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

Re: Issue tracker in subversion

From: David Weintraub <qazwart_at_gmail.com>
Date: Tue, 14 Apr 2009 11:34:12 -0400

2009/4/13 Grégoire Welraeds <gwelraeds_at_eastnets.com>:
> Hello,
> Just a quick question. Is there any plan to make an issue tracker inside
> subversion?
> I have looked at both DITrack and subissue but they both looks like dead
> and/or buggy projects.
> Does anyone have valuable input on this?

I just wanted to say a few words about integrating defect tracking
with version control.

Back in the 1960s, the Germans produced the Amphicar It was a mix of
a car and a boat. You could drive it to the lake and right into the
lake. Well, it turned out that it was both a lousy car and a lousy
boat: <http://www.time.com/time/specials/2007/article/0,28804,1658545_1658498,00.html>

Most integrated version control/defect tracking packages are either
bad defect tracking systems, bad version control systems, or both.

I've used Sablime, PVCS, ClearCase/ClearQuest, StarTeam, and a few
more I can't quite remember. In almost all of them, whatever level of
integration existed between the version control side and the defect
tracking side was a disappointment. You ended up with a lousy version
control system, a lousy defect tracking system, and an integration
that was no better than I could design with a few hooks and a late
night on a coffee buzz.

In ClearCase/ClearQuest, the ClearQuest issue ID was placed on the
revision of the file that was fixed, and the file's revision ID was
included in the ClearQuest side. However, something as simple as
asking "What files and revisions were fixed with ClearQuest issue
#1005?" couldn't be done. Yes, there was a field -- a very hard to
read field that contained the name of the file (as well as the view's
name), but this field simply could not be queried. Nor, could you
easily go to the ClearCase side and say "Show me all the files and
their check in comments that went with issue #1005". Such a query
could take hours. The entire integration depended upon a ClearCase
Perl trigger that had to be executed on everyone's system. We had over
150 users and spent almost all of our time installing software
whenever someone got a new computer.

You're almost always better finding the defect tracking and version
control systems that do the job you need. In almost all cases, a good
defect tracking system and a good version control system will contain
the hooks you need to do the integration. I know that Seapine not only
has a Subversion integration, but also integrates with IDEs which is
even more important. One of the things developers hate is having to go
into another piece of software to do their work. I don't know how well
TestTrack integrates with Subversion, but I take it that it's no worse
than most integrated suites.

David Weintraub
To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_subversion.tigris.org].
Received on 2009-04-14 17:35:08 CEST

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