[Ksummit-2012-discuss] [ATTEND] kernel core dump and "dying breath"

Cong Wang xiyou.wangcong at gmail.com
Thu Jun 21 14:55:07 UTC 2012


On 06/21/2012 10:43 PM, Jonathan Corbet wrote:
> On Thu, 21 Jun 2012 09:35:32 -0400
> "John W. Linville"<linville at tuxdriver.com>  wrote:
>
>> YMMV -- my experience with ABRT is that it provides little useful
>> information by itself and that the reported bugs rarely have reporters
>> interested in following-up with more information.  Plus, it does a
>> terrible job correlating new bug reports with existing bug reports.
>>
>> I would vote against cluttering bugzilla.kernel.org with nearly
>> useless automated bug reports...
>
> <metoo>
>
> Running Rawhide gives one plenty of opportunity to report bugs; in the
> process, I often find the Fedora bugzilla choked with ABRT reports.  I
> can't imagine that flood of stuff is useful to anybody, and it makes
> life hard for people who want to try to avoid putting in duplicate
> reports.  ABRT quality is low, and I can easily see it driving away
> higher-quality reports from people who are actually willing to follow
> up.
>

I think humans still need to be involved to check the duplication of bug 
reports. It is a hard problem for ABRT to detect if two kernel crashes 
are identical, we do need to some clever algorithms here.


More information about the Ksummit-2012-discuss mailing list