Re: A registry for distributed sources and binaries

[Andreas Enge <andreas@enge.fr>]

Hello again,

On Mon, Jul 25, 2016 at 01:21:50AM +1000, Jookia wrote:
> An issue tracker that you can reply to by the web would be much much better,
> because there's less things to go wrong and less ways to be shamed for. I've
> suggested this many times and the only responses I've heard are 'no' and 'let me
> tell you how easy you can use mailing lists', so I give up.

let us start from here. As said, I think we could/should improve our tooling
and processes. And it would certainly be best if those who are most unhappy
with our current ways could make concrete suggestions, that are also
compatible with the ways in which people work who are happy with the current
situation.

There was a trial with patch/patches/patchwork or whatever it is called
(so the complaints have not been ignored!), but I do not feel like this
has been of much use and help.

> Even worse, if I want to reply to an issue on a mailing list that I'm not
> subscribed to, it's difficult. I still haven't figured it out, maybe you can go
> to the archive and download an mbox and look at the reference and ask your mail
> client to reply to it? I don't know.

This is worse with most web based issue trackers, which require you to sign up
and, in the worst case, agree to terms and conditions. For mailing lists, at
least you can sign up without terms and conditions and easily unsubscribe.

Personally, I think that debbugs could also be used to handle patch
submissions. That would make it easier to track the status of patches.
"Bugs" containing applied patches could be closed as "fixed", others
where the submitter disappears could be closed after a while as "wont-fix".
The guix-devel mailing list could be subscribed to the bug tracker, so that
the complete discussion would appear in the list. Also, "bugs" can be
treated (and in particularly closed when the patch is applied, which should
be the most frequent outcome) easily by e-mail. But as far as I know, debbugs
has no web interface.

So do you know a concrete solution? I think it needs to/should satisfy the
following requirements:
1) It must be free software that we can install on our own servers.
   (And it should represent almost zero maintenance burden, so we can
   concentrate on the work rather than on the tools.)
2) It should have a web interface :-), but also be usable from the command
   line; maybe by mail, preferably with some kind of emacs integration.
3) It should be usable without signing up.
There may be more requirements that I do not think of.

> I think I'm done for now. No amount of writing has even started a discussion on
> how to fix things, only that things aren't broken.

I hope that I am starting a discussion with the above. Still I think that
the main problem is that we do not have enough people willing to do enough
boring work. If we find ways to make contributing to the boring and exciting
tasks easier and more pleasant, then I am all for it.

Andreas