Enable precompiled headers for Blink on Windows.

Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the problem
persists with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/60880bc3884f2a65d27b8204b86ca5e863c35d1d
Cr-Commit-Position: refs/heads/master@{#433832}

[sof, 2016-11-21 09:24:42]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:    Debug       93:06        8935714
opera-pch[3]:    Release    108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:    Debug       93:06        8935714
opera-pch[3]:    Release    108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[sof, 2016-11-21 09:24:48]

The CQ bit was checked by sigbjornf@opera.com to run a CQ dry run

[commit-bot: I haz the power, 2016-11-21 09:25:03]

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

[sof, 2016-11-21 09:27:34]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:    Debug       93:06        8935714
opera-pch[3]:    Release    108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[sof, 2016-11-21 09:31:01]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[sof, 2016-11-21 09:33:12]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[sof, 2016-11-21 10:05:35]

sigbjornf@opera.com changed reviewers:
+ thakis@chromium.org

[sof, 2016-11-21 10:05:36]

please take a look.

[commit-bot: I haz the power, 2016-11-21 11:49:22]

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

[commit-bot: I haz the power, 2016-11-21 11:49:24]

Dry run: This issue passed the CQ dry run.

[Nico, 2016-11-21 15:42:25]

Thanks for picking this up, seems like a good week to try this again. The CL
description should call out that there's no reason to believe that things have
improved other than that we've updated to a newer MSVC, and that this we're
basically relanding and hoping for the best. It should include something along
the lines of "sheriffs: if you're seeing build flakiness that looks like
$how_it_looked_last_time, it's likely due to this CL; in that case revert".

Is there something in place to check that blink targets depend on at most one of
these _pch targets?

https://codereview.chromium.org/2520863002/diff/1/build/config/pch.gni
File build/config/pch.gni (right):

https://codereview.chromium.org/2520863002/diff/1/build/config/pch.gni#newcode15
build/config/pch.gni:15: }
disable_precompiled_headers = is_official_build || use_goma

in general, we try to not use negated variable names since `if (!disable_foo)`
is harder to read than `if (enable_foo)` and `if (!enable_foo)` is equally hard
as `if (disable_foo)` (double negation and whatnot. I realize you're just moving
this around, but could this be called enable_precompiled_headers?

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/B...
File third_party/WebKit/Source/BUILD.gn (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/B...
third_party/WebKit/Source/BUILD.gn:73: # This error doesn't happen every time.
In VS2013, it seems if the .pch
Since this mentions 2013 -- do we still need this flag with 2015?

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/win/Precompile-core.cpp (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/win/Precompile-core.cpp:6: * Precompiled header
generator for Windows builds. No include is needed
same

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/win/Precompile-core.h (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/win/Precompile-core.h:5: /*
why c-style comment

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
File third_party/WebKit/Source/platform/BUILD.gn (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
third_party/WebKit/Source/platform/BUILD.gn:1904:
"//third_party/WebKit/Source:blink_pch",
why blink_pch instead of blink_platform_pch?

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
File third_party/WebKit/Source/platform/win/Precompile-platform.h (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
third_party/WebKit/Source/platform/win/Precompile-platform.h:13: // Include
Oilpan's Handle.h by default, and all it refers to.
comments should say why, not (just) what

[sof, 2016-11-21 19:12:21]

The CQ bit was checked by sigbjornf@opera.com to run a CQ dry run

[commit-bot: I haz the power, 2016-11-21 19:13:12]

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

[sof, 2016-11-21 19:18:29]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect. This was with GYP and earlier MSVC toolchains; we have no reason to
believe the problem persists with GN and MSVC2015.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[sof, 2016-11-21 19:24:58]

Thanks for the quick review, thakis.

> The CL description should call out that there's no reason to believe that
things
> have improved other than that we've updated to a newer MSVC, ..

Good idea, added to the description.

>Is there something in place to check that blink targets depend on at most one
> of these _pch targets?

Yes, gn's Target::ResolvePrecompiledHeaders() checks for conflicting
(re)definitions.

https://codereview.chromium.org/2520863002/diff/1/build/config/pch.gni
File build/config/pch.gni (right):

https://codereview.chromium.org/2520863002/diff/1/build/config/pch.gni#newcode15
build/config/pch.gni:15: }
On 2016/11/21 15:42:25, Nico wrote:
> disable_precompiled_headers = is_official_build || use_goma
> 
> in general, we try to not use negated variable names since `if (!disable_foo)`
> is harder to read than `if (enable_foo)` and `if (!enable_foo)` is equally
hard
> as `if (disable_foo)` (double negation and whatnot. I realize you're just
moving
> this around, but could this be called enable_precompiled_headers?

ok, switched to enable_precompiled_headers.

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/B...
File third_party/WebKit/Source/BUILD.gn (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/B...
third_party/WebKit/Source/BUILD.gn:73: # This error doesn't happen every time.
In VS2013, it seems if the .pch
On 2016/11/21 15:42:25, Nico wrote:
> Since this mentions 2013 -- do we still need this flag with 2015?

Hard to conclusively say; I would prefer to handle its potential removal
separately from this bigger change. Added a followup TODO for it.

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/win/Precompile-core.h (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/win/Precompile-core.h:5: /*
On 2016/11/21 15:42:25, Nico wrote:
> why c-style comment

thanks, normalized them all.

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
File third_party/WebKit/Source/platform/BUILD.gn (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
third_party/WebKit/Source/platform/BUILD.gn:1904:
"//third_party/WebKit/Source:blink_pch",
On 2016/11/21 15:42:25, Nico wrote:
> why blink_pch instead of blink_platform_pch?

The sources aren't exercising Blink platform APIs much (=> Oilpan), but I don't
think it is merited to use precompiled headers for this small set. Removed.

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
File third_party/WebKit/Source/platform/win/Precompile-platform.h (right):

https://codereview.chromium.org/2520863002/diff/1/third_party/WebKit/Source/p...
third_party/WebKit/Source/platform/win/Precompile-platform.h:13: // Include
Oilpan's Handle.h by default, and all it refers to.
On 2016/11/21 15:42:25, Nico wrote:
> comments should say why, not (just) what

Adjusted.

[Nico, 2016-11-21 23:02:32]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect. This was with GYP and earlier MSVC toolchains; we have no reason to
believe the problem persists with GN and MSVC2015.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
was fixed with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[Nico, 2016-11-21 23:03:41]

lgtm

https://codereview.chromium.org/2520863002/diff/20001/build/config/pch.gni
File build/config/pch.gni (right):

https://codereview.chromium.org/2520863002/diff/20001/build/config/pch.gni#ne...
build/config/pch.gni:14: enable_precompiled_headers = true
enable_precompiled_headers = !is_official_build && !use_goma

instead of the if

[commit-bot: I haz the power, 2016-11-21 23:51:49]

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

[commit-bot: I haz the power, 2016-11-21 23:51:50]

Dry run: This issue passed the CQ dry run.

[sof, 2016-11-22 06:28:06]

The CQ bit was checked by sigbjornf@opera.com to run a CQ dry run

[commit-bot: I haz the power, 2016-11-22 06:28:22]

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

[sof, 2016-11-22 06:34:42]

sigbjornf@opera.com changed reviewers:
+ jochen@chromium.org, tkent@chromium.org

[sof, 2016-11-22 06:34:43]

tkent@, jochen@: could you have a look as platform owners, please?

https://codereview.chromium.org/2520863002/diff/20001/build/config/pch.gni
File build/config/pch.gni (right):

https://codereview.chromium.org/2520863002/diff/20001/build/config/pch.gni#ne...
build/config/pch.gni:14: enable_precompiled_headers = true
On 2016/11/21 23:03:41, Nico wrote:
> enable_precompiled_headers = !is_official_build && !use_goma
> 
> instead of the if

switched to expression form.

[tkent, 2016-11-22 06:53:53]

lgtm

https://codereview.chromium.org/2520863002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/build/win/Precompile.h (right):

https://codereview.chromium.org/2520863002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/win/Precompile.h:10: // Precompiled header for
WebKit when built on Windows using
WebKit -> Blink?

[sof, 2016-11-22 07:05:44]

https://codereview.chromium.org/2520863002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/build/win/Precompile.h (right):

https://codereview.chromium.org/2520863002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/win/Precompile.h:10: // Precompiled header for
WebKit when built on Windows using
On 2016/11/22 06:53:52, tkent wrote:
> WebKit -> Blink?

thanks, updated.

[sof, 2016-11-22 07:05:49]

The CQ bit was checked by sigbjornf@opera.com

[sof, 2016-11-22 07:05:49]

The patchset sent to the CQ was uploaded after l-g-t-m from thakis@chromium.org,
tkent@chromium.org
Link to the patchset: https://codereview.chromium.org/2520863002/#ps60001
(title: "update comment")

[commit-bot: I haz the power, 2016-11-22 07:06:02]

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

[sof, 2016-11-22 07:06:20]

The CQ bit was unchecked by sigbjornf@opera.com

[sof, 2016-11-22 07:13:50]

The CQ bit was checked by sigbjornf@opera.com

[sof, 2016-11-22 07:13:50]

The patchset sent to the CQ was uploaded after l-g-t-m from tkent@chromium.org,
thakis@chromium.org
Link to the patchset: https://codereview.chromium.org/2520863002/#ps80001
(title: "rebased")

[commit-bot: I haz the power, 2016-11-22 07:13:59]

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

[sof, 2016-11-22 09:07:06]

The CQ bit was unchecked by sigbjornf@opera.com

[sof, 2016-11-22 09:07:53]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2016-11-22 09:08:12]

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, 2016-11-22 10:39:28]

CQ is committing da patch.

Bot data: {"patchset_id": 80001, "attempt_start_ts": 1479805673840810,
"parent_rev": "b6dc208b0cbc31d2c99bb55d038e4d52ecc922e6", "commit_rev":
"85fee106abc263f5c427a2f771126470df009b8a"}

[commit-bot: I haz the power, 2016-11-22 10:39:50]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
was fixed with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
was fixed with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[commit-bot: I haz the power, 2016-11-22 10:39:51]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-11-22 10:43:30]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
was fixed with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
was fixed with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/60880bc3884f2a65d27b8204b86ca5e863c35d1d
Cr-Commit-Position: refs/heads/master@{#433832}
==========

[commit-bot: I haz the power, 2016-11-22 10:43:31]

Patchset 5 (id:??) landed as
https://crrev.com/60880bc3884f2a65d27b8204b86ca5e863c35d1d
Cr-Commit-Position: refs/heads/master@{#433832}

[sof, 2016-11-22 13:45:01]

Description was changed from

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
was fixed with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/60880bc3884f2a65d27b8204b86ca5e863c35d1d
Cr-Commit-Position: refs/heads/master@{#433832}
==========

to

==========
Enable precompiled headers for Blink on Windows.

One reason Blink is slow to compile is that there is a lot of code
included in every compilation unit. This is partly because everything
depends on either LayoutObject.h or Document.h and those in turn
include huge portions of the rest of Blink.

By precompiling LayoutObject.h and Document.h, the compilation of
core/ and modules/ in Blink can be considerably reduced;
some numbers:

@ r433149       config      build (mins)  size (Kb)
------------------------------------------------
master:         Debug       149:30        9410487
master:         Release     176:16        6118938

opera-pch[2]:   Debug       134:59        9337121
opera-pch[2]:   Release     160:42        6110812

opera-pch[3]:   Debug        93:06        8935714
opera-pch[3]:   Release     108:34        5029242

This for a clean build of target 'blink_tests', i.e., building
both chromium and blink parts. The gains are all local to Blink,
clearly. Host is an i7-3770 (4 phys cores); 32G + 256 SSD -
Win7 Pro.

The precompiled header file is judiciously (and forcefully) included
while compiling the core/ + web/ (and some of modules/) sources. Except
for some name disambiguation trivia when compiling the XPath grammar,
no source changes are needed to make this work out.

Note that distributed compilation system disables precompiled headers
globally so this will *not* make trybots faster. But many developers
don't have access to such super powers.

This already landed[1] in the gyp/VS2013 world some time ago but
unclear & unexplained bot failures caused a revert. Now with gn and
VS2015 the world should be a better place. This CL actually takes over
where [2] got stuck / ran out of time, extending its scope quite
considerably (i.e., 40 mins faster builds wrt the above pch numbers.)

[1] https://codereview.chromium.org/1167523007/
[2] https://codereview.chromium.org/2152783002/
[3] this CL.

Note sheriffs: should unexplained Windows build errors surface on the bots,
similar to the ones seen in crbug.com/511945, then please consider this CL a
suspect.
This was with GYP and earlier MSVC toolchains; we have no reason to believe the
problem
persists with GN and MSVC2015, we're just hoping for the best.

R=
BUG=495697
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/60880bc3884f2a65d27b8204b86ca5e863c35d1d
Cr-Commit-Position: refs/heads/master@{#433832}
==========