Changing default Windows compiler to VS2017

Changing default Windows compiler to VS2017

Doing another VS2017 test over the week end now that the PGO issues have
been fixed.

This CL is currently purely for testing purposes and will be reverted by
the end of the week end.

BUG=683729
Review-Url: https://codereview.chromium.org/2877993002
Cr-Commit-Position: refs/heads/master@{#475212}
Committed: https://chromium.googlesource.com/chromium/src/+/833aa96095e1f827ffc05af8bf545a094ffa6e35

[Sébastien Marchand, 2017-05-12 15:58:39]

Description was changed from

==========
Changing default Windows compiler to VS2017

Doing another VS2017 test over the week end now that the PGO issues have
been fixed.

This CL is currently purely for testing purposes and will be reverted by
the end of the week end.

BUG=683729
==========

to

==========
Changing default Windows compiler to VS2017

Doing another VS2017 test over the week end now that the PGO issues have
been fixed.

This CL is currently purely for testing purposes and will be reverted by
the end of the week end.

BUG=683729
==========

[Sébastien Marchand, 2017-05-12 15:58:39]

sebmarchand@chromium.org changed reviewers:
+ brucedawson@chromium.org, thakis@chromium.org

[Sébastien Marchand, 2017-05-12 15:58:52]

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

[commit-bot: I haz the power, 2017-05-12 15:59:36]

Dry run: CQ is trying da patch.

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

[Sébastien Marchand, 2017-05-12 16:45:37]

PTAL, planning to land this around 5PM PDT (8PM EST)

[Nico, 2017-05-12 16:54:25]

lgtm

[brucedawson, 2017-05-12 17:01:58]

lgtm - I'll be monitoring the bots and the canaries that will (hopefully!) be
produced

[commit-bot: I haz the power, 2017-05-12 17:07:58]

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

[commit-bot: I haz the power, 2017-05-12 17:07:59]

Dry run: Try jobs failed on following builders:
  win_chromium_x64_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...)

[brucedawson, 2017-05-12 17:24:54]

FAILED: obj/v8/v8_base/state-values-utils.obj 
ninja -t msvc -e environment.x86 -- E:\b\c\goma_client/gomacc.exe
"E:\b\depot_tools\win_toolchain\vs_files\4e8a360587a3c8ff3fa46aa9271e982bf3e948ec\VC\Tools\MSVC\14.10.25017\bin\HostX64\x86/cl.exe"
/nologo /showIncludes /FC @obj/v8/v8_base/state-values-utils.obj.rsp /c
../../v8/src/compiler/state-values-utils.cc
/Foobj/v8/v8_base/state-values-utils.obj /Fd"obj/v8/v8_base_cc.pdb"
GOMA:cl.exe ../../v8/src/compiler/state-values-utils.cc:*ERROR*:
compiler_proxy:1190ms: reached max number of active fail fallbacks
[20841/42641] CXX obj/v8/v8_base/simplified-operator-reducer.obj

Uggh. Goma flakiness? Or does this mean that goma is not actually correctly
configured for VS 2017 but has been silently struggling on regardless. I haven't
seen this before. Frustrating.

[Sébastien Marchand, 2017-05-12 17:35:18]

It's failing on several
bots: win_chromium_compile_dbg_ng, win_chromium_x64_rel_ng,
win_chromium_rel_ng
 ... So I don't think that it's a flakiness, seems like an issue with Goma.

I'll enable Goma locally to see if I can repro.


On Fri, May 12, 2017 at 1:24 PM, <brucedawson@chromium.org> wrote:

> FAILED: obj/v8/v8_base/state-values-utils.obj
> ninja -t msvc -e environment.x86 -- E:\b\c\goma_client/gomacc.exe
> "E:\b\depot_tools\win_toolchain\vs_files\4e8a360587a3c8ff3fa
> 46aa9271e982bf3e948ec\VC\Tools\MSVC\14.10.25017\bin\HostX64\x86/cl.exe"
> /nologo /showIncludes /FC @obj/v8/v8_base/state-values-utils.obj.rsp /c
> ../../v8/src/compiler/state-values-utils.cc
> /Foobj/v8/v8_base/state-values-utils.obj /Fd"obj/v8/v8_base_cc.pdb"
> GOMA:cl.exe ../../v8/src/compiler/state-values-utils.cc:*ERROR*:
> compiler_proxy:1190ms: reached max number of active fail fallbacks
> [20841/42641] CXX obj/v8/v8_base/simplified-operator-reducer.obj
>
> Uggh. Goma flakiness? Or does this mean that goma is not actually correctly
> configured for VS 2017 but has been silently struggling on regardless. I
> haven't
> seen this before. Frustrating.
>
>
> https://codereview.chromium.org/2877993002/
>



-- 
Sébastien Marchand | Software Developer | sebmarchand@google.com

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[brucedawson, 2017-05-12 17:37:42]

On 2017/05/12 17:35:18, Sébastien Marchand wrote:
> It's failing on several
> bots: win_chromium_compile_dbg_ng, win_chromium_x64_rel_ng,
> win_chromium_rel_ng
>  ... So I don't think that it's a flakiness, seems like an issue with Goma.
> 
> I'll enable Goma locally to see if I can repro.

Check the goma build status page when testing locally to see if goma is working
- sometimes it silently falls back to local compiles and you won't necessarily
realize unless you look at the stats.

[Sébastien Marchand, 2017-05-12 17:50:59]

Yeah, it seems that most task are running locally (local vs goma is at 130
/ 4), I'm not too familiar with Goma, any way to force ninja to always use
Goma?

On Fri, May 12, 2017 at 1:37 PM, <brucedawson@chromium.org> wrote:

> On 2017/05/12 17:35:18, Sébastien Marchand wrote:
> > It's failing on several
> > bots: win_chromium_compile_dbg_ng, win_chromium_x64_rel_ng,
> > win_chromium_rel_ng
> > ... So I don't think that it's a flakiness, seems like an issue with
> Goma.
> >
> > I'll enable Goma locally to see if I can repro.
>
> Check the goma build status page when testing locally to see if goma is
> working
> - sometimes it silently falls back to local compiles and you won't
> necessarily
> realize unless you look at the stats.
>
>
> https://codereview.chromium.org/2877993002/
>



-- 
Sébastien Marchand | Software Developer | sebmarchand@google.com

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[brucedawson, 2017-05-12 18:00:58]

WHen goma is selected it is supposed to do most (all?) compilation with
goma. A fallback could be because it can't find the correct version of the
compiler (but I thought that was set up long ago) or it could mean that
there are goma problems. Adding goma-users@ in case they have some ideas.

On Fri, May 12, 2017 at 10:50 AM, Sébastien Marchand <
sebmarchand@chromium.org> wrote:

> Yeah, it seems that most task are running locally (local vs goma is at 130
> / 4), I'm not too familiar with Goma, any way to force ninja to always use
> Goma?
>
> On Fri, May 12, 2017 at 1:37 PM, <brucedawson@chromium.org> wrote:
>
>> On 2017/05/12 17:35:18, Sébastien Marchand wrote:
>> > It's failing on several
>> > bots: win_chromium_compile_dbg_ng, win_chromium_x64_rel_ng,
>> > win_chromium_rel_ng
>> > ... So I don't think that it's a flakiness, seems like an issue with
>> Goma.
>> >
>> > I'll enable Goma locally to see if I can repro.
>>
>> Check the goma build status page when testing locally to see if goma is
>> working
>> - sometimes it silently falls back to local compiles and you won't
>> necessarily
>> realize unless you look at the stats.
>>
>>
>> https://codereview.chromium.org/2877993002/
>>
>
>
>
> --
> Sébastien Marchand | Software Developer | sebmarchand@google.com
>
>
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[scottmg, 2017-05-12 19:31:49]

My builds locally (clang win goma) just OOM'd my machine due to -j1000
fallbacks, so I think it must just be an outage, rather than a config problem.

[Nico, 2017-05-12 19:32:49]

Isn't goma supposed to fail by default instead of falling back? I thought
we had changed it to be behaved that way early last year :-/

On Fri, May 12, 2017 at 3:31 PM, <scottmg@chromium.org> wrote:

> My builds locally (clang win goma) just OOM'd my machine due to -j1000
> fallbacks, so I think it must just be an outage, rather than a config
> problem.
>
> https://codereview.chromium.org/2877993002/
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[scottmg, 2017-05-12 20:11:28]

On 2017/05/12 19:32:49, Nico wrote:
> Isn't goma supposed to fail by default instead of falling back? I thought
> we had changed it to be behaved that way early last year :-/

Oops, you're right, false alarm. I mistakenly history-completed to the wrong
build dir that didn't have goma enabled.

(The failures on the bots would tend to indicate that there's a threshold for
local fail fallbacks.)

> 
> On Fri, May 12, 2017 at 3:31 PM, <mailto:scottmg@chromium.org> wrote:
> 
> > My builds locally (clang win goma) just OOM'd my machine due to -j1000
> > fallbacks, so I think it must just be an outage, rather than a config
> > problem.
> >
> > https://codereview.chromium.org/2877993002/
> >
> 
> -- 
> You received this message because you are subscribed to the Google Groups
> "Chromium-reviews" group.
> To unsubscribe from this group and stop receiving emails from it, send an
email
> to mailto:chromium-reviews+unsubscribe@chromium.org.

[Sébastien Marchand, 2017-05-12 21:00:23]

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

[commit-bot: I haz the power, 2017-05-12 21:01:16]

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, 2017-05-12 22:13:39]

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

[commit-bot: I haz the power, 2017-05-12 22:13:40]

Dry run: Try jobs failed on following builders:
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[brucedawson, 2017-05-13 00:24:20]

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

[commit-bot: I haz the power, 2017-05-13 00:24:53]

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, 2017-05-13 01:29:39]

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

[commit-bot: I haz the power, 2017-05-13 01:29:40]

Dry run: Try jobs failed on following builders:
  win_chromium_x64_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...)

[brucedawson, 2017-05-13 15:26:33]

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

[commit-bot: I haz the power, 2017-05-13 15:26:50]

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, 2017-05-13 16:16:02]

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

[commit-bot: I haz the power, 2017-05-13 16:16:02]

Dry run: Try jobs failed on following builders:
  win_chromium_compile_dbg_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_comp...)

[brucedawson, 2017-05-16 20:21:28]

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

[commit-bot: I haz the power, 2017-05-16 20:23:37]

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, 2017-05-17 00:10:25]

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

[commit-bot: I haz the power, 2017-05-17 00:10:27]

Dry run: Try jobs failed on following builders:
  win_chromium_x64_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...)

[brucedawson, 2017-05-17 00:16:43]

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

[commit-bot: I haz the power, 2017-05-17 00:17:54]

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, 2017-05-17 07:56:35]

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

[commit-bot: I haz the power, 2017-05-17 07:56:36]

Dry run: This issue passed the CQ dry run.

[Sébastien Marchand, 2017-05-17 14:44:12]

https://codereview.chromium.org/2884613003/ should fix the last PGO failures.

[brucedawson, 2017-05-26 20:05:19]

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

[commit-bot: I haz the power, 2017-05-26 20:05:52]

Dry run: CQ is trying da patch.

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

[brucedawson, 2017-05-26 20:06:18]

On 2017/05/17 14:44:12, Sébastien Marchand wrote:
> https://codereview.chromium.org/2884613003/ should fix the last PGO failures.

I updated the VS 2017 package to use VS 2017 Update 3 Preview. So, when we land
this it will pick up that latest compiler/header-file version. See
crrev.com/2904733004 for details. Try jobs running now...

[commit-bot: I haz the power, 2017-05-26 21:42:18]

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

[commit-bot: I haz the power, 2017-05-26 21:42:19]

Dry run: This issue passed the CQ dry run.

[Nico, 2017-05-27 00:57:46]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2017-05-27 00:58:01]

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, 2017-05-27 01:34:34]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1495846666875050, "parent_rev":
"73de84a84f1f7c8a4aa9e4b7b1816851df82d49b", "commit_rev":
"833aa96095e1f827ffc05af8bf545a094ffa6e35"}

[commit-bot: I haz the power, 2017-05-27 01:34:46]

Description was changed from

==========
Changing default Windows compiler to VS2017

Doing another VS2017 test over the week end now that the PGO issues have
been fixed.

This CL is currently purely for testing purposes and will be reverted by
the end of the week end.

BUG=683729
==========

to

==========
Changing default Windows compiler to VS2017

Doing another VS2017 test over the week end now that the PGO issues have
been fixed.

This CL is currently purely for testing purposes and will be reverted by
the end of the week end.

BUG=683729

Review-Url: https://codereview.chromium.org/2877993002
Cr-Commit-Position: refs/heads/master@{#475212}
Committed:
https://chromium.googlesource.com/chromium/src/+/833aa96095e1f827ffc05af8bf54...
==========

[commit-bot: I haz the power, 2017-05-27 01:34:47]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/chromium/src/+/833aa96095e1f827ffc05af8bf54...

[brucedawson, 2017-05-27 17:03:15]

On 2017/05/27 01:34:47, commit-bot: I haz the power wrote:
> Committed patchset #1 (id:1) as
>
https://chromium.googlesource.com/chromium/src/+/833aa96095e1f827ffc05af8bf54...

A new Chrome canary just popped out (61.0.3113.0 (Official Build) canary
(32-bit) (cohort: 32-Bit)) but it's built from #475195, so still 2015. So,
61.0.3114.0 should be the one.

[brucedawson, 2017-05-28 16:17:44]

On 2017/05/27 17:03:15, brucedawson wrote:
> On 2017/05/27 01:34:47, commit-bot: I haz the power wrote:
> > Committed patchset #1 (id:1) as
> >
>
https://chromium.googlesource.com/chromium/src/+/833aa96095e1f827ffc05af8bf54...
> 
> A new Chrome canary just popped out (61.0.3113.0 (Official Build) canary
> (32-bit) (cohort: 32-Bit)) but it's built from #475195, so still 2015. So,
> 61.0.3114.0 should be the one.

We now have a VS 2017 build of Chrome canary: 61.0.3114.0 (Official Build)
canary (32-bit) (cohort: 32-Bit) - I'll check crash rates and other data on
Tuesday.

[yoichio, 2017-05-29 06:15:16]

I'm not sure this is related but I have not be able to attach debugger with
vs2015 
 on latest debug build.

[Nico, 2017-05-29 13:45:20]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/2909903002/ by thakis@chromium.org.

The reason for reverting is: We have a canary..

[Nico, 2017-05-30 01:05:38]

On Mon, May 29, 2017 at 2:15 AM, <yoichio@chromium.org> wrote:

> I'm not sure this is related but I have not be able to attach debugger with
> vs2015
> on latest debug build.
>

Can you check if it's better again after the revert? If so, can you file a
bug for this?


>
> https://codereview.chromium.org/2877993002/
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[brucedawson, 2017-05-30 03:35:16]

On 2017/05/30 01:05:38, Nico wrote:
> On Mon, May 29, 2017 at 2:15 AM, <mailto:yoichio@chromium.org> wrote:
> 
> > I'm not sure this is related but I have not be able to attach debugger with
> > vs2015
> > on latest debug build.

How did the attaching fail? I'm not aware of any debugger/compiler dependencies
that would prevent attaching, and I just attached to the 32-bit VS 2017 canary
with VS 2015, no problems.

There could be problems with parsing the symbols, especially if you are doing a
local build and are using /debug:fastlink (is_win_fastlink = true). The fastlink
format has changed with VS 2017.

More details would be appreciated.

[yoichio, 2017-05-30 04:34:09]

On 2017/05/30 03:35:16, brucedawson wrote:
> On 2017/05/30 01:05:38, Nico wrote:
> > On Mon, May 29, 2017 at 2:15 AM, <mailto:yoichio@chromium.org> wrote:
> > 
> > > I'm not sure this is related but I have not be able to attach debugger
with
> > > vs2015
> > > on latest debug build.
> 
> How did the attaching fail? I'm not aware of any debugger/compiler
dependencies
> that would prevent attaching, and I just attached to the 32-bit VS 2017 canary
> with VS 2015, no problems.
> 
> There could be problems with parsing the symbols, especially if you are doing
a
> local build and are using /debug:fastlink (is_win_fastlink = true). The
fastlink
> format has changed with VS 2017.
> 
> More details would be appreciated.

My gn args are:
is_component_build = true
enable_nacl = false
is_win_fastlink = true
target_cpu = "x86"
is_debug = true

Yes, I use the frag.
BTW, I started to use VS2017 and it works fine so far. Thanks.

[brucedawson, 2017-05-30 17:43:08]

> My gn args are:
> is_component_build = true
> enable_nacl = false
> is_win_fastlink = true
> target_cpu = "x86"
> is_debug = true

Some debugging problems when using the VS 2015 debugger and is_win_fastlink and
the VS 2017 compiler are known and expected so it sounds like there is no need
for filing a bug.