state of bugzilla (and, current kernel root exploits....)

[Matthew Miller]

On Wed, Mar 30, 2005 at 09:49:27AM -0700, Michal Jaegermann wrote:
> > Done: <https://bugzilla.redhat.com/beta/show_bug.cgi?id=152532>
> By "Done" you mean that you managed to enter that report or that
> old data were imported? I tried to search for all Legacy bugs

Done as in "I did what was suggested, and entered the report." Sorry. Issue
is not done at all.

> I also wonder about "Version" field which is really a distribution
> tag.  It looks like that it can be set only to "core1", or "rhl9"
> or "rhl7.3" in an exclusive way.  If there is a bug which affects
> multiple distributions, which most often is the case although
> patch particulars may differ, then a bug report should be cloned
> across all these?  In particular #152532 clearly affects not
> only rhl9 but it is marked as such.

Yeah, this is a little bit of an issue. For tracking current state properly,
it's better to have separate bugs for each version -- they're not generated
from the same spec file, and they don't end up going to the same place, and
different distro versions will be affected differently.

This can feel a little bit like administrivia, but my experience with using
bugzilla to track the BU Linux project has convinced me that it's really
worth it and makes the whole process less error-prone.

The other downside, of course, is that it's kinda convenient to
"cross-pollinate" between development/QA for closely related issues. For our
own bugzilla-managed BU Linux project, we simply do that by making sure each
bug has a comment pointing to the bug # of the other one.

Another approach would be to add a version named "meta" or "tracking", and
create initial bugs like "CAN-005-468: rh9, fc1, fc2" and then create
individual bugs for each release and have them all block the top-level
tracking bug. This is even *more* administrative overhead, though.

I don't use RHL 7.x or FC1, so I only filed the RHL 9 one; I probably should
have filed the others as well. I can go ahead and do that, but I'll wait a
little bit to see what comes out of this discussion.

-- 
Matthew Miller           mattdm at mattdm.org        <http://www.mattdm.org/>
Boston University Linux      ------>                <http://linux.bu.edu/>