ABRT opt-out (was Re: Summary/Minutes from today's FESCo meeting)

Jiri Moskovcak jmoskovc at redhat.com
Mon Dec 13 18:17:48 UTC 2010


On 12/13/2010 05:43 PM, Kevin Kofler wrote:
> Jiri Moskovcak wrote:
>> This is not a first time when in see this idea and was already answered
>> - we're the distro and we're responsible for the packages, filling all
>> bugs to the upstream will make more harm then good - e.g. crash caused
>> by our patch or by some library which has different upstream or the bug
>> is already fixed in the new upstream version which is not updated in our
>> repositories... and these problems are for maintainers to decide before
>> they forward the ticket to the upstream...
>
> It probably depends on the software (I know that GCC developers, for
> instance, want issues encountered with distro packages to get filed on
> distro trackers), but e.g. Gnash upstream wants to hear about the crashes in
> their software, whereas I don't, because unfortunately I really have no way
> to debug and fix all those crashes. I could spend some time on one or two
> arbitrary ones, but that wouldn't help getting all the others fixed.
>
>          Kevin Kofler
>

- just to make sure we're on the same page - I'm not completely against 
this idea and in fact ABRT was design to be able to this (report bugs in 
different packages to a different bug trackers), right now I see 
following problems

1. Reporting directly to upstream may upset some them
- if we have some policy for that and maintainers negotiate this with 
upstream before we start filling their bugzilla, then we should be ok..
- and maybe we can draw some more attention (and devels..) to ABRT ;-)

2. someone will have to write the specific reporter plugin for reporting 
to different upstream bugtrackers... right now ABRT, can report to: 
bugzilla, kerneloops, e-mail, ftp
(btw even 2 bugzillas don't speak the same XMLRPC :()

J.


More information about the devel mailing list