The ToT bots look decent; trying r301182: https://codereview.chromium.org/2837143002

We probably want r301243 for the android/asan thing too...

That was fast! Looks like it just landed so I'll wait a bit and try again.

Tried r301243: https://codereview.chromium.org/2843513003
There are lit failures on Windows which seem to be tracked upstream in https://bugs.llvm.org/show_bug.cgi?id=32770

The Driver/fuzzer.c test failure I also saw should be fixed by r301246

r301257: https://codereview.chromium.org/2836193002

The windows build got pretty far this time but ran into another problem
https://bugs.llvm.org/show_bug.cgi?id=32777
Reverted in r301274

Trying that in https://codereview.chromium.org/2842663002

goma is ready, running tryjobs on the above

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/09a607ae4637ace31251bd434527772c0aed313f

commit 09a607ae4637ace31251bd434527772c0aed313f
Author: hans <hans@chromium.org>
Date: Tue Apr 25 15:53:03 2017

Roll clang 300839:301274.

BUG= 714769 

Review-Url: https://codereview.chromium.org/2842663002
Cr-Commit-Position: refs/heads/master@{#467001}

[modify] https://crrev.com/09a607ae4637ace31251bd434527772c0aed313f/tools/clang/scripts/update.py

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/cde7b8180ec37363ff3147cfbb40450b29548160

commit cde7b8180ec37363ff3147cfbb40450b29548160
Author: hans <hans@chromium.org>
Date: Tue Apr 25 16:11:19 2017

Revert of Roll clang 300839:301274. (patchset #1 id:1 of https://codereview.chromium.org/2842663002/ )

Reason for revert:
Broke WebKit Mac Builder(dbg)

http://builders/WebKit%20Mac%20Builder%20%28dbg%29/builds/197185

Assertion failed: (LocationKind == Implicit || LocationKind == Unknown)

Original issue's description:
> Roll clang 300839:301274.
>
> BUG= 714769 
>
> Review-Url: https://codereview.chromium.org/2842663002
> Cr-Commit-Position: refs/heads/master@{#467001}
> Committed: https://chromium.googlesource.com/chromium/src/+/09a607ae4637ace31251bd434527772c0aed313f

TBR=thakis@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 714769 

Review-Url: https://codereview.chromium.org/2838903004
Cr-Commit-Position: refs/heads/master@{#467003}

[modify] https://crrev.com/cde7b8180ec37363ff3147cfbb40450b29548160/tools/clang/scripts/update.py

The roll broke the mac debug build:
https://luci-milo.appspot.com/buildbot/chromium.mac/Mac%20Builder%20(dbg)/126873
https://luci-milo.appspot.com/buildbot/chromium.webkit/WebKit%20Mac%20Builder%20(dbg)/197185

I'm surprised the  mac_chromium_compile_dbg_ng trybot didn't catch this.

Did the source change after the tryjob or are we missing some config? I'll investigate.

https://bugs.llvm.org/show_bug.cgi?id=32792 for the assert

mac_chromium_compile_dbg_ng builds with symbol_level = 1 which is why that didn't catch it

Same thing for our ToT bots.

I tried to see if the CQ configs have changed, but it seems the debug bot recipies 'debug_trybot' built with symbol_level=1 'minimal_symbols' for a long time.

I suppose this is fine for the CQ, but we really want symbol_level=2 for our debug ToT bots to cover as much compiler surface as possible.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7fa45855fb213b42550245183ee37ce33d5e0902

commit 7fa45855fb213b42550245183ee37ce33d5e0902
Author: hans <hans@chromium.org>
Date: Tue Apr 25 23:03:14 2017

mb: Don't use minimal_symbols for Clang ToT debug builders

The last Clang roll slipped through the net of ToT and CQ builders though it
would immediately assert on Mac debug builds using the regular symbol level.
Make sure the debug ToT bots don't build with minimal_symbols so we can catch
this in the future.

BUG= 714769 

Review-Url: https://codereview.chromium.org/2838253002
Cr-Commit-Position: refs/heads/master@{#467156}

[modify] https://crrev.com/7fa45855fb213b42550245183ee37ce33d5e0902/tools/mb/mb_config.pyl

Bots looking pretty good this morning. Trying r301382: https://codereview.chromium.org/2844673003

The goma push didn't seem completely successful :-/ Running tryjobs anyway to get whatever signal we can from them.

I gave up on 382 and built and pushed 383 yesterday. That failed due to goma problems, but it seems goma is doing better now, so tryjobs: https://codereview.chromium.org/2846063002

Built 384 (https://codereview.chromium.org/2848923002/) and tried to push that to goma, but failed. I officially give up for this week. Goma bug is b/37783823

For those following along, I still haven't been able to push new packages to goma, and the goma team appears to be on holiday this week.

When rolling past r302190, we need -Wno-microsoft-enum-forward-reference; see https://codereview.chromium.org/2861093003/

Now that I landed it, I realize that we can already pass that flag unconditionally since current clang already knows about it. Maybe we should pass it by default to make the next roll a bit easier.

Yes, please if that's possible.

We would like to have r302239 or higher for issue 716235. Should I file a separate bug for that?

> We would like to have r302239 or higher for issue 716235. Should I file a separate bug for that?

No, let's just track it here.

We need to get past r302578 to fix debug info issues found in  issue 713437 .

r302795 for update.py changes is interesting too (doesn't have to be part of this roll though, not sure if the plan is to build new packages, and if so, at what revision).

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a545baf74563a7bf60c5c6eb2e044e76cad48e0e

commit a545baf74563a7bf60c5c6eb2e044e76cad48e0e
Author: hans <hans@chromium.org>
Date: Thu May 11 19:45:19 2017

Roll clang 300839:301384.

BUG= 714769 

Review-Url: https://codereview.chromium.org/2848923002
Cr-Commit-Position: refs/heads/master@{#471047}

[modify] https://crrev.com/a545baf74563a7bf60c5c6eb2e044e76cad48e0e/tools/clang/scripts/update.py

We need r302272 for issue with cfi, see  bug 718867 .

May also want r303089 for issue 706429.

r303132 would also be nice to fix bug 720551

We need r303267 for a lit test failure

Trying r303273: https://codereview.chromium.org/2884383004

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/027325f153ba7f269b197cac9b43cca8519f08bf

commit 027325f153ba7f269b197cac9b43cca8519f08bf
Author: Hans Wennborg <hans@chromium.org>
Date: Thu May 18 04:35:33 2017

Roll clang 301384:303273.

BUG= 714769 
R=thakis@chromium.org

Review-Url: https://codereview.chromium.org/2884383004 .
Cr-Commit-Position: refs/heads/master@{#472665}

[modify] https://crrev.com/027325f153ba7f269b197cac9b43cca8519f08bf/tools/clang/scripts/update.py

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d6dc38b8713ca8ac2fd49ee08f0966d221298486

commit d6dc38b8713ca8ac2fd49ee08f0966d221298486
Author: treib <treib@chromium.org>
Date: Thu May 18 11:54:59 2017

Revert of Roll clang 301384:303273. (patchset #1 id:1 of https://codereview.chromium.org/2884383004/ )

Reason for revert:
Speculative revert: This might have broken cc_unittests on Linux MSan: https://build.chromium.org/p/chromium.memory/builders/Linux%20MSan%20Tests
(In fact, the same tests failed on the linux_chromium_msan_rel_ng right here.)

The failing tests are related to shader transpiling which uses ANGLE which uses clang. Kind of a long shot, but nothing else looks related.

Original issue's description:
> Roll clang 301384:303273.
>
> BUG= 714769 
> R=thakis@chromium.org
>
> Review-Url: https://codereview.chromium.org/2884383004 .
> Cr-Commit-Position: refs/heads/master@{#472665}
> Committed: https://chromium.googlesource.com/chromium/src/+/027325f153ba7f269b197cac9b43cca8519f08bf

TBR=thakis@chromium.org,hans@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 714769 

Review-Url: https://codereview.chromium.org/2888363002
Cr-Commit-Position: refs/heads/master@{#472771}

[modify] https://crrev.com/d6dc38b8713ca8ac2fd49ee08f0966d221298486/tools/clang/scripts/update.py

Trying r303369: https://codereview.chromium.org/2897453002/

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f646505961b549caf4699c3cda9856c061bec224

commit f646505961b549caf4699c3cda9856c061bec224
Author: hans <hans@chromium.org>
Date: Fri May 19 04:39:40 2017

Roll clang 301384:303369.

BUG= 714769 

Review-Url: https://codereview.chromium.org/2897453002
Cr-Commit-Position: refs/heads/master@{#473087}

[modify] https://crrev.com/f646505961b549caf4699c3cda9856c061bec224/tools/clang/scripts/update.py

It seems to be sticking.