[Intel-gfx] i915 ERRORs and WARN_ON()s

[Daniel Vetter]

Hi Florian,

I fully understand your frustration with all this - you have a bug and
it's annoying and creating bugzilla accounts all over the internet is
indeed not a good way to do things. Unfortunately bugzilla is the
standard, and we don't have the manpower to write our own, nor do the
freedesktop.org admins (who's infrastructure we're using, and who are
all volunteers) have time to transition to other systems.

Also we're positively flooded in bug reports and patch mails (100-200
mails/day on intel-gfx, probably 100 new bugs/months in bugzilla), and
we do expect reporters to e.g. be able to install from source, provide
logs and tons of other things. If creating an account is too much
work, then it's much better to file with someone else who takes care
of the triaging for you (distro or similar), or hire a contractor to
fix your problem for you (there are tons). Bug reports on the m-l that
aren't resolved immediately by pointing at an existing patch are
pretty much guaranteed to get lost.

Again, I fully understand your frustration, but please also take into
account the constraints on our side.

I expect that the technical discussion on the bug itself will continue
in bugzilla. And of course if you want to improve this, help with bug
triaging or anything else then that's very much welcome, but
complaining that there's not an endless supply of free labour that
makes everything work perfectly on both sides is not productive.

Thanks a lot,

Daniel, drm/i915 maintainer


On Sun, May 1, 2016 at 2:11 AM, Florian Zumbiehl <florz at florz.de> wrote:
> Hi,
>
>> > Hu? I don't quite get it. Communicating via email is an unscalable
>> > inconvenience? So, if I asked you to create an account with my todo tracker
>> > instead, would that be more scalable and convenient? Using everyone's
>> > software of choice instead of my own most certainly is neither scalable nor
>> > convenient for me at all.
>>
>> I think you're missing the point here. First of all, your bug report
>> isn't the only bug report. If all bug reports were posted here rather
>> than reported in our bug tracker, it would definitely be more likely
>> that some bugs would get lost. It works for small projects with a low
>> influx of bug reports, but it's less convenient for larger projects.
>> This is the scalability factor.
>
> Did you actually read the email you are replying to completely? If you did,
> you should know what I think a solution that's scalable and convenient for
> both sides would look like, and that it does not consist of you filtering
> bug reports from the mailing lists, so why are you arguing against a
> position that I obviously don't hold?
>
>> Second of all, convenience. Your convenience, while of course a
>> consideration, isn't the primary concern. The convenience of the
>> developers is.  You're reporting one bug, but the developers here
>> have to keep track of, and fix, hundreds (and that's on top of
>> new features, support for new platforms, test cases, and performance
>> improvements).
>
> What do you suppose I do all day long? I get up, report a bug to you, and
> then twiddle my thumbs for the rest of the day? Believe it or not, but
> reporting bugs to you and helping with the debugging isn't the only thing I
> do either.
>
> So, what your argument boils down to is, if I may paraphrase, "our time is
> more valuable than your time, therefore, you should waste your time so we
> don't have to waste ours" (ignoring the fact that you actually wouldn't
> have to). I must say, I am not convinced.
>
>> If we were asking for support for software you develop, you'd obviously
>> be the one to set the rules (and yes, that includes your TODO tracker,
>
> That's not actually how that works. While I am obviously free to set any
> rules I like as to how I engage with people, those don't actually compel
> anyone to obey them. Rather, people who have bugs to report are just as
> free to set any rules they like as to how to engage with people, including
> people they report bugs to. If the two sets of rules are not compatible,
> the bug simply won't get reported/fixed.
>
>> closed forums or what not). Reporting bugs in a bug tracking system
>> isn't exactly a novelty. Rather the opposite.
>
> As mentioned above, I explained in the email that you replied to that I
> don't object to bug tracking systems, so that's kindof a straw man. But
> also, against someone pointing out "That's inefficient and a bit unfair
> because <...>!", "But that's how we've always done it!" isn't really a
> useful argument, is it?
>
> Regards, Florian
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx



-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch