Interesting to see discussion on this list.

On Wed, Dec 16, 2015 at 10:45:05AM +0100, Matěj Cepl wrote:
> My own overview 
> https://matej.ceplovi.cz/blog/current-state-of-the-distributed-issue-tracking.html
> although more than five years old is more or less perfect and not much has
> changed unfortunately. I liked gitissius and the idea of
> data-in-hidden-git-branch (compatibility with other VCS is mostly useless
> these days, I guess), but it was never finished and fixed to the production
> level.

I have a slightly heretical point of view: I do not want my tickets in
the same git repository as the code. Apart from finding it an ugly
solution, I'd really like to not tie tickets and code together. Some
of my projects have more than one git repository. Also, a lot of
tickets are support issues, with large attachments, and there's no
point in making the code repository larger for those.

My approach, in a very-low-priority side project, is approximately
this:

* A ticket represents an issue a user has, or a bug in the project, or
  a similar problem. The goal is to fix the problem.

* A problem is solved primarily via a discussion. A ticket system's
  primary job is to record the discussion, and manage a little
  metadata about it, primarily "is this problem still relevant" (is
  the ticket open).

* All discussion is over email, to avoid making people have to sign up
  to a website. The emails are stored in a per-ticket Maildir. Later,
  a web UI can be added, which will add fake emails to the ticket, but
  that's optional for now.

* Having at least a read-only view of the system via a static
  rendering of the ticket system data is crucial. A more interactive
  one would be nice, of course.

* All ticketing system data is stored in git, and normal git
  operations are used for syncing. The data is stored in a way that
  makes merging unlikely to cause conflicts.

* Syncing with other ticketing systems is not a priority.

I realise those aren't things that fit well into other people's world
view, for which I am afraid I do not apologise at all. I'm writing
this for myself, and my own needs.

I have most of this kind of working, but in a clumsy way. I don't yet
use this myself, I hope to start to some time next year.

http://git.liw.fi/cgi-bin/cgit/cgit.cgi/distix/ for the code. Written
in Python. If anyone wants to help, writing something to generate a
static website from a distix repo would be nice.

-- 
Schrödinger's backup hypothesis: the condition of any backup is
undefined until a restore is attempted. -- andrewsh

Attachment: signature.asc
Description: Digital signature

_______________________________________________
dist-bugs mailing list
[email protected]
http://kitenet.net/cgi-bin/mailman/listinfo/dist-bugs

Reply via email to