[quagga-dev 5536] Re: Bug tracking system for Quagga

Jeffrey Ollie jeff at ocjtech.us
Tue Jul 1 14:13:21 BST 2008


On Tue, Jul 1, 2008 at 7:39 AM,  <paul at clubi.ie> wrote:
> On Mon, 30 Jun 2008, Jeffrey Ollie wrote:
>
>> Almost all Trac functionality is implemented as plugins, some are just not
>> distributed with the main tarball.  That said, the SCM plugins other than
>> the SVN plugin are of varying quality.  An example what you get with the Git
>> plugin is here:
>>
>> https://fedorahosted.org/koji/browser
>
> Interesting, so that's a fairly major project that is using Trac with an SCM
> /other/ than SVN?

Yes, they use Git.  There are many more projects using Git or SCMs
other than SVN at https://fedorahosted.org/web/

> What are the downsides to Trac? Two I've heard elsewhere are:
>
> - new users have to be added via the command line (does the webadmin
>  plugin fix that?)

Trac doesn't really do any user management/authentication on it's own
- it defers all that to Apache (or whatever web server you are using).
 However, "users" are only needed to perform operations that you want
to restrict to certain people.  Depending on how you set up
permissions end users can report bugs without logging in at all.

Fedora has a web system for creating accounts that we've integrated
with Fedora Hosted but that's probably overkill for Quagga.

> - upgrades might be tricky (but same can be said of bugzilla)

I don't have a lot of experience with upgrading Trac projects, but
it's supposed to be as simple as:

1. Install new source code
2. trac-admin /path/to/projenv upgrade

http://trac.edgewall.org/wiki/TracUpgrade

> If we were to try Trac, does it have any bug-data export functionality in
> case things go wrong (e.g. bugzilla has an XML interface (though it doesnt
> export attachments in the version we have, grr..))?

Trac stores all bug data in a SQL database so you'd be able to use the
database dump commands to get the raw data.

Jeff



More information about the Quagga-dev mailing list