An interesting read when discussing what to do about our bugs...

Matej Cepl mcepl at redhat.com
Sat Jan 19 22:19:56 UTC 2008


On 2008-01-19, 18:10 GMT, seth vidal wrote:
> Hmm, is that what the 'upstream' close reason is for? Normally, 
> I close things 'upstream' when I have checked a fix into the 
> upstream code base. Which seems pretty reasonable time to close 
> it to me.

OK, this is something which should be well described in the 
http://fedoraproject.org/wiki/JonStanley/BugWorkflow (Jon?).  
Closing upstream is from the maintainer's point of view pretty 
difficult thing to do. First rule, which I try to always hold 
with Xorg bugs is that the bug should *NEVER* be closed as 
UPSTREAM unless I have number of the upstream bug in the External 
Bugzilla References (or in the comment, if the upstream bugzilla 
is not available among external bugzilla references; fortunately, 
that doesn't happen for the bugs I triage). So, even if I ask 
reporter to file the bug upstream (and I do it less and less), 
I put them in NEEDINFO, and close the bug as UPSTREAM only when 
they come bug with the number of the upstream bug.

I totally understand the reasons why we ask reporters to file the 
bug, and why it is problematic to file a bug ourselves (after 
all, it is THEIR bug, they can reproduce it, test the fix, 
provide additional information, etc.), but still I do it less and 
less. Now I usually try to find a duplicate bug in the upstream 
bugzilla (trip into Mozilla bugzilla always cheers me up -- if 
our bugzilla is mess, and it is, I have no words to describe 
their bugzilla ;-)), and suggest to the reporter to comment on it 
and join the happy crowd around that bug in the upstream 
bugzilla.

The reason is that asking them to file a bug upstream (internally 
called "a request for suicide") is very bad from the marketing 
point of view. I feel that we should treat the bug reporters as 
our most valuable asset (which they are), and to be very careful 
not to alieante them. If there is a person, who is willing to 
file a reasonably thought-through bug to our bugzilla, it is 
a person, which is extremely valuable to us (because that's the 
only QA we have in Fedora, and because open-source software is 
bug driven).

However, I don't think it is necessary to close bug upstream only 
when I have actually fixed it (unless you are a primary developer 
of the particular component, as seth is). Whole point of 
upstreaming bugs (and you can read it in 
https://bugzilla.redhat.com/page.cgi?id=fields.html#upstream 
which is actually pretty good resource to read) is that everybody 
from all distributions can join their efforts about fixing the 
bug, so the issue doesn't have to be fixed multiple times in each 
distro independently. It is always a good idea to write something 
in such sense to the comment by which you either ask reporter to 
file a bug upstream, or by the one by which you close the bug.

Just my 0.02 CZK ;-)

Matěj




More information about the fedora-devel-list mailing list