Add MB configs for ClangToTLinux* bots on FYI

Add MB configs for ClangToTLinux* bots on FYI

The tip-of-tree Clang bots rely on an environment variable
called LLVM_FORCE_HEAD_REVISION to be set in order to know
to fetch the tip-of-tree Clang and use its version in the compile
time defines, rather than the checked-in version.

We don't want GN to use or depend on environment variables,
so this patch modifies the clang update script to also look at
a new command line arg, and adds a GN arg called 'llvm_force_head_revision' to cause the same effect.

However, we need to keep the env var around for `gclient runhooks`
and to keep the existing bots (and devs) working while on GYP,
and, perhaps more importantly, adding a new GYP_DEFINE that
does what we need to build/common.gypi is proving intractable,
so we hack MB to rewrite what the GYP_DEFINE would be if life
made sense back into the env var.

BUG=481692, 582737
R=thakis@chromium.org, brettw@chromium.org

Committed: https://crrev.com/909bad6afa7124f67c4890f3c1bc244179141978
Cr-Commit-Position: refs/heads/master@{#389408}

[Dirk Pranke, 2016-04-22 01:05:39]

The CQ bit was checked by dpranke@chromium.org

[commit-bot: I haz the power, 2016-04-22 01:05:57]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1908293002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1908293002/1

[krasin, 2016-04-22 01:09:13]

krasin@google.com changed reviewers:
+ krasin@google.com

[krasin, 2016-04-22 01:09:14]

lgtm

the comment is not super critical, feel free to ignore / fix later.

https://codereview.chromium.org/1908293002/diff/1/tools/mb/mb_config.pyl
File tools/mb/mb_config.pyl (right):

https://codereview.chromium.org/1908293002/diff/1/tools/mb/mb_config.pyl#newc...
tools/mb/mb_config.pyl:2265: 'ubsan_hack': {
Please, rename it to "ubsan_no_recover_hack" to make it more meaningful.

[Dirk Pranke, 2016-04-22 01:14:24]

https://codereview.chromium.org/1908293002/diff/1/tools/mb/mb_config.pyl
File tools/mb/mb_config.pyl (right):

https://codereview.chromium.org/1908293002/diff/1/tools/mb/mb_config.pyl#newc...
tools/mb/mb_config.pyl:2265: 'ubsan_hack': {
On 2016/04/22 01:09:14, krasin wrote:
> Please, rename it to "ubsan_no_recover_hack" to make it more meaningful.

Unfortunately, that'd push the config name on line 1559 past 80 columns, which I
really want to avoid.

Feel free to suggest a better name for the config as part of a change, though.

[Nico, 2016-04-22 01:26:38]

lgtm

I remember that it's important that LLVM_FORCE_HEAD_REVISION is in the env while
build files are generated -- is there a way to express this in mb? (There's a
bug somewhere where we discuss this in some detail)

[Dirk Pranke, 2016-04-22 01:30:47]

On 2016/04/22 01:26:38, Nico wrote:
> lgtm
> 
> I remember that it's important that LLVM_FORCE_HEAD_REVISION is in the env
while
> build files are generated -- is there a way to express this in mb? (There's a
> bug somewhere where we discuss this in some detail)

Ah, yeah, crbug.com/582737 . We don't currently have a solution for this.

Let me see if I can put together something in another CL so we don't get blocked
on this.

That's unfortunate that bug is not on the radar for the GN migration, which I
will fix now.

[Dirk Pranke, 2016-04-22 01:52:30]

The CQ bit was unchecked by dpranke@chromium.org

[Dirk Pranke, 2016-04-22 02:17:55]

Patchset #2 (id:20001) has been deleted

[Dirk Pranke, 2016-04-22 22:31:01]

dpranke@chromium.org changed reviewers:
+ brettw@chromium.org
- aneeshm@chromium.org, krasin@chromium.org, krasin@google.com

[Dirk Pranke, 2016-04-22 22:31:01]

Okay, this is yet another approach for this stupid problem,
whereby we make MB present a sane interface to the user
and hack the env we pass to GYP so that we don't have
to figure out how to get a new variable working in 
build/common.gypi.

Nico, please take a look and let me know if this looks like it'll work to you.

[Dirk Pranke, 2016-04-22 22:37:50]

Description was changed from

==========
Add MB configs for ClangToTLinux* bots on FYI

This is the CL from https://codereview.chromium.org/1910333002/
with a couple of fixes.

BUG=481692
TBR=aneeshm@chromium.org, thakis@chromium.org, krasin@chromium.org
==========

to

==========
Add MB configs for ClangToTLinux* bots on FYI

The tip-of-tree Clang bots rely on an environment variable
called LLVM_FORCE_HEAD_REVISION to be set in order to know
to fetch the tip-of-tree Clang and use its version in the compile
time defines, rather than the checked-in version.

We don't want GN to use or depend on environment variables,
so this patch modifies the clang update script to also look at
a new command line arg, and adds a GN arg called 'llvm_force_head_revision' to
cause the same effect.

However, we need to keep the env var around for `gclient runhooks`
and to keep the existing bots (and devs) working while on GYP,
and, perhaps more importantly, adding a new GYP_DEFINE that
does what we need to build/common.gypi is proving intractable,
so we hack MB to rewrite what the GYP_DEFINE would be if life
made sense back into the env var.

BUG=481692, 582737
R=thakis@chromium.org, brettw@chromium.org
==========

[brettw, 2016-04-22 23:08:08]

LGTM from my end

[Nico, 2016-04-24 01:52:48]

lgtm with comment

https://codereview.chromium.org/1908293002/diff/80001/build/config/compiler/B...
File build/config/compiler/BUILD.gn (right):

https://codereview.chromium.org/1908293002/diff/80001/build/config/compiler/B...
build/config/compiler/BUILD.gn:246: defines += [
"CR_CLANG_REVISION=$clang_revision" ]
restore this to what it was (except with the added flag, optionally) and

https://codereview.chromium.org/1908293002/diff/80001/build/config/compiler/B...
build/config/compiler/BUILD.gn:991: if (clang_revision != "266460-1") {
make this say

  if (llvm_force_head_revision)

since that's what it effectively checks

[Dirk Pranke, 2016-04-24 20:49:04]

The CQ bit was checked by dpranke@chromium.org

[Dirk Pranke, 2016-04-24 20:49:04]

The patchset sent to the CQ was uploaded after l-g-t-m from krasin@google.com,
thakis@chromium.org, brettw@chromium.org
Link to the patchset: https://codereview.chromium.org/1908293002/#ps100001
(title: "attempt to update w/ review feedback")

[commit-bot: I haz the power, 2016-04-24 20:49:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1908293002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1908293002/100001

[Dirk Pranke, 2016-04-24 20:49:18]

https://codereview.chromium.org/1908293002/diff/80001/build/config/compiler/B...
File build/config/compiler/BUILD.gn (right):

https://codereview.chromium.org/1908293002/diff/80001/build/config/compiler/B...
build/config/compiler/BUILD.gn:246: defines += [
"CR_CLANG_REVISION=$clang_revision" ]
On 2016/04/24 01:52:48, Nico wrote:
> restore this to what it was (except with the added flag, optionally) and

I'm not sure I understand what you're asking here, but perhaps you want the
comment either copied or moved from #104-107 to here? I'll go ahead and do that.

https://codereview.chromium.org/1908293002/diff/80001/build/config/compiler/B...
build/config/compiler/BUILD.gn:991: if (clang_revision != "266460-1") {
On 2016/04/24 01:52:48, Nico wrote:
> make this say
> 
>   if (llvm_force_head_revision)
> 
> since that's what it effectively checks

Okay, it wasn't clear to me if the intent was "ToT" or "not this revision".

[commit-bot: I haz the power, 2016-04-24 20:55:59]

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

[commit-bot: I haz the power, 2016-04-24 20:56:00]

Try jobs failed on following builders:
  cast_shell_linux on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[Dirk Pranke, 2016-04-24 21:18:57]

The CQ bit was checked by dpranke@chromium.org

[Dirk Pranke, 2016-04-24 21:18:57]

The patchset sent to the CQ was uploaded after l-g-t-m from thakis@chromium.org,
brettw@chromium.org, krasin@google.com
Link to the patchset: https://codereview.chromium.org/1908293002/#ps120001
(title: "fix var name")

[commit-bot: I haz the power, 2016-04-24 21:19:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1908293002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1908293002/120001

[commit-bot: I haz the power, 2016-04-24 23:15:05]

Description was changed from

==========
Add MB configs for ClangToTLinux* bots on FYI

The tip-of-tree Clang bots rely on an environment variable
called LLVM_FORCE_HEAD_REVISION to be set in order to know
to fetch the tip-of-tree Clang and use its version in the compile
time defines, rather than the checked-in version.

We don't want GN to use or depend on environment variables,
so this patch modifies the clang update script to also look at
a new command line arg, and adds a GN arg called 'llvm_force_head_revision' to
cause the same effect.

However, we need to keep the env var around for `gclient runhooks`
and to keep the existing bots (and devs) working while on GYP,
and, perhaps more importantly, adding a new GYP_DEFINE that
does what we need to build/common.gypi is proving intractable,
so we hack MB to rewrite what the GYP_DEFINE would be if life
made sense back into the env var.

BUG=481692, 582737
R=thakis@chromium.org, brettw@chromium.org
==========

to

==========
Add MB configs for ClangToTLinux* bots on FYI

The tip-of-tree Clang bots rely on an environment variable
called LLVM_FORCE_HEAD_REVISION to be set in order to know
to fetch the tip-of-tree Clang and use its version in the compile
time defines, rather than the checked-in version.

We don't want GN to use or depend on environment variables,
so this patch modifies the clang update script to also look at
a new command line arg, and adds a GN arg called 'llvm_force_head_revision' to
cause the same effect.

However, we need to keep the env var around for `gclient runhooks`
and to keep the existing bots (and devs) working while on GYP,
and, perhaps more importantly, adding a new GYP_DEFINE that
does what we need to build/common.gypi is proving intractable,
so we hack MB to rewrite what the GYP_DEFINE would be if life
made sense back into the env var.

BUG=481692, 582737
R=thakis@chromium.org, brettw@chromium.org
==========

[commit-bot: I haz the power, 2016-04-24 23:15:07]

Committed patchset #6 (id:120001)

[commit-bot: I haz the power, 2016-04-24 23:16:43]

Description was changed from

==========
Add MB configs for ClangToTLinux* bots on FYI

The tip-of-tree Clang bots rely on an environment variable
called LLVM_FORCE_HEAD_REVISION to be set in order to know
to fetch the tip-of-tree Clang and use its version in the compile
time defines, rather than the checked-in version.

We don't want GN to use or depend on environment variables,
so this patch modifies the clang update script to also look at
a new command line arg, and adds a GN arg called 'llvm_force_head_revision' to
cause the same effect.

However, we need to keep the env var around for `gclient runhooks`
and to keep the existing bots (and devs) working while on GYP,
and, perhaps more importantly, adding a new GYP_DEFINE that
does what we need to build/common.gypi is proving intractable,
so we hack MB to rewrite what the GYP_DEFINE would be if life
made sense back into the env var.

BUG=481692, 582737
R=thakis@chromium.org, brettw@chromium.org
==========

to

==========
Add MB configs for ClangToTLinux* bots on FYI

The tip-of-tree Clang bots rely on an environment variable
called LLVM_FORCE_HEAD_REVISION to be set in order to know
to fetch the tip-of-tree Clang and use its version in the compile
time defines, rather than the checked-in version.

We don't want GN to use or depend on environment variables,
so this patch modifies the clang update script to also look at
a new command line arg, and adds a GN arg called 'llvm_force_head_revision' to
cause the same effect.

However, we need to keep the env var around for `gclient runhooks`
and to keep the existing bots (and devs) working while on GYP,
and, perhaps more importantly, adding a new GYP_DEFINE that
does what we need to build/common.gypi is proving intractable,
so we hack MB to rewrite what the GYP_DEFINE would be if life
made sense back into the env var.

BUG=481692, 582737
R=thakis@chromium.org, brettw@chromium.org

Committed: https://crrev.com/909bad6afa7124f67c4890f3c1bc244179141978
Cr-Commit-Position: refs/heads/master@{#389408}
==========

[commit-bot: I haz the power, 2016-04-24 23:16:44]

Patchset 6 (id:??) landed as
https://crrev.com/909bad6afa7124f67c4890f3c1bc244179141978
Cr-Commit-Position: refs/heads/master@{#389408}

[Dirk Pranke, 2016-04-25 22:12:16]

fixing now ...

https://codereview.chromium.org/1908293002/diff/120001/tools/mb/mb.py
File tools/mb/mb.py (right):

https://codereview.chromium.org/1908293002/diff/120001/tools/mb/mb.py#newcode...
tools/mb/mb.py:1094: env['LLVM_FORCE_HEAD_REVISION'] = 1
this should've been '1'.