Roll Clang 247874:254049

Roll Clang 247874:254049

Also revert https://codereview.chromium.org/1375193005 for bootstrap builds,
else compiler-rt won't build in that config.  See comment 45 on the bug.

This also pins MSan to the old ABI:

MSan had a ABI change upstream. Picking it up in Chromium requires
rebuilding prebuilt_instrumented_libraries with the new clang.
Unfortunately, this fails.

This changes switches clang back to the old MSan ABI until we can
sort this out.

Upstream issue:
https://github.com/google/sanitizers/issues/579

BUG=538614, 559530, 560589
R=thakis@chromium.org

Committed: https://chromium.googlesource.com/chromium/src/+/4fea0f71e935f43777c5f0c0b488910f9e7490ad

[hans, 2015-11-25 05:38:30]

hans@chromium.org changed reviewers:
+ thakis@chromium.org

[hans, 2015-11-25 05:38:30]

Please take a look.

(And if you like, CQ if it looks good, for additional time zone usage :D)

[hans, 2015-11-25 05:39:28]

Description was changed from

==========
Roll Clang 247874:254049

This also pins MSan to the old ABI:

MSan had a ABI change upstream. Picking it up in Chromium requires
rebuilding prebuilt_instrumented_libraries with the new clang.
Unfortunately, this fails.

This changes switches clang back to the old MSan ABI until we can
sort this out.

Upstream issue:
https://github.com/google/sanitizers/issues/579

BUG=559530,560589
==========

to

==========
Roll Clang 247874:254049

This also pins MSan to the old ABI:

MSan had a ABI change upstream. Picking it up in Chromium requires
rebuilding prebuilt_instrumented_libraries with the new clang.
Unfortunately, this fails.

This changes switches clang back to the old MSan ABI until we can
sort this out.

Upstream issue:
https://github.com/google/sanitizers/issues/579

BUG=538614, 559530, 560589
==========

[hans, 2015-11-25 05:40:27]

Description was changed from

==========
Roll Clang 247874:254049

This also pins MSan to the old ABI:

MSan had a ABI change upstream. Picking it up in Chromium requires
rebuilding prebuilt_instrumented_libraries with the new clang.
Unfortunately, this fails.

This changes switches clang back to the old MSan ABI until we can
sort this out.

Upstream issue:
https://github.com/google/sanitizers/issues/579

BUG=538614, 559530, 560589
==========

to

==========
Roll Clang 247874:254049

Also revert https://codereview.chromium.org/1375193005 for bootstrap builds,
else compiler-rt won't build in that config.  See comment 45 on the bug.

This also pins MSan to the old ABI:

MSan had a ABI change upstream. Picking it up in Chromium requires
rebuilding prebuilt_instrumented_libraries with the new clang.
Unfortunately, this fails.

This changes switches clang back to the old MSan ABI until we can
sort this out.

Upstream issue:
https://github.com/google/sanitizers/issues/579

BUG=538614, 559530, 560589
==========

[Nico, 2015-11-25 15:26:02]

lgtm. I have a few nit comments, but I'll hit cq anyways. Looks like the msan
bot failed with "goma has no compiler" and the new msan run started just 2
minutes ago, so I'll probably wait until that's mostly done before I hit cq.

https://codereview.chromium.org/1474923002/diff/1/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/1474923002/diff/1/build/common.gypi#newcode4538
build/common.gypi:4538: '-fPIC',
I suppose these should get a "TODO: Remove once we use the new msan abi,
crbug.com/XXX" (also in the gn file)

https://codereview.chromium.org/1474923002/diff/1/tools/clang/scripts/update.sh
File tools/clang/scripts/update.sh (right):

https://codereview.chromium.org/1474923002/diff/1/tools/clang/scripts/update....
tools/clang/scripts/update.sh:536: # TODO: revert this when instrumented
libraries can be rebuilt with the new ABI.
eugenis already filed a tracking bug for this that this should poitn to

[hans, 2015-11-25 15:37:40]

On 2015/11/25 15:26:02, Nico wrote:
> Looks like the msan
> bot failed with "goma has no compiler" and the new msan run started just 2
> minutes ago

There are 3 msan tryjobs actually. The first one failed with goma problems as
you say. The second one seemed to have browser_test problems:
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...
but I'm not sure what the actual errors is.

I just started the third one to see if that makes any difference.


I'll address your other comments when I get into the office.

[Nico, 2015-11-25 15:42:38]

Oh, ok. The browser_tests problems seem "normal" on that bot. Hitting cq
then.

On Wed, Nov 25, 2015 at 10:37 AM, <hans@chromium.org> wrote:

> On 2015/11/25 15:26:02, Nico wrote:
>
>> Looks like the msan
>> bot failed with "goma has no compiler" and the new msan run started just 2
>> minutes ago
>>
>
> There are 3 msan tryjobs actually. The first one failed with goma problems
> as
> you say. The second one seemed to have browser_test problems:
>
>
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...
> but I'm not sure what the actual errors is.
>
> I just started the third one to see if that makes any difference.
>
>
> I'll address your other comments when I get into the office.
>
> https://codereview.chromium.org/1474923002/
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Nico, 2015-11-25 15:42:38]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2015-11-25 15:43:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1474923002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1474923002/1

[commit-bot: I haz the power, 2015-11-25 17:35:29]

Patchset 1 (id:??) landed as
https://crrev.com/4fea0f71e935f43777c5f0c0b488910f9e7490ad
Cr-Commit-Position: refs/heads/master@{#361689}

[hans, 2015-11-25 17:35:56]

Committed patchset #1 (id:1) manually as
4fea0f71e935f43777c5f0c0b488910f9e7490ad (presubmit successful).

[hans, 2015-11-25 17:55:13]

https://codereview.chromium.org/1474923002/diff/1/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/1474923002/diff/1/build/common.gypi#newcode4538
build/common.gypi:4538: '-fPIC',
On 2015/11/25 15:26:02, Nico wrote:
> I suppose these should get a "TODO: Remove once we use the new msan abi,
> crbug.com/XXX" (also in the gn file)

Done.

https://codereview.chromium.org/1474923002/diff/1/tools/clang/scripts/update.sh
File tools/clang/scripts/update.sh (right):

https://codereview.chromium.org/1474923002/diff/1/tools/clang/scripts/update....
tools/clang/scripts/update.sh:536: # TODO: revert this when instrumented
libraries can be rebuilt with the new ABI.
On 2015/11/25 15:26:02, Nico wrote:
> eugenis already filed a tracking bug for this that this should poitn to

Done.

[drott, 2015-11-26 12:46:15]

This seems to break Mac goma builds for me: 

When I revert this change, my Mac goma builds start to work again, otherwise all
goma builds fail when trying to include stddef.h 

In file included from
../../third_party/webrtc/modules/desktop_capture/screen_capturer.cc:11:
In file included from
../../third_party/webrtc/modules/desktop_capture/screen_capturer.h:14:
In file included from
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/vector:265:
In file included from
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/__bit_reference:15:
In file included from
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/algorithm:624:
In file included from
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/initializer_list:47:
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/cstddef:38:10:
fatal error: 'stddef.h' file not found
#include <stddef.h>
^
1 error generated.

[Nico, 2015-11-26 13:25:07]

drott: You need to update your goma client (~/goma/goma_ctl.py update, or
upgrade, or something like that). Does that help? (I thought goma
autoupdated--maybe it doesn't?)
On Nov 26, 2015 7:46 AM, <drott@chromium.org> wrote:

> This seems to break Mac goma builds for me:
>
> When I revert this change, my Mac goma builds start to work again,
> otherwise all
> goma builds fail when trying to include stddef.h
>
> In file included from
> ../../third_party/webrtc/modules/desktop_capture/screen_capturer.cc:11:
> In file included from
> ../../third_party/webrtc/modules/desktop_capture/screen_capturer.h:14:
> In file included from
>
>
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/vector:265:
> In file included from
>
>
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/__bit_reference:15:
> In file included from
>
>
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/algorithm:624:
> In file included from
>
>
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/initializer_list:47:
>
>
/Users/drott/dev/blink/src/third_party/llvm-build/Release+Asserts/bin/../include/c++/v1/cstddef:38:10:
> fatal error: 'stddef.h' file not found
> #include <stddef.h>
> ^
> 1 error generated.
>
> https://codereview.chromium.org/1474923002/
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[drott, 2015-11-30 11:52:17]

On 2015/11/26 13:25:07, Nico wrote:
> drott: You need to update your goma client (~/goma/goma_ctl.py update, or
> upgrade, or something like that). Does that help? (I thought goma
> autoupdated--maybe it doesn't?)

Thanks for pointer, Nico. That worked in the end. Emphasis on using the .py
script - I used goma_ctl.sh update first, which told me everything was
up-to-date at version 94. But it still failed. Using the .py it worked.
Regarding the auto update, I think so, too - maybe I had it running before
updating my source including this CL.