Allow compiling v8_shell with the 'host' toolset.

Allow compiling v8_shell with the 'host' toolset.

BUG=82437
TEST=Compile browser_tests for Arm against http://codereview.chromium.org/7087014/

Committed: http://code.google.com/p/v8/source/detail?r=8356

[Sheridan Rawlins, 2011-06-20 20:57:23]

I'd like to be able to build v8_shell for the host.  Can you please review this
change to v8.gyp?

Thanks,

-SCR

[Søren Thygesen Gjesse, 2011-06-21 08:19:02]

LGTM

However there are some things you need to consider.

I don't know how you plan to use the shell on the host, but we are planning to
shave down the shell to a bare minimum sample application and going forward use
the developer shell (d8) as the V8 shell.

We have a GYP file for building d8 in src/d8.gyp, so maybe you should consider
that instead.

Also there is a separate GYP for the sample shell in samples/samples.gyp so we
should also consider removing the v8_shell target from this file to avoid
duplication.

http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp
File tools/gyp/v8.gyp (right):

http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp#newcode959
tools/gyp/v8.gyp:959: 'toolsets': ['host'],
Shouldn't we build v8_shell for the target while we are at it?

http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp#newcode993
tools/gyp/v8.gyp:993: 'toolsets': ['host'],
Ditto.

[Sheridan Rawlins, 2011-06-21 14:41:01]

Thanks, yeah I had heard that about preferring d8 over v8_shell.

I haven't been able to build a runnable d8 with gyp as a dependency, though I'll
try again before this commit.

FYI, I intend to use the v8_shell to do --print_json_ast on some javascript to
generate C++ IN_PROC_BROWSER_TEST_F calls for every function for testing WebUI,
which must mock out the C++ handler.

On 2011/06/21 08:19:02, Søren Gjesse wrote:
> LGTM
> 
> However there are some things you need to consider.
> 
> I don't know how you plan to use the shell on the host, but we are planning to
> shave down the shell to a bare minimum sample application and going forward
use
> the developer shell (d8) as the V8 shell.
> 
> We have a GYP file for building d8 in src/d8.gyp, so maybe you should consider
> that instead.
> 
> Also there is a separate GYP for the sample shell in samples/samples.gyp so we
> should also consider removing the v8_shell target from this file to avoid
> duplication.
> 
> http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp
> File tools/gyp/v8.gyp (right):
> 
> http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp#newcode959
> tools/gyp/v8.gyp:959: 'toolsets': ['host'],
> Shouldn't we build v8_shell for the target while we are at it?
> 
> http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp#newcode993
> tools/gyp/v8.gyp:993: 'toolsets': ['host'],
> Ditto.

[Sheridan Rawlins, 2011-06-21 14:41:19]

(sending comment replies)

http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp
File tools/gyp/v8.gyp (right):

http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp#newcode959
tools/gyp/v8.gyp:959: 'toolsets': ['host'],
Building type executable for both causes collision in the makefile.

On 2011/06/21 08:19:02, Søren Gjesse wrote:
> Shouldn't we build v8_shell for the target while we are at it?

http://codereview.chromium.org/7212014/diff/1/tools/gyp/v8.gyp#newcode993
tools/gyp/v8.gyp:993: 'toolsets': ['host'],
Ditto

On 2011/06/21 08:19:02, Søren Gjesse wrote:
> Ditto.