Re: [PATCH v1 19/25] contrib: remove 'diff-highlight'

Jeff King <peff@peff.net> writes:

I do not know how much attention you have been paying, and I suspect that you may be aware of all of the following, but I'll send this out anyway, primarily so that others involved in other subthreads can find out the story behind this.

There is no "prove yourself is worthy or get evicted" purge going on in the contrib/ area. I saw contrib/README referred to a few times in the near-by threads, and I think these patches are done primarily by deliberately misinterpreting one part of it in order to grab attention by many people and also to sabotage the project.

The contrib/README file was written back when Git was still a small and young project that was trying to build an ecosystem by having an area to host stuff that are not core-material for some reason or other (e.g. only useful in some environments, only useful for some workflows, the design or code not up to par to be in core) in my tree to ease discovery and distribution.

There, I wrote:

    I expect that things that start their life in the contrib/ area
    to graduate out of contrib/ once they mature, either by becoming
    projects on their own, or moving to the toplevel directory.  On
    the other hand, I expect I'll be proposing removal of disused
    and inactive ones from time to time.

The purpose the last sentence in that paragraph is there was to protect our codebase and our users from those who see an opportunity to throw their ware in to our tree and go AWOL, by giving me, the maintainer, a "stick" to prod them, saying "You as the primary author are responsible for taking good care of the ware you created by responding to issues (questions, suggestions, bugs, patches) in a prompt manner, or your ware may even get evicted."

Among contrib/ materials we have today, I do not think there is anything that requires me to exercise that "stick". diff-highlight certainly is not. Perhaps subtree is the closest, as I see issues raised from time to time but the original champion seems to be inactive for some time, but even there, I recently saw somebody hinting to volunteer to take it over after sending a patch or two to it, and I do not intend to exercise the "stick" yet.

The sole mention of possible removal from contrib/ is this one:

    http://thread.gmane.org/gmane.comp.version-control.git/248063/focus=248457

in which Felipe said:

    I don't want to do anything for a "contrib" tool.

and I suggested that he has an option to make it a standalone third-party project. With the promotion to the core has already been ruled out in the thread that begins at this one:

    http://thread.gmane.org/gmane.comp.version-control.git/247660/focus=248167

that is one of the only two alternatives I can offer, given that the Git ecosystem has matured enough to let third-party tools flourish on their own merit. "We may want a better plug-in registry for Git" I mentioned in

    http://thread.gmane.org/gmane.comp.version-control.git/248063/focus=248391

was to help us in that direction, but seeing that imerge mentioned in many places I do not even regularly visit with the current "discovery and distribution" infrastructure, perhaps yet another new registry may not even be necessary. I dunno.

In any case, that suggestion to remove not related to the "stick", either, and certeinly not about "prove yourself" purge that does not even exist.

So I think most of these removal patches can safely be ignored.

I agree with you and Jonathan that removal of contrib/vim may be a good idea, but that is not due to "stick" nor "prove yourself", either. Jonathan's proposed alternative $gmane/248506 does a good job of explaining and justifying the change. It is a graduation "by becoming projects on their own" that contrib/README mentions.