On Tue, Jul 19, 2011 at 11:38 AM, Amit Kulkarni <amitk...@gmail.com> wrote:
> Can you elaborate? Where they suck? RT: written in perl, painful to upgrade (painful enough, that we've not touched ours in over a year). Ugly interface, but that's the least of its problems. Without a good way to manage users, access, or set up quickly through the UI, it's easier to try to manipulate the DB tables. Perhaps I'm just doing it wrong(tm). So far I've not had enough time to really track upgrades easily or quickly, and haven't had time to fix all the infrastructure that it sits on (MySQL, perl versions, libs, etc) to ensure an upgrade goes cleanly. The biggest advantage RT provides is easy creation of new tickets through email, but it still takes a human on the other end to actually classify what that ticket is. It's bad enough that at my work, we have a general Operations email, that we then handle tickets in the ops group. It wastes time, but it's easier than dealing with engineering misfiring a ticket. Then there's creating sub-users of a larger account... TRAC: nice integration with SVN, but still limited by a complex ACL system and the fact SVN doesn't provide a good user management system in itself, preferring system users (or PAM auth, LDAP, etc). Trouble is that it's not a good general ticket tracking system, and breaks just often enough to be annoying to admin. Given that I have to deal with at work, I don't have time to babysit TRAC's stupid more often than I care for. TRAC also suffers from trying to please a bunch of different people with different needs at once. Is it floorwax or a dessert topping? Wait, no, it's BOTH! Bugzilla: Perl. OpenSource UI, backend of pain (MySQL, PostgreSQL, or SQLite3!). I've not used it (administered) in a few years, but my experience with it has never been close to what one would call "positive." Painful, breaks in weird ways, and sometimes just had errors. Haven't used Jira yet. So, I have no opinion. I don't think bug tracking needs to be difficult, ugly, or annoying to navigate. The problem is that every bug tracking utility is built to solve problems for a large set of implementors. Not, say, solve one specific need really well. Many violate the prime directive of dealing with software and users: KISS.
