use this form instead of the
10324one shown in the first example.
10325
10326Passing only one of the first two elements as L<C<undef>|/undef EXPR> is
10327equivalent to passing a 0 and will not have the effect described when
10328both are L<C<undef>|/undef EXPR>.  This also triggers an
10329uninitialized warning.
10330
10331On systems that support L<futimes(2)>, you may pass filehandles among the
10332files.  On systems that don't support L<futimes(2)>, passing filehandles raises
10333an exception.  Filehandles must be passed as globs or glob references to be
10334recognized; barewords are considered filenames.
10335
10336Portability issues: L<perlport/utime>.
10337
10338=item values HASH
10339X<values>
10340
10341=item values ARRAY
10342
10343=for Pod::Functions return a list of the values in a hash
10344
10345In list context, returns a list consisting of all the values of the named
10346hash.  In Perl 5.12 or later only, will also return a list of the values of
10347an array; prior to that release, attempting to use an array argument will
10348produce a syntax error.  In scalar context, returns the number of values.
10349
10350Hash entries are returned in an apparently random order.  The actual random
10351order is specific to a given hash; the exact same series of operations
10352on two hashes may result in a different order for each hash.  Any insertion
10353into the hash may change the order, as will any deletion, with the exception
10354that the most recent key returned by L<C<each>|/each HASH> or
10355L<C<keys>|/keys HASH> may be deleted without changing the order.  So
10356long as a given hash is unmodified you may rely on
10357L<C<keys>|/keys HASH>, L<C<values>|/values HASH> and
10358L<C<each>|/each HASH> to repeatedly return the same order
10359as each other.  See L<perlsec/"Algorithmic Complexity Attacks"> for
10360details on why hash order is randomized.  Aside from the guarantees
10361provided here the exact details of Perl's hash algorithm and the hash
10362traversal order are subject to change in any release of Perl.  Tied hashes
10363may behave differently to Perl's hashes with respect to changes in order on
10364insertion and deletion of items.
10365
10366As a side effect, calling L<C<values>|/values HASH> resets the HASH or
10367ARRAY's internal iterator (see L<C<each>|/each HASH>) before yielding the
10368values.  In particular,
10369calling L<C<values>|/values HASH> in void context resets the iterator
10370with no other overhead.
10371
10372Apart from resetting the iterator,
10373C<values @array> in list context is the same as plain C<@array>.
10374(We recommend that you use void context C<keys @array> for this, but
10375reasoned that taking C<values @array> out would require more
10376documentation than leaving it in.)
10377
10378Note that the values are not copied, which means modifying them will
10379modify the contents of the hash:
10380
10381    for (values %hash)      { s/foo/bar/g }  # modifies %hash values
10382    for (@hash{keys %hash}) { s/foo/bar/g }  # same
10383
10384Starting with Perl 5.14, an experimental feature allowed
10385L<C<values>|/values HASH> to take a
10386scalar expression. This experiment has been deemed unsuccessful, and was
10387removed as of Perl 5.24.
10388
10389To avoid confusing would-be users of your code who are running earlier
10390versions of Perl with mysterious syntax errors, put this sort of thing at
10391the top of your file to signal that your code will work I<only> on Perls of
10392a recent vintage:
10393
10394    use v5.12;	# so keys/values/each work on arrays
10395
10396See also L<C<keys>|/keys HASH>, L<C<each>|/each HASH>, and
10397L<C<sort>|/sort SUBNAME LIST>.
10398
10399=item vec EXPR,OFFSET,BITS
10400X<vec> X<bit> X<bit vector>
10401
10402=for Pod::Functions test or set particular bits in a string
10403
10404Treats the string in EXPR as a bit vector made up of elements of
10405width BITS and returns the value of the element specified by OFFSET
10406as an unsigned integer.  BITS therefore specifies the number of bits
10407that are reserved for each element in the bit vector.  This must
10408be a power of two from 1 to 32 (or 64, if your platform supports
10409that).
10410
10411If BITS is 8, "elements" coincide with bytes of the input string.
10412
10413If BITS is 16 or more, bytes of the input string are grouped into chunks
10414of size BITS/8, and each group is converted to a number as with
10415L<C<pack>|/pack TEMPLATE,LIST>/L<C<unpack>|/unpack TEMPLATE,EXPR> with
10416big-endian formats C<n>/C<N> (and analogously for BITS==64).  See
10417L<C<pack>|/pack TEMPLATE,LIST> for details.
10418
10419If bits is 4 or less, the string is broken into bytes, then the bits
10420of each byte are broken into 8/BITS groups.  Bits of a byte are
10421numbered in a little-endian-ish way, as in C<0x01>, C<0x02>,
10422C<0x04>, C<0x08>, C<0x10>, C<0x20>, C<0x40>, C<0x80>.  For example,
10423breaking the single input byte C<chr(0x36)> into two groups gives a list
10424C<(0x6, 0x3)>; breaking it into 4 groups gives C<(0x2, 0x1, 0x3, 0x0)>.
10425
10426L<C<vec>|/vec EXPR,OFFSET,BITS> may also be assigned to, in which case
10427parentheses are needed
10428to give the expression the correct precedence as in
10429
10430    vec($image, $max_x * $x + $y, 8) = 3;
10431
10432If the selected element is outside the string, the value 0 is returned.
10433If an element off the end of the string is written to, Perl will first
10434extend the string with sufficiently many zero bytes.   It is an error
10435to try to write off the beginning of the string (i.e., negative OFFSET).
10436
10437If the string happens to be encoded as UTF-8 internally (and thus has
10438the UTF8 flag set), L<C<vec>|/vec EXPR,OFFSET,BITS> tries to convert it
10439to use a one-byte-per-character internal representation. However, if the
10440string contains characters with values of 256 or higher, a fatal error
10441will occur.
10442
10443Strings created with L<C<vec>|/vec EXPR,OFFSET,BITS> can also be
10444manipulated with the logical
10445operators C<|>, C<&>, C<^>, and C<~>.  These operators will assume a bit
10446vector operation is desired when both operands are strings.
10447See L<perlop/"Bitwise String Operators">.
10448
10449The following code will build up an ASCII string saying C<'PerlPerlPerl'>.
10450The comments show the string after each step.  Note that this code works
10451in the same way on big-endian or little-endian machines.
10452
10453    my $foo = '';
10454    vec($foo,  0, 32) = 0x5065726C; # 'Perl'
10455
10456    # $foo eq "Perl" eq "\x50\x65\x72\x6C", 32 bits
10457    print vec($foo, 0, 8);  # prints 80 == 0x50 == ord('P')
10458
10459    vec($foo,  2, 16) = 0x5065; # 'PerlPe'
10460    vec($foo,  3, 16) = 0x726C; # 'PerlPerl'
10461    vec($foo,  8,  8) = 0x50;   # 'PerlPerlP'
10462    vec($foo,  9,  8) = 0x65;   # 'PerlPerlPe'
10463    vec($foo, 20,  4) = 2;      # 'PerlPerlPe'   . "\x02"
10464    vec($foo, 21,  4) = 7;      # 'PerlPerlPer'
10465                                   # 'r' is "\x72"
10466    vec($foo, 45,  2) = 3;      # 'PerlPerlPer'  . "\x0c"
10467    vec($foo, 93,  1) = 1;      # 'PerlPerlPer'  . "\x2c"
10468    vec($foo, 94,  1) = 1;      # 'PerlPerlPerl'
10469                                   # 'l' is "\x6c"
10470
10471To transform a bit vector into a string or list of 0's and 1's, use these:
10472
10473    my $bits = unpack("b*", $vector);
10474    my @bits = split(//, unpack("b*", $vector));
10475
10476If you know the exact length in bits, it can be used in place of the C<*>.
10477
10478Here is an example to illustrate how the bits actually fall in place:
10479
10480  #!/usr/bin/perl -wl
10481
10482  print <<'EOT';
10483                                    0         1         2         3
10484                     unpack("V",$_) 01234567890123456789012345678901
10485  ------------------------------------------------------------------
10486  EOT
10487
10488  for $w (0..3) {
10489      $width = 2**$w;
10490      for ($shift=0; $shift < $width; ++$shift) {
10491          for ($off=0; $off < 32/$width; ++$off) {
10492              $str = pack("B*", "0"x32);
10493              $bits = (1<<$shift);
10494              vec($str, $off, $width) = $bits;
10495              $res = unpack("b*",$str);
10496              $val = unpack("V", $str);
10497              write;
10498          }
10499      }
10500  }
10501
10502  format STDOUT =
10503  vec($_,@#,@#) = @<< == @######### @>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
10504  $off, $width, $bits, $val, $res
10505  .
10506  __END__
10507
10508Regardless of the machine architecture on which it runs, the
10509example above should print the following table:
10510
10511                                    0         1         2         3
10512                     unpack("V",$_) 01234567890123456789012345678901
10513  ------------------------------------------------------------------
10514  vec($_, 0, 1) = 1   ==          1 10000000000000000000000000000000
10515  vec($_, 1, 1) = 1   ==          2 01000000000000000000000000000000
10516  vec($_, 2, 1) = 1   ==          4 00100000000000000000000000000000
10517  vec($_, 3, 1) = 1   ==          8 00010000000000000000000000000000
10518  vec($_, 4, 1) = 1   ==         16 00001000000000000000000000000000
10519  vec($_, 5, 1) = 1   ==         32 00000100000000000000000000000000
10520  vec($_, 6, 1) = 1   ==         64 00000010000000000000000000000000
10521  vec($_, 7, 1) = 1   ==        128 00000001000000000000000000000000
10522  vec($_, 8, 1) = 1   ==        256 00000000100000000000000000000000
10523  vec($_, 9, 1) = 1   ==        512 00000000010000000000000000000000
10524  vec($_,10, 1) = 1   ==       1024 00000000001000000000000000000000
10525  vec($_,11, 1) = 1   ==       2048 00000000000100000000000000000000
10526  vec($_,12, 1) = 1   ==       4096 00000000000010000000000000000000
10527  vec($_,13, 1) = 1   ==       8192 00000000000001000000000000000000
10528  vec($_,14, 1) = 1   ==      16384 00000000000000100000000000000000
10529  vec($_,15, 1) = 1   ==      32768 00000000000000010000000000000000
10530  vec($_,16, 1) = 1   ==      65536 00000000000000001000000000000000
10531  vec($_,17, 1) = 1   ==     131072 00000000000000000100000000000000
10532  vec($_,18, 1) = 1   ==     262144 00000000000000000010000000000000
10533  vec($_,19, 1) = 1   ==     524288 00000000000000000001000000000000
10534  vec($_,20, 1) = 1   ==    1048576 00000000000000000000100000000000
10535  vec($_,21, 1) = 1   ==    2097152 00000000000000000000010000000000
10536  vec($_,22, 1) = 1   ==    4194304 00000000000000000000001000000000
10537  vec($_,23, 1) = 1   ==    8388608 00000000000000000000000100000000
10538  vec($_,24, 1) = 1   ==   16777216 00000000000000000000000010000000
10539  vec($_,25, 1) = 1   ==   33554432 00000000000000000000000001000000
10540  vec($_,26, 1) = 1   ==   67108864 00000000000000000000000000100000
10541  vec($_,27, 1) = 1   ==  134217728 00000000000000000000000000010000
10542  vec($_,28, 1) = 1   ==  268435456 00000000000000000000000000001000
10543  vec($_,29, 1) = 1   ==  536870912 00000000000000000000000000000100
10544  vec($_,30, 1) = 1   == 1073741824 00000000000000000000000000000010
10545  vec($_,31, 1) = 1   == 2147483648 00000000000000000000000000000001
10546  vec($_, 0, 2) = 1   ==          1 10000000000000000000000000000000
10547  vec($_, 1, 2) = 1   ==          4 00100000000000000000000000000000
10548  vec($_, 2, 2) = 1   ==         16 00001000000000000000000000000000
10549  vec($_, 3, 2) = 1   ==         64 00000010000000000000000000000000
10550  vec($_, 4, 2) = 1   ==        256 00000000100000000000000000000000
10551  vec($_, 5, 2) = 1   ==       1024 00000000001000000000000000000000
10552  vec($_, 6, 2) = 1   ==       4096 00000000000010000000000000000000
10553  vec($_, 7, 2) = 1   ==      16384 00000000000000100000000000000000
10554  vec($_, 8, 2) = 1   ==      65536 00000000000000001000000000000000
10555  vec($_, 9, 2) = 1   ==     262144 00000000000000000010000000000000
10556  vec($_,10, 2) = 1   ==    1048576 00000000000000000000100000000000
10557  vec($_,11, 2) = 1   ==    4194304 00000000000000000000001000000000
10558  vec($_,12, 2) = 1   ==   16777216 00000000000000000000000010000000
10559  vec($_,13, 2) = 1   ==   67108864 00000000000000000000000000100000
10560  vec($_,14, 2) = 1   ==  268435456 00000000000000000000000000001000
10561  vec($_,15, 2) = 1   == 1073741824 00000000000000000000000000000010
10562  vec($_, 0, 2) = 2   ==          2 01000000000000000000000000000000
10563  vec($_, 1, 2) = 2   ==          8 00010000000000000000000000000000
10564  vec($_, 2, 2) = 2   ==         32 00000100000000000000000000000000
10565  vec($_, 3, 2) = 2   ==        128 00000001000000000000000000000000
10566  vec($_, 4, 2) = 2   ==        512 00000000010000000000000000000000
10567  vec($_, 5, 2) = 2   ==       2048 00000000000100000000000000000000
10568  vec($_, 6, 2) = 2   ==       8192 00000000000001000000000000000000
10569  vec($_, 7, 2) = 2   ==      32768 00000000000000010000000000000000
10570  vec($_, 8, 2) = 2   ==     131072 00000000000000000100000000000000
10571  vec($_, 9, 2) = 2   ==     524288 00000000000000000001000000000000
10572  vec($_,10, 2) = 2   ==    2097152 00000000000000000000010000000000
10573  vec($_,11, 2) = 2   ==    8388608 00000000000000000000000100000000
10574  vec($_,12, 2) = 2   ==   33554432 00000000000000000000000001000000
10575  vec($_,13, 2) = 2   ==  134217728 00000000000000000000000000010000
10576  vec($_,14, 2) = 2   ==  536870912 00000000000000000000000000000100
10577  vec($_,15, 2) = 2   == 2147483648 00000000000000000000000000000001
10578  vec($_, 0, 4) = 1   ==          1 10000000000000000000000000000000
10579  vec($_, 1, 4) = 1   ==         16 00001000000000000000000000000000
10580  vec($_, 2, 4) = 1   ==        256 00000000100000000000000000000000
10581  vec($_, 3, 4) = 1   ==       4096 00000000000010000000000000000000
10582  vec($_, 4, 4) = 1   ==      65536 00000000000000001000000000000000
10583  vec($_, 5, 4) = 1   ==    1048576 00000000000000000000100000000000
10584  vec($_, 6, 4) = 1   ==   16777216 00000000000000000000000010000000
10585  vec($_, 7, 4) = 1   ==  268435456 00000000000000000000000000001000
10586  vec($_, 0, 4) = 2   ==          2 01000000000000000000000000000000
10587  vec($_, 1, 4) = 2   ==         32 00000100000000000000000000000000
10588  vec($_, 2, 4) = 2   ==        512 00000000010000000000000000000000
10589  vec($_, 3, 4) = 2   ==       8192 00000000000001000000000000000000
10590  vec($_, 4, 4) = 2   ==     131072 00000000000000000100000000000000
10591  vec($_, 5, 4) = 2   ==    2097152 00000000000000000000010000000000
10592  vec($_, 6, 4) = 2   ==   33554432 00000000000000000000000001000000
10593  vec($_, 7, 4) = 2   ==  536870912 00000000000000000000000000000100
10594  vec($_, 0, 4) = 4   ==          4 00100000000000000000000000000000
10595  vec($_, 1, 4) = 4   ==         64 00000010000000000000000000000000
10596  vec($_, 2, 4) = 4   ==       1024 00000000001000000000000000000000
10597  vec($_, 3, 4) = 4   ==      16384 00000000000000100000000000000000
10598  vec($_, 4, 4) = 4   ==     262144 00000000000000000010000000000000
10599  vec($_, 5, 4) = 4   ==    4194304 00000000000000000000001000000000
10600  vec($_, 6, 4) = 4   ==   67108864 00000000000000000000000000100000
10601  vec($_, 7, 4) = 4   == 1073741824 00000000000000000000000000000010
10602  vec($_, 0, 4) = 8   ==          8 00010000000000000000000000000000
10603  vec($_, 1, 4) = 8   ==        128 00000001000000000000000000000000
10604  vec($_, 2, 4) = 8   ==       2048 00000000000100000000000000000000
10605  vec($_, 3, 4) = 8   ==      32768 00000000000000010000000000000000
10606  vec($_, 4, 4) = 8   ==     524288 00000000000000000001000000000000
10607  vec($_, 5, 4) = 8   ==    8388608 00000000000000000000000100000000
10608  vec($_, 6, 4) = 8   ==  134217728 00000000000000000000000000010000
10609  vec($_, 7, 4) = 8   == 2147483648 00000000000000000000000000000001
10610  vec($_, 0, 8) = 1   ==          1 10000000000000000000000000000000
10611  vec($_, 1, 8) = 1   ==        256 00000000100000000000000000000000
10612  vec($_, 2, 8) = 1   ==      65536 00000000000000001000000000000000
10613  vec($_, 3, 8) = 1   ==   16777216 00000000000000000000000010000000
10614  vec($_, 0, 8) = 2   ==          2 01000000000000000000000000000000
10615  vec($_, 1, 8) = 2   ==        512 00000000010000000000000000000000
10616  vec($_, 2, 8) = 2   ==     131072 00000000000000000100000000000000
10617  vec($_, 3, 8) = 2   ==   33554432 00000000000000000000000001000000
10618  vec($_, 0, 8) = 4   ==          4 00100000000000000000000000000000
10619  vec($_, 1, 8) = 4   ==       1024 00000000001000000000000000000000
10620  vec($_, 2, 8) = 4   ==     262144 00000000000000000010000000000000
10621  vec($_, 3, 8) = 4   ==   67108864 00000000000000000000000000100000
10622  vec($_, 0, 8) = 8   ==          8 00010000000000000000000000000000
10623  vec($_, 1, 8) = 8   ==       2048 00000000000100000000000000000000
10624  vec($_, 2, 8) = 8   ==     524288 00000000000000000001000000000000
10625  vec($_, 3, 8) = 8   ==  134217728 00000000000000000000000000010000
10626  vec($_, 0, 8) = 16  ==         16 00001000000000000000000000000000
10627  vec($_, 1, 8) = 16  ==       4096 00000000000010000000000000000000
10628  vec($_, 2, 8) = 16  ==    1048576 00000000000000000000100000000000
10629  vec($_, 3, 8) = 16  ==  268435456 00000000000000000000000000001000
10630  vec($_, 0, 8) = 32  ==         32 00000100000000000000000000000000
10631  vec($_, 1, 8) = 32  ==       8192 00000000000001000000000000000000
10632  vec($_, 2, 8) = 32  ==    2097152 00000000000000000000010000000000
10633  vec($_, 3, 8) = 32  ==  536870912 00000000000000000000000000000100
10634  vec($_, 0, 8) = 64  ==         64 00000010000000000000000000000000
10635  vec($_, 1, 8) = 64  ==      16384 00000000000000100000000000000000
10636  vec($_, 2, 8) = 64  ==    4194304 00000000000000000000001000000000
10637  vec($_, 3, 8) = 64  == 1073741824 00000000000000000000000000000010
10638  vec($_, 0, 8) = 128 ==        128 00000001000000000000000000000000
10639  vec($_, 1, 8) = 128 ==      32768 00000000000000010000000000000000
10640  vec($_, 2, 8) = 128 ==    8388608 00000000000000000000000100000000
10641  vec($_, 3, 8) = 128 == 2147483648 00000000000000000000000000000001
10642
10643=item wait
10644X<wait>
10645
10646=for Pod::Functions wait for any child process to die
10647
10648Behaves like L<wait(2)> on your system: it waits for a child
10649process to terminate and returns the pid of the deceased process, or
10650C<-1> if there are no child processes.  The status is returned in
10651L<C<$?>|perlvar/$?> and
10652L<C<${^CHILD_ERROR_NATIVE}>|perlvar/${^CHILD_ERROR_NATIVE}>.
10653Note that a return value of C<-1> could mean that child processes are
10654being automatically reaped, as described in L<perlipc>.
10655
10656If you use L<C<wait>|/wait> in your handler for
10657L<C<$SIG{CHLD}>|perlvar/%SIG>, it may accidentally wait for the child
10658created by L<C<qx>|/qxE<sol>STRINGE<sol>> or L<C<system>|/system LIST>.
10659See L<perlipc> for details.
10660
10661Equivalent to L<C<waitpid(-1, 0)>|/waitpid PID,FLAGS>.
10662
10663Portability issues: L<perlport/wait>.
10664
10665=item waitpid PID,FLAGS
10666X<waitpid>
10667
10668=for Pod::Functions wait for a particular child process to die
10669
10670Waits for a particular child process to terminate and returns the pid of
10671the deceased process, or C<-1> if there is no such child process.  A
10672non-blocking wait (with L<WNOHANG|POSIX/C<WNOHANG>> in FLAGS) can return 0 if
10673there are child processes matching PID but none have terminated yet.
10674The status is returned in L<C<$?>|perlvar/$?> and
10675L<C<${^CHILD_ERROR_NATIVE}>|perlvar/${^CHILD_ERROR_NATIVE}>.
10676
10677A PID of C<0> indicates to wait for any child process whose process group ID is
10678equal to that of the current process.  A PID of less than C<-1> indicates to
10679wait for any child process whose process group ID is equal to -PID.  A PID of
10680C<-1> indicates to wait for any child process.
10681
10682If you say
10683
10684    use POSIX ":sys_wait_h";
10685
10686    my $kid;
10687    do {
10688        $kid = waitpid(-1, WNOHANG);
10689    } while $kid > 0;
10690
10691or
10692
10693    1 while waitpid(-1, WNOHANG) > 0;
10694
10695then you can do a non-blocking wait for all pending zombie processes (see
10696L<POSIX/WAIT>).
10697Non-blocking wait is available on machines supporting either the
10698L<waitpid(2)> or L<wait4(2)> syscalls.  However, waiting for a particular
10699pid with FLAGS of C<0> is implemented everywhere.  (Perl emulates the
10700system call by remembering the status values of processes that have
10701exited but have not been harvested by the Perl script yet.)
10702
10703Note that on some systems, a return value of C<-1> could mean that child
10704processes are being automatically reaped.  See L<perlipc> for details,
10705and for other examples.
10706
10707Portability issues: L<perlport/waitpid>.
10708
10709=item wantarray
10710X<wantarray> X<context>
10711
10712=for Pod::Functions get void vs scalar vs list context of current subroutine call
10713
10714Returns true if the context of the currently executing subroutine or
10715L<C<eval>|/eval EXPR> is looking for a list value.  Returns false if the
10716context is
10717looking for a scalar.  Returns the undefined value if the context is
10718looking for no value (void context).
10719
10720    return unless defined wantarray; # don't bother doing more
10721    my @a = complex_calculation();
10722    return wantarray ? @a : "@a";
10723
10724L<C<wantarray>|/wantarray>'s result is unspecified in the top level of a file,
10725in a C<BEGIN>, C<UNITCHECK>, C<CHECK>, C<INIT> or C<END> block, or
10726in a C<DESTROY> method.
10727
10728This function should have been named wantlist() instead.
10729
10730=item warn LIST
10731X<warn> X<warning> X<STDERR>
10732
10733=for Pod::Functions print debugging info
10734
10735Emits a warning, usually by printing it to C<STDERR>.  C<warn> interprets
10736its operand LIST in the same way as C<die>, but is slightly different
10737in what it defaults to when LIST is empty or makes an empty string.
10738If it is empty and L<C<$@>|perlvar/$@> already contains an exception
10739value then that value is used after appending C<"\t...caught">.  If it
10740is empty and C<$@> is also empty then the string C<"Warning: Something's
10741wrong"> is used.
10742
10743By default, the exception derived from the operand LIST is stringified
10744and printed to C<STDERR>.  This behaviour can be altered by installing
10745a L<C<$SIG{__WARN__}>|perlvar/%SIG> handler.  If there is such a
10746handler then no message is automatically printed; it is the handler's
10747responsibility to deal with the exception
10748as it sees fit (like, for instance, converting it into a
10749L<C<die>|/die LIST>).  Most
10750handlers must therefore arrange to actually display the
10751warnings that they are not prepared to deal with, by calling
10752L<C<warn>|/warn LIST>
10753again in the handler.  Note that this is quite safe and will not
10754produce an endless loop, since C<__WARN__> hooks are not called from
10755inside one.
10756
10757You will find this behavior is slightly different from that of
10758L<C<$SIG{__DIE__}>|perlvar/%SIG> handlers (which don't suppress the
10759error text, but can instead call L<C<die>|/die LIST> again to change
10760it).
10761
10762Using a C<__WARN__> handler provides a powerful way to silence all
10763warnings (even the so-called mandatory ones).  An example:
10764
10765    # wipe out *all* compile-time warnings
10766    BEGIN { $SIG{'__WARN__'} = sub { warn $_[0] if $DOWARN } }
10767    my $foo = 10;
10768    my $foo = 20;          # no warning about duplicate my $foo,
10769                           # but hey, you asked for it!
10770    # no compile-time or run-time warnings before here
10771    $DOWARN = 1;
10772
10773    # run-time warnings enabled after here
10774    warn "\$foo is alive and $foo!";     # does show up
10775
10776See L<perlvar> for details on setting L<C<%SIG>|perlvar/%SIG> entries
10777and for more
10778examples.  See the L<Carp> module for other kinds of warnings using its
10779C<carp> and C<cluck> functions.
10780
10781=item write FILEHANDLE
10782X<write>
10783
10784=item write EXPR
10785
10786=item write
10787
10788=for Pod::Functions print a picture record
10789
10790Writes a formatted record (possibly multi-line) to the specified FILEHANDLE,
10791using the format associated with that file.  By default the format for
10792a file is the one having the same name as the filehandle, but the
10793format for the current output channel (see the
10794L<C<select>|/select FILEHANDLE> function) may be set explicitly by
10795assigning the name of the format to the L<C<$~>|perlvar/$~> variable.
10796
10797Top of form processing is handled automatically:  if there is insufficient
10798room on the current page for the formatted record, the page is advanced by
10799writing a form feed and a special top-of-page
10800format is used to format the new
10801page header before the record is written.  By default, the top-of-page
10802format is the name of the filehandle with C<_TOP> appended, or C<top>
10803in the current package if the former does not exist.  This would be a
10804problem with autovivified filehandles, but it may be dynamically set to the
10805format of your choice by assigning the name to the L<C<$^>|perlvar/$^>
10806variable while that filehandle is selected.  The number of lines
10807remaining on the current page is in variable L<C<$->|perlvar/$->, which
10808can be set to C<0> to force a new page.
10809
10810If FILEHANDLE is unspecified, output goes to the current default output
10811channel, which starts out as STDOUT but may be changed by the
10812L<C<select>|/select FILEHANDLE> operator.  If the FILEHANDLE is an EXPR,
10813then the expression
10814is evaluated and the resulting string is used to look up the name of
10815the FILEHANDLE at run time.  For more on formats, see L<perlform>.
10816
10817Note that write is I<not> the opposite of
10818L<C<read>|/read FILEHANDLE,SCALAR,LENGTH,OFFSET>.  Unfortunately.
10819
10820=item y///
10821
10822=for Pod::Functions transliterate a string
10823
10824The transliteration operator.  Same as
10825L<C<trE<sol>E<sol>E<sol>>|/trE<sol>E<sol>E<sol>>.  See
10826L<perlop/"Quote-Like Operators">.
10827
10828=back
10829
10830=head2 Non-function Keywords by Cross-reference
10831
10832=head3 perldata
10833
10834=over
10835
10836=item __DATA__
10837
10838=item __END__
10839
10840These keywords are documented in L<perldata/"Special Literals">.
10841
10842=back
10843
10844=head3 perlmod
10845
10846=over
10847
10848=item BEGIN
10849
10850=item CHECK
10851
10852=item END
10853
10854=item INIT
10855
10856=item UNITCHECK
10857
10858These compile phase keywords are documented in L<perlmod/"BEGIN, UNITCHECK, CHECK, INIT and END">.
10859
10860=back
10861
10862=head3 perlobj
10863
10864=over
10865
10866=item DESTROY
10867
10868This method keyword is documented in L<perlobj/"Destructors">.
10869
10870=back
10871
10872=head3 perlop
10873
10874=over
10875
10876=item and
10877
10878=item cmp
10879
10880=item eq
10881
10882=item ge
10883
10884=item gt
10885
10886=item isa
10887
10888=item le
10889
10890=item lt
10891
10892=item ne
10893
10894=item not
10895
10896=item or
10897
10898=item x
10899
10900=item xor
10901
10902These operators are documented in L<perlop>.
10903
10904=back
10905
10906=head3 perlsub
10907
10908=over
10909
10910=item AUTOLOAD
10911
10912This keyword is documented in L<perlsub/"Autoloading">.
10913
10914=back
10915
10916=head3 perlsyn
10917
10918=over
10919
10920=item else
10921
10922=item elsif
10923
10924=item for
10925
10926=item foreach
10927
10928=item if
10929
10930=item unless
10931
10932=item until
10933
10934=item while
10935
10936These flow-control keywords are documented in L<perlsyn/"Compound Statements">.
10937
10938=item elseif
10939
10940The "else if" keyword is spelled C<elsif> in Perl.  There's no C<elif>
10941or C<else if> either.  It does parse C<elseif>, but only to warn you
10942about not using it.
10943
10944See the documentation for flow-control keywords in L<perlsyn/"Compound
10945Statements">.
10946
10947=back
10948
10949=over
10950
10951=item default
10952
10953=item given
10954
10955=item when
10956
10957These flow-control keywords related to the experimental switch feature are
10958documented in L<perlsyn/"Switch Statements">.
10959
10960=back
10961
10962=over
10963
10964=item try
10965
10966=item catch
10967
10968=item finally
10969
10970These flow-control keywords related to the experimental C<try> feature are
10971documented in L<perlsyn/"Try Catch Exception Handling">.
10972
10973=back
10974
10975=over
10976
10977=item defer
10978
10979This flow-control keyword related to the experimental C<defer> feature is
10980documented in L<perlsyn/"defer blocks">.
10981
10982=back
10983
10984=over
10985
10986=item ADJUST
10987
10988This class-related phaser block is documented in L<perlclass>.
10989
10990=back
10991
10992=cut
10993