Re: Trying out GitLab (was Re: In support of Jonas Bernoulli's Magit)

["Ævar Arnfjörð Bjarmason" <avarab@gmail.com>]

On Sat, Jul 08 2017, Eli Zaretskii jotted:

>> From: Dmitry Gutov <dgutov@yandex.ru>
>> Date: Sat, 8 Jul 2017 14:13:44 +0300
>> Cc: Toon Claes <toon@iotcl.com>, emacs-devel@gnu.org
>>
>> On 7/8/17 2:02 PM, Ævar Arnfjörð Bjarmason wrote:
>>
>> > Usually when I discover bugs in things part of Emacs itself I just work
>> > around them, since starting to contribute to the project would require
>> > signing papers etc.,
>>
>> You are not helping our case, unfortunately: even if Emacs switches to
>> GitLab or something similar (which I hope it will), contributing will
>> still require signing papers.
>
> It's actually more than that: patches submitted to Emacs need to
> conform to our coding and various other standards: include
> properly-formatted commit log messages, documentation, and (where
> appropriate) tests, etc.  Patch review could require cleanup changes
> etc.
>
> People who find this too much of an effort should perhaps describe
> their solution without showing any actual code, so that someone else
> could implement it.  This way, copyright assignment is not an issue,
> and chances are the change will be eventually made.

   "My project's requirements are so onerous that some find it not worth
   their time to contribute at all, but it's OK because those some
   people can instead describe in prose what their not-good-enough
   contributions should look like instead, surely that'll yield the same
   end result.".

I think that's a fair paraphrasing of the point you made. Of course
everything that adds extra friction to contributing is an issue, and
results in fewer contributions.

Anyway, I'm not trying to get into some flamewar about the relative
merits of FSF's copyright policies or Emacs's contribution policy on
this list. I just wanted to point out that Toon Claes's point that
started this thread had been lost midway through.

Which is that when people talk about having project X on Github or
Gitlab they don't mean so in the narrow sense of having it be mirrored
there in some capacity, or using some narrow subset of features like CI.

They mean that they'd like project X to be as easy to contribute to as
they're used to from contributing to other projects on those platforms.

> Of course, it is better to actually sign papers, which is a one-time
> hassle.
>
>> > which would be a huge hurdle compared to what I'm
>> > fixing (none of my Magit contributions are major, just small fixes here
>> > & there).
>>
>> Small fixes (up to ~15 lines in total) don't require copyright assignment.
>
> Yes.  But that limit is quickly exhausted.