sed: use warn()/err() where appropriate

Use warn()/err() instead of sed's homegrown warning()/error() for
things other than parser problems. The warning()/error() functions
display the file and lin

sed: use warn()/err() where appropriate

Use warn()/err() instead of sed's homegrown warning()/error() for
things other than parser problems. The warning()/error() functions
display the file and line number in addition to the error message.
This also removes of the COMPILE/FATAL argument to error() since
now all calls to error() are for compilation/parsing issues.
OK op@ espie@

show more ...

Rename mf_fgets() and cu_fgets() to mf_getline() and cu_getline().
These functions now use getline(), not fgets(). From espie@

spelling fixes; from paul tagliamonte
amendments to his diff are noted on tech

As an extension to POSIX, for consistency with our behaviour for
the "b" and "t" commands with a label, and for compatibility with
GNU sed, also accept ";" followed by another command after "b"
and "

As an extension to POSIX, for consistency with our behaviour for
the "b" and "t" commands with a label, and for compatibility with
GNU sed, also accept ";" followed by another command after "b"
and "t" commands without a label: branch to the end of the script
instead of erroring out. Parsing is unchanged.

Missing feature reported by Lars dot Nooden at gmail dot com on bugs@.
OK martijn@ millert@

show more ...

Improve consistency of the substitution command further.

When the opening square bracket ('[') is abused as the delimiter, the regular
expression contains a bracket expression, and the bracket expre

Improve consistency of the substitution command further.

When the opening square bracket ('[') is abused as the delimiter, the regular
expression contains a bracket expression, and the bracket expression contains
another opening square bracket (sick! - i mean, sic!), then do not require
escaping that innermost bracket and treat a preceding backslash as a literal
backslash character, in accordance with POSIX:

$ printf 'x[x\\x\n' | sed 's[\[[][R[g'
xRx\x
$ printf 'x[x\\x\n' | sed 's[\[\[][R[g'
xRxRx

While here, also make the implementation more readable and insert
some comments.

Joint work with martijn@ (started during g2k18) and OK martijn@.

show more ...

When using '[' as the delimiter in sed(1) s/// (don't do that, of course)
and then including '[' in the regular expression by prepending a backslash
to it, remove the backslash before feeding the RE

When using '[' as the delimiter in sed(1) s/// (don't do that, of course)
and then including '[' in the regular expression by prepending a backslash
to it, remove the backslash before feeding the RE to the RE engine, just
like we already do it for other special characters like .^$*+?{(|.
This makes sed 's[\[xy][...[' treat the xy thingy as a char class.
Joint work with martijn@, OK guenther@ martijn@

show more ...

Fix array index by signed char; from martijn@

Fix sign compare warnings; OK martijn@

Fix 2 bugs introduced by previous.
Pointed out by kshe59 <at> zoho <dot> eu

OK millert@

Fix and change y command in the following ways:
- When 'n' is used as a delimiter escaping 'n' will remain a newline
instead of becoming a 'n' character. This is how POSIX specifies how this
shou

Fix and change y command in the following ways:
- When 'n' is used as a delimiter escaping 'n' will remain a newline
instead of becoming a 'n' character. This is how POSIX specifies how this
should work. Other implementations tested also do this wrong.
- '[' and maybe other characters are not special during the parsing of the
y command and don't need to be matched or treated special in any way.
- POSIX specifies that a backslash followed by anything other than the
delimiter, 'n', and another backslash, as well as repeating characters in
string1 are unspecified. Since the various implementations handle these
situations in opposing ways choose to error out on them to prevent people
falling into the pitfall of expecting identical behaviour on various
implementations.

Inspired by the sed.1 patch by kshe59 <at> zoho <dot> eu
Feedback and OK millert@
Manpage bits OK jmc@

show more ...

Make the r command filename obligatory, similar to what FreeBSD and NetBSD
do for several years.
While here make corresponding error message for missing read and write file
consistent between command

Make the r command filename obligatory, similar to what FreeBSD and NetBSD
do for several years.
While here make corresponding error message for missing read and write file
consistent between commands/flag, and shrink the the code of the w flag of
the s command by making it use the same code as the w command.

Prompted by a larger diff by kshe59 <at> zoho <dot> eu
OK millert@

show more ...

Apply stricter pledge rules if possible. These are based on the usage of
the 'w' and 'r' functions and the 'w' flag to the 's' function.

If non of the above is used and input is being read from stdi

Apply stricter pledge rules if possible. These are based on the usage of
the 'w' and 'r' functions and the 'w' flag to the 's' function.

If non of the above is used and input is being read from stdin, we drop
all the way down to stdio!

Original inspiration by benno@.

OK millert@

show more ...

Split error() into error() and warning() so error() can be marked __dead to
appease gcc.

ok procter@ deraadt@

Tidy up error() levels: merge COMPILE and COMPILE2, and kill ERROR (unused)

ok tobias@

Rename err() to error() to prevent confusion with the stdlib function.

Discussed with jca@.

POSIX specifies that that multiple '!' characters preceding a
function should be treated as a single negation. From FreeBSD
via Liviu Daia.

Bounds check the file path used in the 'w' command. Modified version
of a diff from Sebastien Marie to prevent a crash found by Sebastien
with the afl fuzzer.

add a xreallocarray() like the existing fatal xmalloc(), and use it to
detect potential integer.
ok doug

unsigned char for ctype
ok krw okan

NUL-terminate buffer in compiler_text() after realloc, not before.
Now that we have dynamically sized buffers there may not be room
for the NUL until after realloc.

Follow POSIX (IEEE Std 1003.1, 2004 Edition) in the implementation
of the y (translate) command.

"If a backslash character is immediately followed by a backslash
character in string1 or string2, the

Follow POSIX (IEEE Std 1003.1, 2004 Edition) in the implementation
of the y (translate) command.

"If a backslash character is immediately followed by a backslash
character in string1 or string2, the two backslash characters shall
be counted as a single literal backslash character"

From FreeBSD; ok millert@ halex@

show more ...

Allow [ to be used as a delimiter.

From FreeBSD; ok millert@ halex@

rcsid[] and sccsid[] and copyright[] are essentially unmaintained (and
unmaintainable). these days, people use source. these id's do not provide
any benefit, and do hurt the small install media
(th

rcsid[] and sccsid[] and copyright[] are essentially unmaintained (and
unmaintainable). these days, people use source. these id's do not provide
any benefit, and do hurt the small install media
(the 33,000 line diff is essentially mechanical)
ok with the idea millert, ok dms

show more ...

When adding text due to an a, c, or i command, don't eat the space(s) at
the beginning of the lines since the addition is supposed to be "verbatim".

From NetBSD, also matches the bahaviour of GNU se

When adding text due to an a, c, or i command, don't eat the space(s) at
the beginning of the lines since the addition is supposed to be "verbatim".

From NetBSD, also matches the bahaviour of GNU sed

ok millert@

show more ...

add -E flag to compile regular expressions using the extended POSIX
syntax. The -E flag is compatible with the other BSDs and OSX. -r is
also provided as an alias for compatibility with GNU sed.

fee

add -E flag to compile regular expressions using the extended POSIX
syntax. The -E flag is compatible with the other BSDs and OSX. -r is
also provided as an alias for compatibility with GNU sed.

feedback from jmc@ and millert@
ok millert@ pyr@ henning@ marco@

show more ...