[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Bugzilla and SVN integration

From: Luis Malheiro <luis.malheiro_at_rerum.com.br>
Date: 2004-02-05 23:33:29 CET

anton wrote:

> Luis Malheiro wrote:
> >
> > I was wondering, has anyone integrated Bugzilla and Subversion in
> > someway already?
> >
> what do you mean by "integration"? what are your requirements? i would
> be interested myself to learn about possible scenarios of integrating
> cms and bug-tracking system (in its most basic form - storing bug
> numbers in svn properties as the files are changed).
I personally would like to register the actions copy/commit/merge
against a particular bug/issue in Bugzilla, that would include the user
responsible for the action and the the resulting log message.
I also would like to authenticate the SVN users using the same passwords
and user ids stored in the Bugzilla database (MySQL). As a bonus, that
would allow my users to change their passwords through the Bugzilla.
I'm assuming https/http as the protocol for accessing Subversion and I
believe that I could use "mod_auth_mysql" to authenticate users.
I expect the majority of my users to be using TortoiseSVN or another
MSWindows client.

My guess is that the copy/commit events could be collected through a
commit hook, but I don't know what to do about the merge.

I personally favor a SCM model where I would have task branches, one per
bug/issue registered in Bugzilla, and a trunk branch kept by a
merge/build manager responsible for integrating the task branches. I'm
not sure that will be the proper model to use with SVN, but if that
proves itself a reasonable process, I would identify the tasks by the
SVN branches, for example: any action in the branch
"/project/tasks/bug0009977" would be registered against the bug 9977 in
Bugzilla. In that way, my users wouldn't have to deal with properties
and the process turns to be very verbose. Having said that, the use of
properties to hold the bug/issue number could result in a more generic

My view here is heavily influenced by my experience in the shops where I
used ClearCase. One problem I saw in those places was the use of
out-of-the-box integration between the the version control system and
the bug tracker, that would be too generic and would require us to type
a set of bug numbers every time we did a checkout or a commit in a
private branch, which by the process had a bug number as its name... I
found that I bit annoying and would like my users to have a more
transparent system.

Any thoughts on that?

Luis Malheiro

To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Feb 5 23:33:50 2004

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.