xref: /openbsd-src/gnu/usr.bin/perl/t/re/regexp.t (revision ae3cb403620ab940fbaabb3055fac045a63d56b7)
1#!./perl
2
3# The tests are in a separate file 't/re/re_tests'.
4# Each line in that file is a separate test.
5# There are five columns, separated by tabs.
6# An optional sixth column is used to give a reason, only when skipping tests
7#
8# Column 1 contains the pattern, optionally enclosed in C<''>.
9# Modifiers can be put after the closing C<'>.
10#
11# Column 2 contains the string to be matched.
12#
13# Column 3 contains the expected result:
14# 	y	expect a match
15# 	n	expect no match
16# 	c	expect an error
17#	T	the test is a TODO (can be combined with y/n/c)
18#	M	skip test on miniperl (combine with y/n/c/T)
19#	B	test exposes a known bug in Perl, should be skipped
20#	b	test exposes a known bug in Perl, should be skipped if noamp
21#	t	test exposes a bug with threading, TODO if qr_embed_thr
22#       s       test should only be run for regex_sets_compat.t
23#       S       test should not be run for regex_sets_compat.t
24#       a       test should only be run on ASCII platforms
25#       e       test should only be run on EBCDIC platforms
26#
27# Columns 4 and 5 are used only if column 3 contains C<y> or C<c>.
28#
29# Column 4 contains a string, usually C<$&>.
30#
31# Column 5 contains the expected result of double-quote
32# interpolating that string after the match, or start of error message.
33#
34# Column 6, if present, contains a reason why the test is skipped.
35# This is printed with "skipped", for harness to pick up.
36#
37# Column 7 can be used for comments
38#
39# \n in the tests are interpolated, as are variables of the form ${\w+}.
40#
41# Blanks lines are treated as PASSING tests to keep the line numbers
42# linked to the test number.
43#
44# If you want to add a regular expression test that can't be expressed
45# in this format, don't add it here: put it in re/pat.t instead.
46#
47# Note that the inputs get passed on as "m're'", so the re bypasses the lexer.
48# This means this file cannot be used for testing anything that the lexer
49# handles; in 5.12 this means just \N{NAME} and \N{U+...}.
50#
51# Note that columns 2,3 and 5 are all enclosed in double quotes and then
52# evalled; so something like a\"\x{100}$1 has length 3+length($1).
53#
54# \x... and \o{...} constants are automatically converted to the native
55# character set if necessary.  \[0-7] constants aren't
56
57my ($file, $iters);
58BEGIN {
59    $iters = shift || 1;	# Poor man performance suite, 10000 is OK.
60
61    # Do this open before any chdir
62    $file = shift;
63    if (defined $file) {
64	open TESTS, $file or die "Can't open $file";
65    }
66
67    chdir 't' if -d 't';
68    @INC = qw '../lib ../ext/re';
69    if (!defined &DynaLoader::boot_DynaLoader) { # miniperl
70	print("1..0 # Skip Unicode tables not built yet\n"), exit
71	    unless eval 'require "unicore/Heavy.pl"';
72    }
73}
74
75sub _comment {
76    return map { /^#/ ? "$_\n" : "# $_\n" }
77           map { split /\n/ } @_;
78}
79
80sub convert_from_ascii {
81    my $string = shift;
82
83    #my $save = $string;
84    # Convert \x{...}, \o{...}
85    $string =~ s/ (?<! \\ ) \\x\{ ( .*? ) } / "\\x{" . sprintf("%X", utf8::unicode_to_native(hex $1)) .  "}" /gex;
86    $string =~ s/ (?<! \\ ) \\o\{ ( .*? ) } / "\\o{" . sprintf("%o", utf8::unicode_to_native(oct $1)) .  "}" /gex;
87
88    # Convert \xAB
89    $string =~ s/ (?<! \\ ) \\x ( [A-Fa-f0-9]{2} ) / "\\x" . sprintf("%02X", utf8::unicode_to_native(hex $1)) /gex;
90
91    # Convert \xA
92    $string =~ s/ (?<! \\ ) \\x ( [A-Fa-f0-9] ) (?! [A-Fa-f0-9] ) / "\\x" . sprintf("%X", utf8::unicode_to_native(hex $1)) /gex;
93
94    #print STDERR __LINE__, ": $save\n$string\n" if $save ne $string;
95    return $string;
96}
97
98use strict;
99use warnings FATAL=>"all";
100use vars qw($bang $ffff $nulnul); # used by the tests
101use vars qw($qr $skip_amp $qr_embed $qr_embed_thr $regex_sets); # set by our callers
102
103
104
105if (!defined $file) {
106    open TESTS, 're/re_tests' or die "Can't open re/re_tests: $!";
107}
108
109my @tests = <TESTS>;
110
111close TESTS;
112
113$bang = sprintf "\\%03o", ord "!"; # \41 would not be portable.
114$ffff  = chr(0xff) x 2;
115$nulnul = "\0" x 2;
116my $OP = $qr ? 'qr' : 'm';
117
118$| = 1;
119printf "1..%d\n# $iters iterations\n", scalar @tests;
120
121my $test;
122TEST:
123foreach (@tests) {
124    $test++;
125    if (!/\S/ || /^\s*#/ || /^__END__$/) {
126        chomp;
127        my ($not,$comment)= split /\s*#\s*/, $_, 2;
128        $comment ||= "(blank line)";
129        print "ok $test # $comment\n";
130        next;
131    }
132    chomp;
133    s/\\n/\n/g unless $regex_sets;
134    my ($pat, $subject, $result, $repl, $expect, $reason, $comment) = split(/\t/,$_,7);
135    if (!defined $subject) {
136        die "Bad test definition on line $test: $_\n";
137    }
138    $reason = '' unless defined $reason;
139    my $input = join(':',$pat,$subject,$result,$repl,$expect);
140
141    # the double '' below keeps simple syntax highlighters from going crazy
142    $pat = "'$pat'" unless $pat =~ /^[:''\/]/;
143    $pat =~ s/(\$\{\w+\})/$1/eeg;
144    $pat =~ s/\\n/\n/g unless $regex_sets;
145    $pat = convert_from_ascii($pat) if ord("A") != 65;
146
147    $subject = convert_from_ascii($subject) if ord("A") != 65;
148    $subject = eval qq("$subject"); die $@ if $@;
149
150    $expect = convert_from_ascii($expect) if ord("A") != 65;
151    $expect  = eval qq("$expect"); die $@ if $@;
152    $expect = $repl = '-' if $skip_amp and $input =~ /\$[&\`\']/;
153
154    my $todo_qr = $qr_embed_thr && ($result =~ s/t//);
155    my $skip = ($skip_amp ? ($result =~ s/B//i) : ($result =~ s/B//));
156    ++$skip if $result =~ s/M// && !defined &DynaLoader::boot_DynaLoader;
157    if ($result =~ s/ ( [Ss] ) //x) {
158        if (($1 eq 'S' && $regex_sets) || ($1 eq 's' && ! $regex_sets)) {
159            $skip++;
160            $reason = "Test not valid for $0";
161        }
162    }
163    if ($result =~ s/a// && ord("A") != 65) {
164        $skip++;
165        $reason = "Test is only valid for ASCII platforms.  $reason";
166    }
167    if ($result =~ s/e// && ord("A") != 193) {
168        $skip++;
169        $reason = "Test is only valid for EBCDIC platforms.  $reason";
170    }
171    $reason = 'skipping $&' if $reason eq  '' && $skip_amp;
172    $result =~ s/B//i unless $skip;
173    my $todo= $result =~ s/T// ? " # TODO" : "";
174    my $testname= $test;
175    if ($comment) {
176        $comment=~s/^\s*(?:#\s*)?//;
177        $testname .= " - $comment" if $comment;
178    }
179    if (! $skip && $regex_sets) {
180
181        # If testing regex sets, change the [bracketed] classes into
182        # (?[bracketed]).  But note that '\[' and '\c[' don't introduce such a
183        # class.  (We don't bother looking for an odd number of backslashes,
184        # as this hasn't been needed so far.)
185        if ($pat !~ / (?<!\\c) (?<!\\) \[ /x) {
186            $skip++;
187            $reason = "Pattern doesn't contain [brackets]";
188        }
189        else { # Use non-regex features of Perl to accomplish this.
190            my $modified = "";
191            my $in_brackets = 0;
192
193            # Go through the pattern character-by-character.  We also add
194            # blanks around each token to test the /x parts of (?[ ])
195            my $pat_len = length($pat);
196      CHAR: for (my $i = 0; $i < $pat_len; $i++) {
197                my $curchar = substr($pat, $i, 1);
198                if ($curchar eq '\\') {
199                    $modified .= " " if $in_brackets;
200                    $modified .= $curchar;
201                    $i++;
202
203                    # Get the character the backslash is escaping
204                    $curchar = substr($pat, $i, 1);
205                    $modified .= $curchar;
206
207                    # If the character following that is a '{}', treat the
208                    # entire amount as a single token
209                    if ($i < $pat_len -1 && substr($pat, $i+1, 1) eq '{') {
210                        my $j = index($pat, '}', $i+2);
211                        if ($j < 0) {
212                            last unless $in_brackets;
213                            if ($result eq 'c') {
214                                $skip++;
215                                $reason = "Can't handle compilation errors with unmatched '{'";
216                            }
217                            else {
218                                print "not ok $testname # Problem in $0; original = '$pat'; mod = '$modified'\n";
219                                next TEST;
220                            }
221                        }
222                        $modified .= substr($pat, $i+1, $j - $i);
223                        $i = $j;
224                    }
225                    elsif ($curchar eq 'x') {
226
227                        # \x without brackets is supposed to be followed by 2
228                        # hex digits.  Take up to 2, and then add a blank
229                        # after the last one.  This avoids getting errors from
230                        # (?[ ]) for run-ons, like \xabc
231                        my $j = $i + 1;
232                        for (; $j < $i + 3 && $j < $pat_len; $j++) {
233                            my $curord = ord(substr($pat, $j, 1));
234                            if (!(($curord >= ord("A") && $curord <= ord("F"))
235                                 || ($curord >= ord("a") && $curord <= ord("f"))
236                                 || ($curord >= ord("0") && $curord <= ord("9"))))
237                            {
238                                $j++;
239                                last;
240                            }
241                        }
242                        $j--;
243                        $modified .= substr($pat, $i + 1, $j - $i) . " ";
244                        $i = $j;
245                    }
246                    elsif (ord($curchar) >= ord('0')
247                           && (ord($curchar) <= ord('7')))
248                    {
249                        # Similarly, octal constants have up to 3 digits.
250                        my $j = $i + 1;
251                        for (; $j < $i + 3 && $j < $pat_len; $j++) {
252                            my $curord = ord(substr($pat, $j, 1));
253                            if (! ($curord >= ord("0") &&  $curord <= ord("7"))) {
254                                $j++;
255                                last;
256                            }
257                        }
258                        $j--;
259                        $modified .= substr($pat, $i + 1, $j - $i);
260                        $i = $j;
261                    }
262
263                    next;
264                } # End of processing a backslash sequence
265
266                if (! $in_brackets  # Skip (?{ })
267                    && $curchar eq '('
268                    && $i < $pat_len - 2
269                    && substr($pat, $i+1, 1) eq '?'
270                    && substr($pat, $i+2, 1) eq '{')
271                {
272                    $skip++;
273                    $reason = "Pattern contains '(?{'";
274                    last;
275                }
276
277                # Closing ']'
278                if ($curchar eq ']' && $in_brackets) {
279                    $modified .= " ] ])";
280                    $in_brackets = 0;
281                    next;
282                }
283
284                # A regular character.
285                if ($curchar ne '[') {
286                    if (! $in_brackets) {
287                        $modified .= $curchar;
288                    }
289                    else {
290                        $modified .= " $curchar ";
291                    }
292                    next;
293                }
294
295                # Here is a '['; If not in a bracketed class, treat as the
296                # beginning of one.
297                if (! $in_brackets) {
298                    $in_brackets = 1;
299                    $modified .= "(?[ [ ";
300
301                    # An immediately following ']' or '^]' is not the ending
302                    # of the class, but is to be treated literally.
303                    if ($i < $pat_len - 1
304                        && substr($pat, $i+1, 1) eq ']')
305                    {
306                        $i ++;
307                        $modified .= " ] ";
308                    }
309                    elsif ($i < $pat_len - 2
310                            && substr($pat, $i+1, 1) eq '^'
311                            && substr($pat, $i+2, 1) eq ']')
312                    {
313                        $i += 2;
314                        $modified .= " ^ ] ";
315                    }
316                    next;
317                }
318
319                # Here is a plain '[' within [ ].  Could mean wants to
320                # match a '[', or it could be a posix class that has a
321                # corresponding ']'.  Absorb either
322
323                $modified .= ' [';
324                last if $i >= $pat_len - 1;
325
326                $i++;
327                $curchar = substr($pat, $i, 1);
328                if ($curchar =~ /[:=.]/) {
329                    for (my $j = $i + 1; $j < $pat_len; $j++) {
330                        next unless substr($pat, $j, 1) eq ']';
331                        last if $j - $i < 2;
332                        if (substr($pat, $j - 1, 1) eq $curchar) {
333                            # Here, is a posix class
334                            $modified .= substr($pat, $i, $j - $i + 1) . " ";
335                            $i = $j;
336                            next CHAR;
337                        }
338                    }
339                }
340
341                # Here wasn't a posix class, just process normally
342                $modified .= " $curchar ";
343            }
344
345            if ($in_brackets && ! $skip) {
346                if ($result eq 'c') {
347                    $skip++;
348                    $reason = "Can't figure out where to put the (?[ and ]) since is a compilation error";
349                }
350                else {
351                    print "not ok $testname # Problem in $0; original = '$pat'; mod = '$modified'\n";
352                    next TEST;
353                }
354            }
355
356            # Use our modified pattern instead of the original
357            $pat = $modified;
358        }
359    }
360
361    for my $study ('', 'study $subject', 'utf8::upgrade($subject)',
362		   'utf8::upgrade($subject); study $subject') {
363	# Need to make a copy, else the utf8::upgrade of an already studied
364	# scalar confuses things.
365	my $subject = $subject;
366	my $c = $iters;
367	my ($code, $match, $got);
368        if ($repl eq 'pos') {
369            $code= <<EOFCODE;
370                $study;
371                pos(\$subject)=0;
372                \$match = ( \$subject =~ m${pat}g );
373                \$got = pos(\$subject);
374EOFCODE
375        }
376        elsif ($qr_embed) {
377            $code= <<EOFCODE;
378                my \$RE = qr$pat;
379                $study;
380                \$match = (\$subject =~ /(?:)\$RE(?:)/) while \$c--;
381                \$got = "$repl";
382EOFCODE
383        }
384        elsif ($qr_embed_thr) {
385            $code= <<EOFCODE;
386		# Can't run the match in a subthread, but can do this and
387	 	# clone the pattern the other way.
388                my \$RE = threads->new(sub {qr$pat})->join();
389                $study;
390                \$match = (\$subject =~ /(?:)\$RE(?:)/) while \$c--;
391                \$got = "$repl";
392EOFCODE
393        }
394        else {
395            $code= <<EOFCODE;
396                $study;
397                \$match = (\$subject =~ $OP$pat) while \$c--;
398                \$got = "$repl";
399EOFCODE
400        }
401        $code = "no warnings 'experimental::regex_sets';$code" if $regex_sets;
402        #$code.=qq[\n\$expect="$expect";\n];
403        #use Devel::Peek;
404        #die Dump($code) if $pat=~/\\h/ and $subject=~/\x{A0}/;
405	{
406	    # Probably we should annotate specific tests with which warnings
407	    # categories they're known to trigger, and hence should be
408	    # disabled just for that test
409	    no warnings qw(uninitialized regexp deprecated);
410	    eval $code;
411	}
412	chomp( my $err = $@ );
413	if ( $skip ) {
414	    print "ok $testname # skipped", length($reason) ? ".  $reason" : '', "\n";
415	    next TEST;
416	}
417	elsif ($result eq 'c') {
418	    if ($err !~ m!^\Q$expect!) { print "not ok $testname$todo (compile) $input => '$err'\n"; next TEST }
419	    last;  # no need to study a syntax error
420	}
421	elsif ( $todo_qr ) {
422	    print "not ok $testname # TODO", length($reason) ? " - $reason" : '', "\n";
423	    next TEST;
424	}
425	elsif ($@) {
426	    print "not ok $testname$todo $input => error '$err'\n", _comment("$code\n$@\n"); next TEST;
427	}
428	elsif ($result =~ /^n/) {
429	    if ($match) { print "not ok $testname$todo ($study) $input => false positive\n"; next TEST }
430	}
431	else {
432	    if (!$match || $got ne $expect) {
433	        eval { require Data::Dumper };
434                no warnings "utf8"; # But handle should be utf8
435		if ($@ || !defined &DynaLoader::boot_DynaLoader) {
436		    # Data::Dumper will load on miniperl, but fail when used in
437		    # anger as it tries to load B. I'd prefer to keep the
438		    # regular calls below outside of an eval so that real
439		    # (unknown) failures get spotted, not ignored.
440		    print "not ok $testname$todo ($study) $input => '$got', match=$match\n", _comment("$code\n");
441		}
442		else { # better diagnostics
443		    my $s = Data::Dumper->new([$subject],['subject'])->Useqq(1)->Dump;
444		    my $g = Data::Dumper->new([$got],['got'])->Useqq(1)->Dump;
445		    print "not ok $testname$todo ($study) $input => '$got', match=$match\n", _comment("$s\n$g\n$code\n");
446		}
447		next TEST;
448	    }
449	}
450    }
451    print "ok $testname$todo\n";
452}
453
4541;
455