Talk:Patch (Unix)

Needs work
This initial version of the article needs work, especially the second paragraph. While I understand it, it could be stated better. I probably won't be doing the work here, though. I am not a Unix-head. :-) &mdash;Frecklefoot 16:21, 19 Feb 2004 (UTC)

I am :) And I think at least the -p1 option is something you usually do not want. You'd rather want to use -p0. -andy 217.2.48.207 22:03, 17 Jun 2005 (UTC)

GNU utilities for Win32
I do not believe you can get the package from the page mentioned. The download link stops me at "Forbidden". This link has a working version: http://gnuwin32.sourceforge.net/

I'll leave this for a few days if someone more experienced wants to comment, otherwise, I'll go change the link myself. 66.71.75.74 23:14, 20 February 2007 (UTC)


 * I made the change. As you said, unxtools is no longer available.
 * Also, the GnuWin32 tools are newer and better than those once available from unxtools. 60.241.171.136 00:52, 29 April 2007 (UTC)

Readability of diff files
This article says:


 * Context diffs are often easier to read by people unfamiliar with the unified diff format, but are generally significantly larger than unidiffs.

But the article on Diff says the opposite:


 * The unified format (or unidiff) inherits the technical improvements made by the context format, but produces a smaller diff and a more human-readable result. —Preceding unsigned comment added by 172.202.212.171 (talk) 13:16, 7 October 2007 (UTC)
 * Context diffs should be avoided in favor of unified context diffs because of the time stamp format. Plain context diffs contain local time marks with localized text while unified context diffs use digits only and add a GMT offset to permit to parse an always correct time stamp. Schily (talk) 13:30, 4 June 2015 (UTC)


 * It's subjective, and comments regarding relative readability should be rephrased for NPOV. For either case, reading and interpreting the diff output requires some (minimal) experience. Tedickey 20:33, 7 October 2007 (UTC)

answers.com
http://www.answers.com/topic/patch-unix?cat=technology

this is the same information. what gives? —Preceding unsigned comment added by Mustardy (talk • contribs) 18:19, 21 February 2008 (UTC)
 * Yes, answers.com use Wikipedia articles, as they are entitled to under the terms of the GFDL. See WP:Mirrors and forks. EdC (talk) 23:40, 21 February 2008 (UTC)

IBM patch
The comment reminds me to note that Univac had a simple batch editor (same issue), e.g, "slp". DEC used the same name for a similar program (DEC borrowed more names from Univac than from IBM, though "symbiont" never took hold). Tedickey (talk) 00:30, 12 March 2008 (UTC)

POSIX compliant (sic)
Schily's existing promotional edits make derogatory remarks in the change comments about other implementations, while failing to provide reliable sources, e.g., to substantiate "POSIX compliant" on work to which he has made some changes. The place to carry on a discussion is on the talk page, rather than use the change comments to make personal attacks on other editors TEDickey (talk) 00:43, 4 June 2015 (UTC)

For instance, suitable sources would be found here &mdash; if they exist. TEDickey (talk) 01:01, 4 June 2015 (UTC)


 * Rather than attacking other people and trying to make your typical promotional edits, you should learn how to use reliable sources. If you did ever try to inform yourself about "patch" on POSIX certified platforms, you did of course know that these platforms (e.g. Solaris, HP-UX. AIX) do not use and cannot use the non-POSIX gpatch as their official patch implementation; they instead use the closed source implementation original owned by IBM, Sun and the OSF. Sun meanwhile merged with Oracle and the OSF merged with the OpenGroup. My impression is that you revert the text because you don't like to see that gpatch happily ignores POSIX rules for patch even when called with --posix.


 * Is there really no chance that you become a collaborative and constructive member of the WP community? Schily (talk) 10:22, 4 June 2015 (UTC)

That's a lot of words without making any points. Are you going to provide a reliable source for any of that? TEDickey (talk) 00:12, 5 June 2015 (UTC)
 * Once you learned what a reliable source is, it may make sense to continue this discussion. For now, you again try to waste my time. Schily (talk) 12:58, 8 June 2015 (UTC)

Git repositories
I see that there are links to git repositories, but they are for the GNU project, while the page title ("Patch (Unix)") mentions Unix. So shouldn't there be some precision to bring to the link, such as "Official specification: ", and change "Repository" to "GNU implementation repository: "? Remember tha GNU is not Unix, and thus GNU patch is not an official implementation reference. Sedrikov (talk) 09:14, 7 August 2019 (UTC)

toybox patch
Another from-scratch implementation of patch originated in toybox, written in 2007 (https://github.com/landley/toybox/commit/42ecbab08101) and ported from there to busybox in 2010 (http://lists.busybox.net/pipermail/busybox/2010-August/073112.html). The two have been kept roughly in sync ever since. Earlier this year there was an infodump on the toybox mailing list about how patch works: http://lists.landley.net/pipermail/toybox-landley.net/2019-January/010049.html. Toybox patch is the version Android is using to build itself: http://lists.landley.net/pipermail/toybox-landley.net/2019-December/011322.html — Preceding unsigned comment added by 2605:6000:E947:6500:6680:99FF:FE6F:CB54 (talk) 19:14, 20 December 2019 (UTC)

wall-vs-gnu
The version of patch in schilytools started with an early version of the gnu fork, rather than being just Wall's original version. (And given the origin of the comment in this topic, the origin for the others is dubious). See discussion in the diffstat page TEDickey (talk) 19:06, 8 May 2023 (UTC)