Fix double-building of v8 in GN builds when setting just v8_target_cpu.

Fix double-building of v8 in GN builds when setting just v8_target_cpu.

Because of the somewhat strange way default toolchains and custom
toolchains and user-specified arguments work in GN, if you did a v8
build that just set v8_target_cpu, you could end up building two
identical copies of v8 (see the comments in the change for more).

This CL identifies that case and fixes it.

R=machenbach@chromium.org
BUG=629825
Committed: https://crrev.com/3536db45c9409c9aadc4eee6004cf337c0588cdb
Cr-Commit-Position: refs/heads/master@{#37926}

[Michael Achenbach, 2016-07-21 07:13:00]

The CQ bit was checked by machenbach@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-07-21 07:13:10]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-21 07:35:31]

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

[commit-bot: I haz the power, 2016-07-21 07:35:32]

Dry run: This issue passed the CQ dry run.

[Michael Achenbach, 2016-07-21 08:31:25]

lgtm - landing this to be able to add work on top of it now...

[Michael Achenbach, 2016-07-21 08:31:30]

The CQ bit was checked by machenbach@chromium.org

[commit-bot: I haz the power, 2016-07-21 08:31:46]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-21 09:05:47]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-07-21 09:06:46]

Description was changed from

==========
Fix double-building of v8 in GN builds when setting just v8_target_cpu.

Because of the somewhat strange way default toolchains and custom
toolchains and user-specified arguments work in GN, if you did a v8
build that just set v8_target_cpu, you could end up building two
identical copies of v8 (see the comments in the change for more).

This CL identifies that case and fixes it.

R=machenbach@chromium.org
BUG=629825
==========

to

==========
Fix double-building of v8 in GN builds when setting just v8_target_cpu.

Because of the somewhat strange way default toolchains and custom
toolchains and user-specified arguments work in GN, if you did a v8
build that just set v8_target_cpu, you could end up building two
identical copies of v8 (see the comments in the change for more).

This CL identifies that case and fixes it.

R=machenbach@chromium.org
BUG=629825

Committed: https://crrev.com/3536db45c9409c9aadc4eee6004cf337c0588cdb
Cr-Commit-Position: refs/heads/master@{#37926}
==========

[commit-bot: I haz the power, 2016-07-21 09:06:47]

Patchset 1 (id:??) landed as
https://crrev.com/3536db45c9409c9aadc4eee6004cf337c0588cdb
Cr-Commit-Position: refs/heads/master@{#37926}

[Michael Achenbach, 2016-07-21 10:52:39]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/2167113002/ by machenbach@chromium.org.

The reason for reverting is: Breaks:
https://build.chromium.org/p/client.v8.fyi/builders/V8%20Android%20GN%20%28db...

And also the trybot:
https://build.chromium.org/p/tryserver.chromium.android/builders/android_clan....

[Michael Achenbach, 2016-07-21 11:03:41]

Sorry for committing too fast. I manually added android_clang_dbg_recipe in
order to catch problems like that and saw that it failed with and without patch.
The failures are different though, should have looked more closely.

Looks like it uses target toolchain as snapshot toolchain for mkpeephole. Maybe
that would affect mksnapshot too, but peephole ran first, not sure.

[Dirk Pranke, 2016-07-21 17:20:48]

On 2016/07/21 11:03:41, Michael Achenbach (slow) wrote:
> Sorry for committing too fast. I manually added android_clang_dbg_recipe in
> order to catch problems like that and saw that it failed with and without
patch.
> The failures are different though, should have looked more closely.
> 
> Looks like it uses target toolchain as snapshot toolchain for mkpeephole.
Maybe
> that would affect mksnapshot too, but peephole ran first, not sure.

Ah, hmm, looks like this breaks the normal cross-compile case, and it wasn't
obvious to me because I was just testing x86.

I'll have to think about this some more.

[Dirk Pranke, 2016-07-22 01:54:02]

On 2016/07/21 17:20:48, Dirk Pranke wrote:
> On 2016/07/21 11:03:41, Michael Achenbach (slow) wrote:
> > Sorry for committing too fast. I manually added android_clang_dbg_recipe in
> > order to catch problems like that and saw that it failed with and without
> patch.
> > The failures are different though, should have looked more closely.
> > 
> > Looks like it uses target toolchain as snapshot toolchain for mkpeephole.
> Maybe
> > that would affect mksnapshot too, but peephole ran first, not sure.
> 
> Ah, hmm, looks like this breaks the normal cross-compile case, and it wasn't
> obvious to me because I was just testing x86.
> 
> I'll have to think about this some more.

Revised fix posted to https://codereview.chromium.org/2171083003/ .