llvm dev - Oct 2011 - [LLVMdev] Reminder: LLVM 3.0 Branching Friday!

This is just a reminder to say that we will be branching for the LLVM 3.0
release Friday!

	07:00:00 p.m. Friday October 14, 2011 PDT
	02:00:00 a.m. Saturday October 15, 2011 GMT

Now is the time to look at the buildbots and see what fixes they need:

	http://lab.llvm.org:8011/console

As of this writing, we have:

• several test failures on llvm-gcc self-host:

	http://lab.llvm.org:8011/builders/llvm-gcc-i386-linux-selfhost/builds/25

• a miscompile in one of the llvm-gcc testers:

	http://lab.llvm.org:8011/builders/llvm-x86_64-linux-checks/builds/55

Comparing stages 2 and 3
warning: ./cc1plus-checksum.o differs
Bootstrap comparison failure!
./cp/pt.o differs

• many test failures on mingw32:

	http://lab.llvm.org:8011/builders/clang-native-mingw32-win7/builds/5

Please take some time to look at these and determine if they are real problems
or can be ignored. Having the buildbots green at the time we branch is the main
goal! :-)

Pay close attention to buildbot failures! If you check-in something and one of
the buildbots breaks, please either fix is quickly or revert the patch until a
fix can be submitted.

With your help, we will have a very smooth 3.0 release.

Thank you!
-bw

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20111013/30ba04b4/attachment.html>

Hi Bill,
> As of this writing, we have:
>
> • several test failures on llvm-gcc self-host:
>
> http://lab.llvm.org:8011/builders/llvm-gcc-i386-linux-selfhost/builds/25
I can reproduce this here.  I'll poke at it this weekend sometime.
> • a miscompile in one of the llvm-gcc testers:
>
> http://lab.llvm.org:8011/builders/llvm-x86_64-linux-checks/builds/55
>
> Comparing stages 2 and 3
> warning: ./cc1plus-checksum.o differs
> Bootstrap comparison failure!
> ./cp/pt.o differs
I bisected this to commit 139841 by Andrew Trick.  He said he would look at it.

Ciao, Duncan.

On Oct 14, 2011, at 6:31 AM, Duncan Sands wrote:
>> • a miscompile in one of the llvm-gcc testers:
>> 
>> http://lab.llvm.org:8011/builders/llvm-x86_64-linux-checks/builds/55
>> 
>> Comparing stages 2 and 3
>> warning: ./cc1plus-checksum.o differs
>> Bootstrap comparison failure!
>> ./cp/pt.o differs
> 
> I bisected this to commit 139841 by Andrew Trick.  He said he would look at
it.
I was able to reproduce the problem, but only during a parallel make of the
entire llvm-gcc tree and not consistently. I have not been able to reproduce yet
with a single build step, which is making this quite painful.

I did carefully inspect all of the differences between the .o files and proved
to myself that they are functionally equivalent. It appears to be random codegen
behavior. But I don't understand yet why self-hosting the front end could
affect it.

Duncan, if you provide the differing .o files I can take a look and see if I
notice a pattern. Better yet, provide an address-stripped asm diff in case my
binutils are inadequate. We can move this discussion to PR11133.

You mentioned that you need a loaded system to reproduce it. I agree, but
can't understand the relationship between system load and codegen. Addresses
are randomized either way.

-Andy

Bill Wendling <wendling at apple.com> writes:
> This is just a reminder to say that we will be branching for the LLVM 3.0
release Friday!
>
> 07:00:00 p.m. Friday October 14, 2011 PDT
> 02:00:00 a.m. Saturday October 15, 2011 GMT
Hi Bill,

I have some pending patches for 3.0 only.  Will you be sending out an
announcement when the split is finished and such patches can be
committed?

Thanks!

                             -Dave

greened at obbligato.org (David A. Greene) writes:
> I have some pending patches for 3.0 only.
Of course I meant post-3.0 only.

                           -Dave

2011/10/14 Bill Wendling <wendling at apple.com>:
> • many test failures on mingw32:
> http://lab.llvm.org:8011/builders/clang-native-mingw32-win7/builds/5
FYI, I have not met. (--enable-optimized).
http://bb.pgr.jp/builders/clang-i686-msys
I have suppressed a few failures with local patch on my builder, though.

...Takumi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: clang.patch
Type: application/octet-stream
Size: 741 bytes
Desc: not available
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20111015/b014ca32/attachment.obj>