Turn off SSE2 when building for ARM or MIPS.

Turn off SSE2 when building for ARM or MIPS. 

Disable SSE2 when building the code for ARM or MIPS. 

Original patch by <petarj@mips.com>;
at https://chromiumcodereview.appspot.com/10947012/

BUG=https://code.google.com/p/chromium/issues/detail?id=130022 
TEST=None.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=158392

[tony, 2012-09-24 17:39:22]

I found the problem in https://chromiumcodereview.appspot.com/10947012/ .

The problem was in qcms.gyp, the variables block is at the same nesting level as
the code added to common.gypi.  In qcms.gyp, when it reads the disable_sse2
variable, it was getting the value from the inner variables block in
common.gypi.

To work around this, I moved the variables block in qcms.gyp into the target so
it would get the value after the condition in common.gypi is resolved.

[Lei Zhang, 2012-09-24 17:51:18]

On 2012/09/24 17:39:22, tony wrote:
> I found the problem in https://chromiumcodereview.appspot.com/10947012/ .
> 
> The problem was in qcms.gyp, the variables block is at the same nesting level
as
> the code added to common.gypi.  In qcms.gyp, when it reads the disable_sse2
> variable, it was getting the value from the inner variables block in
> common.gypi.
> 
> To work around this, I moved the variables block in qcms.gyp into the target
so
> it would get the value after the condition in common.gypi is resolved.

Have you tried putting the target_arch conditional in the conditional block
around line 389 instead, and then keep the 'disable_sse2%': '<(disable_sse2)' on
line 622? Would that avoid having to put the variable block one level lower in
qcms.gyp?

[tony, 2012-09-24 18:44:58]

On 2012/09/24 17:51:18, Lei Zhang wrote:
> On 2012/09/24 17:39:22, tony wrote:
> > I found the problem in https://chromiumcodereview.appspot.com/10947012/ .
> > 
> > The problem was in qcms.gyp, the variables block is at the same nesting
level
> as
> > the code added to common.gypi.  In qcms.gyp, when it reads the disable_sse2
> > variable, it was getting the value from the inner variables block in
> > common.gypi.
> > 
> > To work around this, I moved the variables block in qcms.gyp into the target
> so
> > it would get the value after the condition in common.gypi is resolved.
> 
> Have you tried putting the target_arch conditional in the conditional block
> around line 389 instead, and then keep the 'disable_sse2%': '<(disable_sse2)'
on
> line 622? Would that avoid having to put the variable block one level lower in
> qcms.gyp?

I already suggested that to petarj :)
https://chromiumcodereview.appspot.com/10947012/#msg9
Doesn't work because target_arch is defined in the same level.  We could move
target_arch up one level, but that's a big yak to shave.

[Lei Zhang, 2012-09-24 18:46:53]

lgtm

Oh right. :)

[commit-bot: I haz the power, 2012-09-24 19:06:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tony@chromium.org/10972012/1

[commit-bot: I haz the power, 2012-09-24 21:31:25]

Change committed as 158392