Initial attempt to add support for using gyp to build V8 itself...

Initial attempt to add support for using gyp to build V8 itself

This is based on the structore used in chromium with a script wrapping the call to gyp itself and the default processing of common.gypi.

It is possible to build all our targets on Intel Linux for all architectures (ia32, x64 and ARM simulator). When this is committed I wil take a look at Windows.

See the README.txt file in the changelist for the current way of using it.

Committed: http://code.google.com/p/v8/source/detail?r=6000

[Evan Martin, 2010-12-09 22:03:13]

http://codereview.chromium.org/5701001/diff/1/build/common.gypi
File build/common.gypi (right):

http://codereview.chromium.org/5701001/diff/1/build/common.gypi#newcode56
build/common.gypi:56: },{'target_defaults': {'ccflags':['-xxx']}}],
I'd rewrite this line as
 }, { # else: not linux
   'target_defaults': {'ccflags':['-xxx']}
 }],

http://codereview.chromium.org/5701001/diff/1/tools/gyp/samples.gyp
File tools/gyp/samples.gyp (right):

http://codereview.chromium.org/5701001/diff/1/tools/gyp/samples.gyp#newcode44
tools/gyp/samples.gyp:44: # This could be gotten by not setting chromium_code,
if that's OK.
We normally only set chromium_code for code that is in the chromium tree.  Since
v8 better knows its build settings, I think it can leave chromium_code at 0.

[Søren Thygesen Gjesse, 2010-12-10 14:21:24]

As the CL description says it is possible to build all our targets on Intel
Linux for all architectures (ia32, x64 and ARM simulator).

Evan:
The way I try to determine the host architecture in common.gypi does not work, I
have to do

  $ GYP_DEFINES=target_arch=ia32; build/gyp_v8

or

  $ GYP_DEFINES=target_arch=x64; build/gyp_v8

Also in cctest.gyp I have some stuff which is duplicated in v8.gyp. Will the
best refactoring be to move taht to an include (e.g. arch_defines.gypi)?

Mads:
What do you think?

http://codereview.chromium.org/5701001/diff/1/build/common.gypi
File build/common.gypi (right):

http://codereview.chromium.org/5701001/diff/1/build/common.gypi#newcode56
build/common.gypi:56: },{'target_defaults': {'ccflags':['-xxx']}}],
On 2010/12/09 22:03:13, Evan Martin wrote:
> I'd rewrite this line as
>  }, { # else: not linux
>    'target_defaults': {'ccflags':['-xxx']}
>  }],

This was testing leftovers, removed.

http://codereview.chromium.org/5701001/diff/1/tools/gyp/samples.gyp
File tools/gyp/samples.gyp (right):

http://codereview.chromium.org/5701001/diff/1/tools/gyp/samples.gyp#newcode44
tools/gyp/samples.gyp:44: # This could be gotten by not setting chromium_code,
if that's OK.
On 2010/12/09 22:03:13, Evan Martin wrote:
> We normally only set chromium_code for code that is in the chromium tree. 
Since
> v8 better knows its build settings, I think it can leave chromium_code at 0.

Done.

[Søren Thygesen Gjesse, 2010-12-10 14:22:10]

Adding Mads as reviewer.

[Mads Ager (chromium), 2010-12-13 13:09:42]

LGTM

I would be fine with committing this as a work in progress. We need to make it
easier to use before switching to it for real. Switching completely to gyp
generated build files would be great.

http://codereview.chromium.org/5701001/diff/18001/build/README.txt
File build/README.txt (right):

http://codereview.chromium.org/5701001/diff/18001/build/README.txt#newcode12
build/README.txt:12: To generate Makefiles and build 32-bit version on Linux:
As we just discussed offline I think it would be nice for V8 if we can simply
create a Makefile (VS project, XCode project) that contains targets for
shell-ia32 shell-x64 and shell-sim-arm. This way you don't have to deal with the
GYP_DEFINES.

I'm fine with doing this as a followup change though.

http://codereview.chromium.org/5701001/diff/18001/build/gyp_v8
File build/gyp_v8 (right):

http://codereview.chromium.org/5701001/diff/18001/build/gyp_v8#newcode3
build/gyp_v8:3: # Copyright 2008 the V8 project authors. All rights reserved.
2010

[Evan Martin, 2010-12-13 17:20:58]

I agree with Mads, this LGTM

http://codereview.chromium.org/5701001/diff/18001/build/common.gypi
File build/common.gypi (right):

http://codereview.chromium.org/5701001/diff/18001/build/common.gypi#newcode65
build/common.gypi:65: 'cflags': [ '-Wall', '-pthread', '-fno-rtti',
'-fno-exceptions'],
space before ] for symmetry

http://codereview.chromium.org/5701001/diff/18001/tools/gyp/v8.gyp
File tools/gyp/v8.gyp (right):

http://codereview.chromium.org/5701001/diff/18001/tools/gyp/v8.gyp#newcode650
tools/gyp/v8.gyp:650: '-lpthread',
I always get this wrong, but I have the vague recollection that there a -pthread
flag that passes -lpthread as well as set some other settings...