Commit Graph

65 Commits

Author SHA1 Message Date
Warner Losh
fbbd9655e5 Renumber copyright clause 4
Renumber cluase 4 to 3, per what everybody else did when BSD granted
them permission to remove clause 3. My insistance on keeping the same
numbering for legal reasons is too pedantic, so give up on that point.

Submitted by:	Jan Schaumann <jschauma@stevens.edu>
Pull Request:	https://github.com/freebsd/freebsd/pull/96
2017-02-28 23:42:47 +00:00
Xin LI
8e2106f0cb Restore r312404:
Use S_ISREG instead of manual & (also it's better to compare the
result from & and the pattern instead of just assuming it's one bit
value).

Pointed out by Tianjie Mao <tjmao tjmao com>.

MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D4827
2017-02-05 08:51:41 +00:00
Xin LI
1bcffc747f Revert r312404 as we need to do an exp-run and fix existing ports that
rely on the previous behavior.

Requested by:	antonie (portmgr)
2017-01-20 06:45:06 +00:00
Xin LI
805ca1775e Use S_ISREG instead of manual & (also it's better to compare the
result from & and the pattern instead of just assuming it's one bit
value).

Pointed out by Tianjie Mao <tjmao tjmao com>.

MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D4827
2017-01-19 08:01:35 +00:00
Pedro F. Giffuni
249a8a8003 sed(1): Revert r303047 "cleanup" and therefore r303572.
While big, the change was meant to have no effect on behavior and instead
so far we have found two regressions: one in the etcupdate tests and
another one in the games/openttd port[1].

Revert to a known working state. We will likely have to split the patch in
functional parts before bringing back the changes.

PR:		195929
Reported by:	danfe, madpilot [1]
2016-08-02 15:35:53 +00:00
Enji Cooper
637a1a7446 Fix regression with /i caused by r303047
'\n' was specifically added to -e arguments prior to r303047. Restore
historical behavior which in turn fixes usr.sbin/etcupdate/preworld_test:main .

The fix is being committed to address the issue in the short term and may be
iterated upon as noted in bug 211399

Discussed with:		mi, pfg
Differential Revision:	https://reviews.freebsd.org/D7368
PR:			195929, 211399 [*]
MFC after:		18 days
X-MFC with:		r303047
Reported by:		Jenkins
Sponsored by:		EMC / Isilon Storage Division
2016-07-31 05:31:09 +00:00
Baptiste Daroussin
a43d43d574 Remove usage of _WITH_GETLINE from usr.bin 2016-07-30 01:07:47 +00:00
Ed Schouten
6718ba1851 Don't call basename() and dirname() in an unportable way.
POSIX allows these functions to modify their input buffer, so that they
have storage for the return value. Pull copies of the filename before
calling these utility functions.
2016-07-28 15:17:12 +00:00
Pedro F. Giffuni
b4932d18ee sed(1): Assorted cleanups and simplifications.
Const-ify several variables, make it build cleanly with WARNS level 5.

Submitted by:	mi
PR:		195929
MFC after:	1 month
2016-07-19 22:56:40 +00:00
Pedro F. Giffuni
11398c77da Revert r299279:
Simplify redundant malloc'ing in sed -e.

It is causing havoc in the ports tree:

===>  Configuring for wxsvg-1.5.7
sed: 1: "/gcc_dir=\\`/s/gcc /$CC /": bad flag in substitute command: '/'
*** Error code 1

===>  Patching for vips-8.3.1
sed: 1: "1s|^#![[:space:]]*/usr/ ...": bad flag in substitute command: 's'
*** Error code 1

PR:		195929
Reported by:	danilo
2016-05-10 02:02:50 +00:00
Pedro F. Giffuni
d74b9e1311 Simplify redundant malloc'ing in sed -e.
When encountering an -e argument, sed currently mallocs a string to COPY
the optarg -- with '\n' appended. The appendage does not seem necessary --
indeed, the same call to add_compunit processing the sole command (given
without -e) passes the *argv verbatim: without making a copy, and without
appending newline.

This matches what is done in other BSDs.

Submitted by:	Mikhail T.
PR:		195929
MFC after:	2 weeks
2016-05-09 18:53:46 +00:00
Pedro F. Giffuni
76c76d54f8 sed: fix pasto from previous r277802.
MFC after:	3 days
2015-01-27 19:46:19 +00:00
Pedro F. Giffuni
fc04dce059 Fix resource leak and dereference after NULL.
process.c:
Protect access against NULL.

main.c:
Prevent outfile overwrite resource leak.

CID:	271181
CID:	1006930

Obtained from:	NetBSD
MFC after:	3 days
2015-01-27 18:58:24 +00:00
Pedro F. Giffuni
707c9cc5d1 sed: Address warnings with clang and gcc48.
MFC after:	2 weeks
2015-01-08 16:33:15 +00:00
Jean-Sébastien Pédron
96d68291af sed(1): Don't force a newline on last line, if input stream doesn't have one
While here, change how we check if the current line is the last one.
Before, we just checked if there were more files after the current one.
Now, we check the actual content of those files: they files may not have
a line at all. This matches the definition of the "last line" by the
Open Group.

The new behavior is closer to GNU sed.

PR:		160745
Phabric:	https://phabric.freebsd.org/D431
Reviewed by:	jilles
Approved by:	jilles
Exp-run by:	antoine
2014-08-08 17:29:01 +00:00
Pedro F. Giffuni
f243100087 More sed(1) usage fixing: the extension -i is not optional.
Pointed out by:	jmallet
MFC after:	3 days
2014-07-03 00:31:13 +00:00
Pedro F. Giffuni
9b788a2e5b Fix sed(1) usage: the extension -i is not an optional.
MFC after:	3 days
2014-07-02 23:07:01 +00:00
Pedro F. Giffuni
90d81f30ec Drop ifdef nonsense: just use setvbuf(3).
Pointed out by:		kib, bde
2014-06-21 14:07:33 +00:00
Pedro F. Giffuni
fc39ce9ee1 Add -u (unbuffered output) after GNU sed.
Obtained from:	NetBSD
MFC after:	1 week
2014-06-20 21:41:30 +00:00
Pedro F. Giffuni
1a2a4fc8ce sed: use getline() instead of fgetln().
In BSD, fgetln() available in libc but in Illumos the Solaris port had to
include it internally. It also seems to have caused problems [1].

Aid portability by using getline() instead.

Reference:
https://www.illumos.org/issues/3820 [1]

Submitted by:	Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.com>
Reviewed by:	dds
MFC after:	2 weeks
2013-06-26 04:14:19 +00:00
Jilles Tjoelker
919e14f01e sed: Try hard links to make -i target available continually.
When creating a backup file, sed renamed the original before renaming the
changed copy into place, leading to a short time when no file with the
original name was present (usually only visible on SMP systems). Try
creating the backup file using a hard link instead, avoiding this problem.
If creating the hard link fails for any reason, fall back to the old rename
method.

When not creating a backup file, sed already renamed the changed copy onto
the original. This remains unchanged.

I am not adding the suppression of redundant fchown/fchmod to this commit,
because FreeBSD appears to check this in the kernel (for msdosfs at least).

PR:		bin/153261
Submitted by:	Pedro F. Giffuni
Reviewed by:	dds (older version)
Obtained from:	Illumos
MFC after:	2 weeks
2011-01-08 00:03:18 +00:00
Warner Losh
701d73b65c Make -r mean exactly the same thing as -E for increased compatibility
with GNU sed.

MFC after:	7 days
2010-03-31 17:40:13 +00:00
Xin LI
821df508e8 Revert most part of 200420 as requested, as more review and polish is
needed.
2009-12-13 03:14:06 +00:00
Xin LI
6f2d322192 Remove unneeded header includes from usr.bin/ except contributed code.
Tested with:	make universe
2009-12-11 23:35:38 +00:00
David Malone
d3e5e11ce6 WARNS fixes:
1) Add missing parens around assignment that is compared to zero.
2) Make some variables that only take non-negative values unsigned.
3) Some casts/type changes to fix other constness warnings.
4) Make one variable a const char *.
5) Make sure termwidth is positive, it doesn't make sense for it to be negative.

Approved by:	dds
2008-02-09 09:12:02 +00:00
Diomidis Spinellis
a511e63958 Fix visibility of "inplace".
Found by: CScout
2007-11-08 08:57:51 +00:00
Yaroslav Tykhiy
26a5710c40 Don't forget to clear out the hold space for each subsequent file
when in -i mode so that each file gets a clean context of its own.

Add a regression test for the bug.

Tested with:	regression tests
2007-06-12 12:05:24 +00:00
Yaroslav Tykhiy
f6703c9c0a Change the semantics of -i (in-place editing) so that it treats
each file independently from other files.  The new semantics are
desired in the most of practical cases, e.g.: delete lines 5-9
from each file.

Keep the previous semantics of -i under a new option, -I, which
uses a single continuous address space covering all files to edit
in-place -- they are too cool to just drop them.

Add regression tests for -i and -I.

Approved by:	dds
Compared with:	GNU sed
Discussed on:	-hackers
MFC after:	2 weeks
2007-04-21 01:21:36 +00:00
Stefan Farfeleder
7c4847b792 The function inplace_edit() doesn't exist anymore, remove the prototype.
Submitted by:	Leonardo Chiquitto Filho
2005-09-24 13:18:30 +00:00
Gleb Smirnoff
1ed86f6c90 Add option -l, to make output line buffered.
PR:		bin/78692
Submitted by:	Eugene Grosbein
MFC after:	2 weeks
2005-05-10 13:40:50 +00:00
Stefan Farfeleder
a3eb24c68b Include missing headers. 2005-02-09 09:13:36 +00:00
Diomidis Spinellis
30fd73fb81 Per letter dated July 22, 1999 remove 3rd clause of Berkeley derived
software (original contributor).

Reviewed by:	imp
2004-08-09 15:29:41 +00:00
Tim J. Robbins
81a8648adb Make the 'y' (translate) command aware of multibyte characters. 2004-07-14 10:06:22 +00:00
Brian Somers
7bbdbe1482 Plug a file descriptor leak.
When sed is asked to inline-edit files, it forgets to close the temporary
file and runs out of descriptors for long command lines (assuming you reset
kern.maxfilesperproc to something sane that's less than the number of files
passed to sed).
2004-06-03 22:33:50 +00:00
Dag-Erling Smørgrav
591c337e01 Fix the case where the file is not in the current directory.
Discovered by:	Vladimir Kravchenko <jimson@mostcom.ru>
Pointy hat to:	des
2003-11-04 22:39:25 +00:00
Dag-Erling Smørgrav
f54cda1467 Reimplement in-place editing in a slightly less disgusting manner. Also,
make an effort to preserve the ownership and mode of the file we are
editing.

Sponsored by:	Registrar AS
2003-11-04 13:09:16 +00:00
Dag-Erling Smørgrav
e6478125c8 ANSIfy 2003-11-04 12:15:20 +00:00
Dag-Erling Smørgrav
9df25dda8c Correct style bugs in the in-place editing code. 2003-11-04 12:07:09 +00:00
Tony Finch
d1691a2ff6 Fix some bugs in in-place editing:
(1) errors from freopen were not reported correctly
(2) large files were not handled correctly
(3) read-only files broke things

MFC after:	1 week
2002-07-30 19:42:18 +00:00
Tim J. Robbins
8523e9a677 Avoid checking whether each line is the last line of the file when we don't
need to know. Instead, check when we are trying to match a "$" address.
This does not change the way sed processes regular files, but makes it behave
more sensibly when used interactively.

PR:		40101
MFC after:	2 weeks
2002-07-03 14:38:27 +00:00
Tim J. Robbins
ed92199d30 Don't store newlines at the end of each line in the hold/pattern spaces,
instead add the newline when the pattern space is printed. Make the `G' and
`H' commands add a newline to the space before the data, remove bogus
addition of newline from `x' command.

PR:		29790, 38195
2002-06-22 01:42:26 +00:00
Tim J. Robbins
254fac850d Correct the handling of (for example) the N command by only zeroing the
input space in mf_fgets() if we reach the end of all input files.
2002-06-16 08:44:39 +00:00
Tim J. Robbins
6689fb2bbc Exit with non-zero status if an input file can not be opened. This was
broken by rev. 1.22.
2002-06-14 02:20:05 +00:00
Maxim Sobolev
8701af62ee Fix one serious bug and one potential problem with in-place editing code:
- original version of code worked incorrectly when more than one
  input files were specified - it was moving the last line from the 1st file
  to be the first line of the 2nd, last line of the 2nd to be the first
  line of the 3rd and so on;

- use mmap()->write() to create temporary file instead of
  malloc()->read()->write(), which was not only slower, but also did not
  bother to free allocated memory once backup file was created, potentially
  leading to memory exhausting when regex is applied to a big file or a large
  number of small ones.
2002-06-14 01:28:52 +00:00
Tim J. Robbins
3af4dcb223 If a file argument cannot be processed, process the remaining ones
and exit non-zero (SUSv3)
2002-05-24 10:58:21 +00:00
Brian Feldman
e7b663eff2 o Clean up tmp file usage a little by using mkstemp(3) instead of
mktemp(3).  It would be amazingly unlikely, but the former method
  could result in a symlink attack.  A better solution would use
  ${TMPDIR}, though.
o Make sed not overwrite old backup files with no warning.
2002-05-14 23:15:42 +00:00
Juli Mallett
90f7fe63e9 Oops, update usage() to have -i. 2002-05-07 23:33:44 +00:00
Juli Mallett
5d16412db7 Add handling for any nil-length string passed to -i for the backup extension.
Add a note that this is kinda-sorta dangerous to the manual page.
2002-05-07 23:32:26 +00:00
Juli Mallett
d88e9d841a Un-shadow the `fname' variable. 2002-05-07 23:06:47 +00:00
Juli Mallett
839af0c103 Add a -i option to sed(1) to do inplace editing, to give us an alternative to
Perl for such things.  The key difference to Perl is that a backup extension
*MUST* be specified, because on one hand it isn't recommended to have options
which optionally take a parameter, and on the other hand, it'd be slightly
unpleasent to implement proper handling for that.

The difference between this and the version posted to developers@ is that it
does handle multiple files in argv after the getopt(3) handling "correctly",
in that the inplace editing-specific code has been moved out to a function,
and that function is used beyond the first file in our linked list.

This option has been documented as FreeBSD-specific in the manpage.

Reviewed by:	developers@ (got feedback from: des, fanf, sobomax, roberto,
		obrien)
MFC after:	1 week
2002-05-07 18:32:18 +00:00