Roll Clang 218707:220284 (+cherry-picks of 220340, 220403 and 220407)

Roll Clang 218707:220284 (+cherry-picks of 220340, 220403 and 220407)

Also:
- Update the build script to work with the new Android NDK
- Make the package script fail if the update script fails.

BUG=431803
R=thakis@chromium.org

Committed: https://crrev.com/18dd7aa2a0a236c14fb6889d526d3fe6e1dfa08c
Cr-Commit-Position: refs/heads/master@{#304081}

[hans, 2014-11-13 00:46:04]

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

[hans, 2014-11-13 00:46:05]

The trybots look unusually green :) I expect the failures on mac_asan_64 to be
the same on a whitespace change (https://codereview.chromium.org/405473002).

Please take a look.

[Nico, 2014-11-13 00:53:03]

lgtm; the 4.8 question below is maybe interesting though

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/pack...
File tools/clang/scripts/package.sh (right):

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/pack...
tools/clang/scripts/package.sh:76: set -o pipefail
nice!

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
File tools/clang/scripts/update.sh (right):

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:255: done
Since we're likely keeping this cherrypicking nonsense, should this just do `svn
revert -R ${LLVM_DIR} ${CLANG_DIR}`?

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:331: # Presumably the llvm-config flags aren't
quite right.
Is this a merge? If not, is there a bug for this?

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:959: --toolchain=arm-linux-androideabi-4.8 \
What's the 4.8? The gcc version? I think the chrome/android bots use 4.9, is
that important for anything?

[hans, 2014-11-13 01:00:16]

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/pack...
File tools/clang/scripts/package.sh (right):

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/pack...
tools/clang/scripts/package.sh:76: set -o pipefail
On 2014/11/13 00:53:03, Nico wrote:
> nice!

Yeah, this has been bugging me for a long time :)

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
File tools/clang/scripts/update.sh (right):

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:255: done
On 2014/11/13 00:53:03, Nico wrote:
> Since we're likely keeping this cherrypicking nonsense, should this just do
`svn
> revert -R ${LLVM_DIR} ${CLANG_DIR}`?

I'd certainly like that. I thought we'd been avoiding reverting everything so
far to facilitate building with local modifications, but maybe that's not a
common use case?

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:331: # Presumably the llvm-config flags aren't
quite right.
On 2014/11/13 00:53:03, Nico wrote:
> Is this a merge? If not, is there a bug for this?

It's not a merge. I should figure out what's going on here.

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:959: --toolchain=arm-linux-androideabi-4.8 \
On 2014/11/13 00:53:03, Nico wrote:
> What's the 4.8? The gcc version? I think the chrome/android bots use 4.9, is
> that important for anything?

Yes, it's the gcc version. We use this to build the asan runtime. Previously we
were using 4.6 (as some kind of default), but after the NDK roll that isn't
available anymore. I don't think it matters which version we use as long as it
can build the asan runtime.

[Nico, 2014-11-13 01:02:37]

(don't do the below in this cl though, i think)

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
File tools/clang/scripts/update.sh (right):

https://codereview.chromium.org/718513002/diff/40001/tools/clang/scripts/upda...
tools/clang/scripts/update.sh:255: done
> I'd certainly like that. I thought we'd been avoiding reverting everything so
> far to facilitate building with local modifications, but maybe that's not a
> common use case?

Yes, that was the motivation, but listing 15+ files here is pretty silly and
makes it unpredictably dangerous to work in the third_party checkout, I suppose.

[hans, 2014-11-13 18:28:02]

The tryjobs all look good except that PPAPINaClNewlibTest.URLLoader3 in
browser_tests failed on mac_asan_64, but not on the whitespace change. The test
passes for me locally though, so I assume it was just a bot hickup.

[hans, 2014-11-13 18:28:11]

The CQ bit was checked by hans@chromium.org

[commit-bot: I haz the power, 2014-11-13 18:29:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/718513002/60001

[commit-bot: I haz the power, 2014-11-13 18:34:24]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-11-13 18:34:25]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[hans, 2014-11-13 18:44:29]

The CQ bit was checked by hans@chromium.org

[commit-bot: I haz the power, 2014-11-13 18:45:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/718513002/60001

[commit-bot: I haz the power, 2014-11-13 18:50:51]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-11-13 18:50:53]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[hans, 2014-11-13 21:17:41]

Committed patchset #4 (id:60001) manually as
18dd7aa2a0a236c14fb6889d526d3fe6e1dfa08c (presubmit successful).

[commit-bot: I haz the power, 2014-11-13 21:17:42]

Patchset 4 (id:??) landed as
https://crrev.com/18dd7aa2a0a236c14fb6889d526d3fe6e1dfa08c
Cr-Commit-Position: refs/heads/master@{#304081}

[skobes, 2014-11-13 23:05:00]

It looks like this change may have regressed sizes on Linux x64:
http://build.chromium.org/p/chromium/builders/Linux%20x64/builds/73463

Can you investigate and/or revert?

[DaleCurtis, 2014-11-13 23:14:30]

A revert of this CL (patchset #4 id:60001) has been created in
https://codereview.chromium.org/725773002/ by dalecurtis@chromium.org.

The reason for reverting is: Sizes regression, http://crbug.com/433067.

[amineer_google, 2014-11-13 23:37:01]

On 2014/11/13 23:14:30, DaleCurtis wrote:
> A revert of this CL (patchset #4 id:60001) has been created in
> https://codereview.chromium.org/725773002/ by mailto:dalecurtis@chromium.org.
> 
> The reason for reverting is: Sizes regression, http://crbug.com/433067.

Could this roll also be the reason precise32/64 continuous builders are failing
with the following?

FAILED: cd ../../chrome; flock -- /tmp/linux_package_lock bash
../out/Release/installer/rpm/build.sh -o../out/Release -b ../out/Release -a x64
-c unstable
Staging common install files in '/tmp/rpm.build.AMGthY'...
Staging RPM install files in '/tmp/rpm.build.AMGthY'...
Packaging x86_64...
23a24,25
> libgcc_s.so.1()(64bit)
> libgcc_s.so.1(GCC_4.0.0)(64bit)

ERROR: Shared library dependencies changed!
If this is intentional, please update:
chrome/installer/linux/rpm/expected_deps_i386
chrome/installer/linux/rpm/expected_deps_x86_64

[Nico, 2014-11-13 23:40:39]

On Thu, Nov 13, 2014 at 3:37 PM, <amineer@google.com> wrote:

> On 2014/11/13 23:14:30, DaleCurtis wrote:
>
>> A revert of this CL (patchset #4 id:60001) has been created in
>> https://codereview.chromium.org/725773002/ by mailto:
>> dalecurtis@chromium.org.
>>
>
>  The reason for reverting is: Sizes regression, http://crbug.com/433067.
>>
>
> Could this roll also be the reason precise32/64 continuous builders are
> failing
> with the following?
>

It's been reverted, so if that's the case they should green up.


>
> FAILED: cd ../../chrome; flock -- /tmp/linux_package_lock bash
> ../out/Release/installer/rpm/build.sh -o../out/Release -b ../out/Release
> -a x64
> -c unstable
> Staging common install files in '/tmp/rpm.build.AMGthY'...
> Staging RPM install files in '/tmp/rpm.build.AMGthY'...
> Packaging x86_64...
> 23a24,25
>
>> libgcc_s.so.1()(64bit)
>> libgcc_s.so.1(GCC_4.0.0)(64bit)
>>
>
> ERROR: Shared library dependencies changed!
> If this is intentional, please update:
> chrome/installer/linux/rpm/expected_deps_i386
> chrome/installer/linux/rpm/expected_deps_x86_64
>
> https://codereview.chromium.org/718513002/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.