[Buildroot] [autobuild.buildroot.net] Build results for 2013-01-28
gilles.talis at gmail.com
Tue Jan 29 19:46:08 UTC 2013
2013/1/29 Thomas Petazzoni <thomas.petazzoni at free-electrons.com>
> Dear Gilles Talis,
> On Tue, 29 Jan 2013 11:30:45 -0800, Gilles Talis wrote:
> > I could not find in buildroot documentation how these build failures
> > were tracked and what the process was to triage and fix them.
> > These issues do not appear in the bug tracking system.
> > I assume this question has been asked many times, but I could not
> > find the answer to it.
> > Can you briefly describe how build failures are handled?
> We have autobuilders machines that are continuously building random
> package configurations, from a set of predefined toolchain
> configurations. Those run 24/7, and all the autobuilder machines report
> the results of their build on http://autobuild.buildroot.org.
> And then, once a day, an automatic e-mail is sent to the Buildroot
> mailing list, with the list of build that failed.
> There is for now no connection to the bug tracker. It is up to the
> Buildroot community to have a look at the daily e-mail, pick some build
> failure to work on, and submit patches to fix it. When a build failure
> takes a long time to fix, we often send an e-mail to the list to say
> we're working on it, or we discuss it on the Buildroot IRC channel.
> There is no formal process for taking care of a build failure, except
> by sending one or several patches to fix.
> At the moment, we cannot easily integrate this with the bug tracker,
> because a lot of build failures are duplicate. If we submitted one bug
> for each build failure, we would have gazillions of duplicate bugs in
> the bug tracker. It is not easy to automatically detect when two build
> failures are the same, of when two build failures are different. So we
> leave this analysis to Buildroot developers/contributors.
> Do these explanations answer your questions? If you need more details,
> do not hesitate to ask.
> Best regards,
> Thomas Petazzoni, Free Electrons
> Kernel, drivers, real-time and embedded Linux
> development, consulting, training and support.
Thanks a lot for this clear explanation. This answer my questions. Process
seems fair enough to me. My only question for the moment would be how do we
make sure two (or more) people are not working on fixing the same issue? If
I want to fix a specific failure, should I send an email to everyone saying
"hey, I am working on build failure X"? What is the usual process here?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the buildroot