Roll clang again |
||||||||
Issue descriptionTracking bug for the next Clang roll. Previous Clang roll: https://crbug.com/834010
,
Apr 27 2018
,
Apr 28 2018
Reid fixed the performance problem with -fast-isel-sink-local-values and enabled it by default in r331088, so we should drop the explicit flag when rolling.
,
Apr 30 2018
When rolling past r331214, we can drop the -Xclang prefix for the -fmerge-all-constants flag with clang-cl.
,
May 1 2018
,
May 2 2018
,
May 4 2018
,
May 7 2018
,
May 7 2018
,
May 8 2018
Trying a roll here: https://chromium-review.googlesource.com/c/chromium/src/+/1049654
,
May 8 2018
There's a goma problem :-( Internal bug b/79412175
,
May 8 2018
While GOMA is blocking the roll, I got a chance to land one more change needed for the code coverage: https://reviews.llvm.org/D46478
,
May 8 2018
That'll likely have to wait for the next roll. We already have packages, it only needs some goma side tweaking to get them deployed.
,
May 8 2018
That's sad, but I guess we can deploy coverage tools bundle separately, using another revision. Is that allowed?
,
May 8 2018
I don't remember how the coverage bundle is distributed. Is it part of the compiler package? If so, that's not possible. Else, if there are no ABI dependencies between compiler and coverage package, it seems fine if update.py doesn't need to be hacked up. I don't see the "that's sad" part though: we can just roll clang again; if we just rolled a follow-up roll usually just works without issues.
,
May 8 2018
The coverage bundle is being packaged and uploaded here: https://cs.chromium.org/chromium/src/tools/clang/scripts/package.py?type=cs&sq=package:chromium&l=339 There might be some dependencies between coverage tools and compiler/linker, as they both use ProfileData. Sure, but I guess the next clang roll will happen in a couple of weeks or so?
,
May 8 2018
We try to roll every two weeks or so, or on demand. You're not blocked on us, you can try rolling yourself. Several others have done so when they wanted something upstream. There's an md file in docs/ describing the process.
,
May 8 2018
Ok, thanks. Just in case the roll doesn't stick this time, please consider including r331801.
,
May 9 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/8f8c2cf49b99861dbf22ba86a68170248fd0f978 commit 8f8c2cf49b99861dbf22ba86a68170248fd0f978 Author: Hans Wennborg <hans@chromium.org> Date: Wed May 09 12:26:39 2018 Roll clang 330570:331747. Bug: 837440 Change-Id: I216c457b53e795d229228e03df8dad42ac78dc07 Reviewed-on: https://chromium-review.googlesource.com/1049654 Reviewed-by: Nico Weber <thakis@chromium.org> Commit-Queue: Hans Wennborg <hans@chromium.org> Cr-Commit-Position: refs/heads/master@{#557156} [modify] https://crrev.com/8f8c2cf49b99861dbf22ba86a68170248fd0f978/build/config/compiler/BUILD.gn [modify] https://crrev.com/8f8c2cf49b99861dbf22ba86a68170248fd0f978/tools/clang/plugins/ChromeClassTester.cpp [modify] https://crrev.com/8f8c2cf49b99861dbf22ba86a68170248fd0f978/tools/clang/scripts/update.py
,
May 10 2018
|
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by h...@chromium.org
, Apr 26 2018