As a starting point, we want r329009, but we should probably wait to pick up an LLD fix for the installer tests.

Marking this as blocked on an upstream fix for  https://crbug.com/827082 , then we can flip the blocker relationship.

From previous bug: Want r328905 for  issue 801601 .

r329009 is for  issue 822202 

r329227 for  issue 827082 

I think the bots look decent; started a package here: https://chromium-review.googlesource.com/#/c/chromium/src/+/997735

The roll from #8 insists on failing telemetry_perf_unittests on android_n5x_swarming_rel. It doesn't fail on a whitespace change: https://chromium-review.googlesource.com/c/chromium/src/+/998155

The test runs on swarming so we might be able to bisect.

Could also try rolling further and see if it goes away.

Probably worth trying both.

If we roll past r329300 we have to start manually passing -fmerge-all-constants.

I added -fmerge-all-constants for issue 829795, so that's already done.

We need at least r329581 for Code Coverage fixes and improvements.

We also want r329651 (to reland https://chromium-review.googlesource.com/c/chromium/src/+/981451 )

I can give it a try

fyi you can check https://ci.chromium.org/p/chromium/g/chromium.clang/console to see how tot fares

I'm also working on this.

Fixed some lit tests yesterday: r329558, r329569, D45361

Built a package yesterday for testing: https://chromium-review.googlesource.com/#/c/chromium/src/+/1000784

It still has the telemetry_perf_unittests failure :-(

hans@ please feel free to take this issue

I had to revert the fix for  https://crbug.com/793819  (VS set next statement) in upstream LLVM due to compile time issues. I did it in r329822. If you roll clang past this revision, you will need to add '-mllvm -fast-isel-sink-local-values' to Chromium cflags to avoid regressing that bug. I'll try to put together a proper fix at some point, but I'm out tomorrow, so that may not happen before the next clang roll.

The LLVM bug for the compile time regression is https://llvm.org/pr37010. In practice, enabling code coverage is what tends to hit these long compiles.

There's a lot of red on the ToT waterfall, but the stuff I looked at already seems fixed.

Running tryjobs on a package Nico build yesterday here:
https://chromium-review.googlesource.com/c/chromium/src/+/1012028

And I also prepared a package from this morning here in case we need it:
https://chromium-review.googlesource.com/c/chromium/src/+/1011614 (not pushed to goma yet)

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1f109d9c516640dda836493b24fd033da64411ff

commit 1f109d9c516640dda836493b24fd033da64411ff
Author: Hans Wennborg <hans@chromium.org>
Date: Fri Apr 13 13:23:26 2018

Roll clang 328716:329921.

Bug:  828582 
Change-Id: I85864722c577268d9c95a0d9c37cf7197b04e5d3
Reviewed-on: https://chromium-review.googlesource.com/1012028
Reviewed-by: Nico Weber <thakis@chromium.org>
Commit-Queue: Hans Wennborg <hans@chromium.org>
Cr-Commit-Position: refs/heads/master@{#550608}
[modify] https://crrev.com/1f109d9c516640dda836493b24fd033da64411ff/build/config/compiler/BUILD.gn
[modify] https://crrev.com/1f109d9c516640dda836493b24fd033da64411ff/tools/clang/scripts/update.py

Is this fixed after the cl in c#31?

Yes, this one is done.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1f109d9c516640dda836493b24fd033da64411ff

commit 1f109d9c516640dda836493b24fd033da64411ff
Author: Hans Wennborg <hans@chromium.org>
Date: Fri Apr 13 13:23:26 2018

Roll clang 328716:329921.

Bug:  828582 
Change-Id: I85864722c577268d9c95a0d9c37cf7197b04e5d3
Reviewed-on: https://chromium-review.googlesource.com/1012028
Reviewed-by: Nico Weber <thakis@chromium.org>
Commit-Queue: Hans Wennborg <hans@chromium.org>
Cr-Commit-Position: refs/heads/master@{#550608}
[modify] https://crrev.com/1f109d9c516640dda836493b24fd033da64411ff/build/config/compiler/BUILD.gn
[modify] https://crrev.com/1f109d9c516640dda836493b24fd033da64411ff/tools/clang/scripts/update.py

Next clang roll:  https://crbug.com/834010