GN: Do not pass /safeseh to ml64.exe

GN: Do not pass /safeseh to ml64.exe

The 64-bit MSVC assembler (ml64.exe) does not accept the /safeseh
switch.  Only the 32-bit x86 MSVC assembler (ml.exe) does.  The
64-bit assembler prints a warning about an unrecognized option,
which is ulgy though it does not break the build.

BUG= none
R=scottmg@chromium.org

Committed: https://crrev.com/a176b5bc5cca245c32f8963579fd6cdca4e6ea25
Cr-Commit-Position: refs/heads/master@{#362539}

[scottmg, 2015-12-01 21:44:44]

https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn
File build/config/win/BUILD.gn (right):

https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn#n...
build/config/win/BUILD.gn:15: if (current_cpu == "x86") {
Are target_cpu and current_cpu the same here?

[Roland McGrath, 2015-12-01 21:49:23]

https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn
File build/config/win/BUILD.gn (right):

https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn#n...
build/config/win/BUILD.gn:15: if (current_cpu == "x86") {
On 2015/12/01 21:44:44, scottmg wrote:
> Are target_cpu and current_cpu the same here?

Not necessarily.  current_cpu is the proper one to test; it says what cpu this
config applies to.  target_cpu is what the overall build's so-called target is,
but that will not vary in different toolchain contexts as current_cpu does.  See
below in the same block where current_cpu is tested to choose -m32/-m64 options
for Clang.

[scottmg, 2015-12-01 21:49:31]

lgtm with that.

https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn
File build/config/win/BUILD.gn (right):

https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn#n...
build/config/win/BUILD.gn:15: if (current_cpu == "x86") {
On 2015/12/01 21:44:44, scottmg wrote:
> Are target_cpu and current_cpu the same here?

From `gn help target_cpu` and `gn help current_cpu`, I think target_cpu makes
more sense.

[scottmg, 2015-12-01 21:50:22]

On 2015/12/01 21:49:23, Roland McGrath wrote:
> https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn
> File build/config/win/BUILD.gn (right):
> 
>
https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn#n...
> build/config/win/BUILD.gn:15: if (current_cpu == "x86") {
> On 2015/12/01 21:44:44, scottmg wrote:
> > Are target_cpu and current_cpu the same here?
> 
> Not necessarily.  current_cpu is the proper one to test; it says what cpu this
> config applies to.  target_cpu is what the overall build's so-called target
is,
> but that will not vary in different toolchain contexts as current_cpu does. 
See
> below in the same block where current_cpu is tested to choose -m32/-m64
options
> for Clang.

Oh! I see. lgtm as is then.

[Roland McGrath, 2015-12-01 21:51:00]

On 2015/12/01 21:49:31, scottmg wrote:
> lgtm with that.
> 
> https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn
> File build/config/win/BUILD.gn (right):
> 
>
https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn#n...
> build/config/win/BUILD.gn:15: if (current_cpu == "x86") {
> On 2015/12/01 21:44:44, scottmg wrote:
> > Are target_cpu and current_cpu the same here?
> 
> From `gn help target_cpu` and `gn help current_cpu`, I think target_cpu makes
> more sense.

I'm quite sure you're mistaken here.  Please explain your rationale.  Evidently
the 'gn help' text needs to change so as not to mislead you in this way.

[Roland McGrath, 2015-12-01 21:51:25]

The CQ bit was checked by mcgrathr@chromium.org

[commit-bot: I haz the power, 2015-12-01 21:52:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1488853003/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1488853003/1

[scottmg, 2015-12-01 21:57:26]

On 2015/12/01 21:51:00, Roland McGrath wrote:
> On 2015/12/01 21:49:31, scottmg wrote:
> > lgtm with that.
> > 
> > https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn
> > File build/config/win/BUILD.gn (right):
> > 
> >
>
https://codereview.chromium.org/1488853003/diff/1/build/config/win/BUILD.gn#n...
> > build/config/win/BUILD.gn:15: if (current_cpu == "x86") {
> > On 2015/12/01 21:44:44, scottmg wrote:
> > > Are target_cpu and current_cpu the same here?
> > 
> > From `gn help target_cpu` and `gn help current_cpu`, I think target_cpu
makes
> > more sense.
> 
> I'm quite sure you're mistaken here.  Please explain your rationale. 
Evidently
> the 'gn help' text needs to change so as not to mislead you in this way.

Yeah, I only read the headings.

"target_cpu: The desired cpu architecture for the build." sounds like the target
and "current_cpu: The processor architecture of the current toolchain." sounds
just like the architecture for which the tools were compiled.

But I think they're fine on a more careful reading.

[commit-bot: I haz the power, 2015-12-01 22:54:03]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-12-01 23:02:35]

Description was changed from

==========
GN: Do not pass /safeseh to ml64.exe

The 64-bit MSVC assembler (ml64.exe) does not accept the /safeseh
switch.  Only the 32-bit x86 MSVC assembler (ml.exe) does.  The
64-bit assembler prints a warning about an unrecognized option,
which is ulgy though it does not break the build.

BUG= none
R=scottmg@chromium.org
==========

to

==========
GN: Do not pass /safeseh to ml64.exe

The 64-bit MSVC assembler (ml64.exe) does not accept the /safeseh
switch.  Only the 32-bit x86 MSVC assembler (ml.exe) does.  The
64-bit assembler prints a warning about an unrecognized option,
which is ulgy though it does not break the build.

BUG= none
R=scottmg@chromium.org

Committed: https://crrev.com/a176b5bc5cca245c32f8963579fd6cdca4e6ea25
Cr-Commit-Position: refs/heads/master@{#362539}
==========

[commit-bot: I haz the power, 2015-12-01 23:02:35]

Patchset 1 (id:??) landed as
https://crrev.com/a176b5bc5cca245c32f8963579fd6cdca4e6ea25
Cr-Commit-Position: refs/heads/master@{#362539}