arm: fix native compiles

arm: fix native compiles

Landed through http://codereview.chromium.org/546030

[piman, 2010-01-12 02:36:40]

This is to fix http://code.google.com/p/chromium/issues/detail?id=31274

It needs http://codereview.chromium.org/547019 first.

[bradn, 2010-01-12 04:31:41]

I'm not clear on why you wouldn't always want -m32 (as it would have done
before), that's all this appears to do?

How does this relate to the r7 thing?

rubberstamp LG if it makes it go away though.

[bradn, 2010-01-12 04:34:41]

nm, wrong bug in mind...
LGTM

[Søren Thygesen Gjesse, 2010-01-12 08:20:41]

LGTM

http://codereview.chromium.org/549018/diff/1/2
File tools/gyp/v8.gyp (right):

http://codereview.chromium.org/549018/diff/1/2#newcode200
tools/gyp/v8.gyp:200: # The ARM assembler assumes the host is 32 bits, so force
building
Shouldn't this "ARM assembler" be "ARM simulator"?

http://codereview.chromium.org/549018/diff/1/2#newcode427
tools/gyp/v8.gyp:427: # The ARM assembler assumes the host is 32 bits, so force
building
Ditto.

http://codereview.chromium.org/549018/diff/1/2#newcode595
tools/gyp/v8.gyp:595: # The ARM assembler assumes the host is 32 bits, so force
building
Ditto.

[piman, 2010-01-12 08:42:17]

http://codereview.chromium.org/549018/diff/1/2
File tools/gyp/v8.gyp (right):

http://codereview.chromium.org/549018/diff/1/2#newcode200
tools/gyp/v8.gyp:200: # The ARM assembler assumes the host is 32 bits, so force
building
On 2010/01/12 08:20:41, Søren Gjesse wrote:
> Shouldn't this "ARM assembler" be "ARM simulator"?

assembler-arm-inl.h is the file that fails to compile if -m32 isn't on.

[Søren Thygesen Gjesse, 2010-01-12 11:11:13]

On 2010/01/12 08:42:17, piman wrote:
> http://codereview.chromium.org/549018/diff/1/2
> File tools/gyp/v8.gyp (right):
> 
> http://codereview.chromium.org/549018/diff/1/2#newcode200
> tools/gyp/v8.gyp:200: # The ARM assembler assumes the host is 32 bits, so
force
> building
> On 2010/01/12 08:20:41, Søren Gjesse wrote:
> > Shouldn't this "ARM assembler" be "ARM simulator"?
> 
> assembler-arm-inl.h is the file that fails to compile if -m32 isn't on.

Thats right, but if we are compiling the ARM assembler for a x64 target it
implies that the ARM simulator will be used to execute the code it will
generate. Anyway keep the comment as is as mentioning the simulator might create
more confusion.

[piman, 2010-01-13 02:19:22]

On 2010/01/12 11:11:13, Søren Gjesse wrote:
> On 2010/01/12 08:42:17, piman wrote:
> > http://codereview.chromium.org/549018/diff/1/2
> > File tools/gyp/v8.gyp (right):
> > 
> > http://codereview.chromium.org/549018/diff/1/2#newcode200
> > tools/gyp/v8.gyp:200: # The ARM assembler assumes the host is 32 bits, so
> force
> > building
> > On 2010/01/12 08:20:41, Søren Gjesse wrote:
> > > Shouldn't this "ARM assembler" be "ARM simulator"?
> > 
> > assembler-arm-inl.h is the file that fails to compile if -m32 isn't on.
> 
> Thats right, but if we are compiling the ARM assembler for a x64 target it
> implies that the ARM simulator will be used to execute the code it will
> generate. Anyway keep the comment as is as mentioning the simulator might
create
> more confusion.

Ok. Do you mind checking this one in as well ? Thanks !

[Søren Thygesen Gjesse, 2010-01-13 07:37:39]

On 2010/01/13 02:19:22, piman wrote:
> On 2010/01/12 11:11:13, Søren Gjesse wrote:
> > On 2010/01/12 08:42:17, piman wrote:
> > > http://codereview.chromium.org/549018/diff/1/2
> > > File tools/gyp/v8.gyp (right):
> > > 
> > > http://codereview.chromium.org/549018/diff/1/2#newcode200
> > > tools/gyp/v8.gyp:200: # The ARM assembler assumes the host is 32 bits, so
> > force
> > > building
> > > On 2010/01/12 08:20:41, Søren Gjesse wrote:
> > > > Shouldn't this "ARM assembler" be "ARM simulator"?
> > > 
> > > assembler-arm-inl.h is the file that fails to compile if -m32 isn't on.
> > 
> > Thats right, but if we are compiling the ARM assembler for a x64 target it
> > implies that the ARM simulator will be used to execute the code it will
> > generate. Anyway keep the comment as is as mentioning the simulator might
> create
> > more confusion.
> 
> Ok. Do you mind checking this one in as well ? Thanks !

Landed through http://codereview.chromium.org/546030.