Re: Bug tracker

Dave Page wrote:
> On Thu, Sep 25, 2008 at 10:14 AM, Mickael Deloison <mdeloison(at)gmail(dot)com> wrote:
>> Hi,
>>
>> This is a suggestion: with the two Google Summer of Code projects (GQB
>> and pgScript) I think that installing a bug tracker (and a feature
>> request list) for pgAdmin should be a good idea. If there is a bug in
>> one part of pgAdmin (GQB, pgScript or pgAdmin) we would be able to:
>> 1) Categorize the bug or the feature
>> 2) Send an email automatically to the person supposed to correct the
>> bug / add the feature
>> 3) Tell when the bug is fixed / the feature added
>>
>> What do you think about that?
>
> It's something Magnus and I have talked about - the problem is there

We did. Let me state for the record that I am in favor of doing it. And
I got the impression that Dave is as well, but not unreservedly (if
that's a word).

> are few good trackers out there imho. Bugzilla in particular which
> many people seem to like is vastly over complicated for anything we'd
> do.

Yes. I definitely would not like to see Bugzilla. I want to see
something that's really simple, but I do think that it would be very
helpful.

I know Dave doesn't entirely agree with me, but I like the tracker
functionality in Trac. Mainly because it's simple, and it does most of
what I want. I'm not saying it has to be Trac, absolutely not, but
that's about the level of complexity that I think is reasonable.

Another option that is worth looking at is Roundup, I think. For an
example of what it looks like, look at http://bugs.python.org. I haven't
installed it myself, but I'm told it's nice and easy to install and use.

One thing, btw, I find important - it should be an actual issue tracker
and *not* a support forum. We don't want a bazillion feature requests or
people not understanding how things are supposed to work to be
registered in there.

> We have also been discussing another topic which may be work rolling
> into this discussion - that is, a move to git over svn (note that the
> decision there is a yes/no one as far as I'm concerned - I have no
> intention of looking at anything else).

I'd say +1 on the move. I've been working a bit with git lately, and I
can certainly see the potential. I think it would particularly be good
for things like these gsoc projects - and it would help in the review of
such big projects as well.

The pgadmin code is also, from a SCM perspective, fairly simple. Thus,
the conversion shouldn't be too difficult.

Another question around it is if we want/need an integrated environment
if we go the step. For example, Trac has a very good integration with
SVN. The integration with git is not at all as good yet.

Personally, I don't think we need this. We could just use gitweb.

//Magnus