Re: Git mirrors

[Óscar Fuentes]

Eli Zaretskii <address@hidden> writes:

[snip]

>> If some Emacs hackers feel more at home with
>> git/mercurial/whatever-other-Free-tool, why not simply provide it?
>
> Because it takes the effort from other useful development tasks.

Eli, please don't do selective quoting. This is what I said, which makes
your sentence above irrelevant:

Óscar Fuentes <address@hidden> writes:

> It is the other way around: the infrastructure must adapt to the needs
> of users, whenever there are enough technical and human resources. If
> some Emacs hackers feel more at home with
> git/mercurial/whatever-other-Free-tool, why not simply provide it?

"enough technical and human resources" == we have the machines, the
bandwith and someone is willing to do the job.

> And because providing every possible tool, just because someone feels
> more at home with it, when reasonable alternatives exist, is not the
> norm in Free Software projects.  What is not required from projects
> that use git should not be required from projects that use bzr.

You talk about "required" as if someone were asking for a rule:
"projects who use bzr must provide git mirrors." That's not the case and
you know it.

Why not support as much as possible as long as someone wishes to do the
work? MSDOS anyone, Eli? You want it, you do the work, it keeps you
contributing, and that's good for Emacs. Why it is different with those
who feel more comfortable (and productive!) using a different VC tool?
That means more contributions to Emacs. How could that be "not
beneficial" for GNU?

And if no project who uses git offers bzr too, maybe that's because such
request is never made, because there is just a small group of users who
strongly prefer bzr over git/mercurial, with most projects having none,
and possibly too because offering bzr support is not as straightforward
as offering git, as we are experiencing, or the contrary, creating a bzr
mirror in Launchpad wich updates daily is easy enough. Not because there
is some kind of attitude of git users towards bzr users that should be
corresponded.