Tom Lane wrote:
> "Magnus Hagander" <mha@sollentuna.net> writes:
> >> ... Red Hat's present bugzilla system
> >> could be described that way --- and while I can't say I'm in
> >> love with it, I can deal with it.
>
> > Doesn't bugzilla insist on sending you the complete bug every time?
>
> Nope, it just sends the changes/additions. Other than the lack of a
> direct email input method, I find BZ quite usable. Josh was just
> complaining that its source code is a mess (dunno, haven't looked)
> but other than that I think it's a definite possibility, just because
> so many people are already familiar with it.
Have you tried to use debbugs? I agree with Greg Stark that it's a
better fit for our current procedure, while enabling better
traceability.
For an example, see http://bugs.debian.org. There are three links there
pointing to pages on how to use the system. Entering a bug number shows
detail; for example try entering 330514 which is a PostgreSQL bug. You
can add more detail to a bug by mailing <bug-number>@bugs.debian.org.
You can close a bug by mailing <bug-number>-done@bugs.debian.org. You
can of course clone bugs, retarget to a different package, merge bugs,
etc.
It's controllable by email -- in fact, I think email is the only
controlling interface. You can get reports using the web frontend. You
can get an mbox via HTTP for a particular bug, which you can later open
with your email client if you like. (And respond to it, etc).
We would have to determine what constitutes a "package" (probably one
for each contrib module, one for each interface, one for the backend,
etc; or we could have separate package for "optimizer", "rewriter",
"transaction system", one for each access method, etc), what "tags"
there are, what "versions", etc.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support