Prefer setvbuf() to setlinebuf() for portability; ok deraadt@

unsigned char casts for ctype
ok krw

Fix confusing patch(1) output on failure when using -C. From
dragonflybsd via Loganaden Velvindron. OK mikeb@ and deraadt@

Send normal output to stdout instead of everything to stderr. This
is currently done in the latest versions of GNU patch.

Please watch out for output appearing out-of-order.

Discussed with millert

Send normal output to stdout instead of everything to stderr. This
is currently done in the latest versions of GNU patch.

Please watch out for output appearing out-of-order.

Discussed with millert and deraadt.

OK deraadt

show more ...

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 ...

If no patch can be found in input, in addition to complaining on stderr,
have the exit code indicate failure, not success.

Discrepancy with GNU patch pointed out by Arfrever Frehtes Taifersar Arahes

If no patch can be found in input, in addition to complaining on stderr,
have the exit code indicate failure, not success.

Discrepancy with GNU patch pointed out by Arfrever Frehtes Taifersar Arahesis.

Help from and ok djm@, "sounds good to me" millert@

show more ...

The code to detect multiple applied diffs to empty files I introduced 4
years ago is not correct if the diff has an empty context for other
reasons, so revert that change. Problem found by Joerg Sonn

The code to detect multiple applied diffs to empty files I introduced 4
years ago is not correct if the diff has an empty context for other
reasons, so revert that change. Problem found by Joerg Sonnenberger;
ok tedu@ millert@

show more ...

adds a detailed synopsis to the man page; sorts options in
the description list and adds missing arguments; syncs usage()

written with a lot of useful advice by jmc@

ok by jmc@

type police and assorted cleanup. From Lionel Fourquaux; ok ray@

Write reject file in unified format if the original diff was in
unified format.
ok deraadt@

remove unused variable

Properly detect if a patch already has been applied, even if there
are no eols at eofs involved. Tested by quite some people.

ok deraadt@

Unbreak handling of standard diffs (which have an empty context).

ok @millert

Teach patch(1) to detect if a diff creating a file (or one adding
lines to an empty file) already has been applied; be a bit less
chatty about empty contexts. Testing hshoexer@ ok millert@

o fairly major man page reorganization
o add --posix option for strict POSIX conformance
o change filename determination to match GNU patch
Tested by various people, including a full ports build by n

o fairly major man page reorganization
o add --posix option for strict POSIX conformance
o change filename determination to match GNU patch
Tested by various people, including a full ports build by naddy@

show more ...

Do not add an extra newline at the end if the last line of the input
file contains no newline and the diff does not touch the last line.
Contributions from millert@.

ok millert@ tedu@

Warn if the diff is a context or unified diff and the context
is empty. In this case, a previously applied patch cannot be detected.

ok millert@ tedu@

- use stdbool.h instead of roll-your-own booleans
- fix some -Wall warnings
- fix asserts: in some cases remove them, in other cases they have
become Internal errors or detection of malformed patch f

- use stdbool.h instead of roll-your-own booleans
- fix some -Wall warnings
- fix asserts: in some cases remove them, in other cases they have
become Internal errors or detection of malformed patch files.
- fix some free() related code

ok millert@ tedu@

show more ...

Historically, patch would treat a bare -p as -p0. This contradicts
POSIX and GNU patch has also removed this, so we will too. No
objections on icb (no one even seemed to know about this "feature").

Print a maximum of one invalid line number warning per patch in a patch file.
Thanks to espie@ for spotting the problem.

ok millert@ henning@ espie@

Add POSIX -i option; tedu@ OK

o add pathnames.h
o ignore empty TMPDIR environment variable
o strip any trailing slashes from TMPDIR
otto@ OK

minor knf and cleanups; otto ok

Be more exact on how backup files are handled.
ok millert@ tedu@

Make patch(1) exit value match POSIX and be consistent with diff.
Comments and OK from otto@