[llvm] r178854 - Buildbot fix for r178851: mistake was in wrong TargetRegisterInfo::getRegClass usage.

[Renato Golin]

On 7 April 2013 13:03, Stepan Dyatkovskiy <stpworld at narod.ru> wrote:

> OK. Which time is most active in ARM direction? I asked because I want to
> sync it with my sleep time. Just, don't want any accidents like this
> anymore :-)
>

If you mean time zones, there's at least UK, China, India and California
time with plenty of ARM folks to ask questions, but PST has a larger volume
all round (not just ARM). Plus, there's always the odd chap during
weekends. ;)

You should be covered all day long, I wouldn't change your sleep patterns
just because of that. I think it's safer to have a decent test
infrastructure than rely on others to spot the problems for you. For ARM, a
simple Pandaboard is enough to get you going on both check-all and
test-suite, though a Chromebook is much faster if you can find it where you
live. I haven't tried the ODroid A15, but it's supposed to be the same
hardware as the Chromebook, and IIRC, shipped worldwide.

I don't expect every one to run tests on ARM, but expect every ARM patch to
have been tested on ARM, and that includes the test-suite, not just the
example in which it was based on. I also expect people to send at least one
regression test (check-all) on all but the most extremely simple and/or
already covered by existing tests. Adding the example/test in which you
found the bug/improvement/feature to the test-suite would be great, too.

By all means, I don't want to discourage you to send patches, but we need
to keep the minimum infrastructure running smoothly or things can go bad
really quickly.

Thanks for your work and persistence, looking forward to new patches! ;)

regards,
--renato
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130407/387864b8/attachment.html>