gyp changes - add support for building Native Client for 64-bit Windows...

Split more targets from chrome.gyp to add support for building Native Client for 64-bit Windows. 

This CL is required for supporting Native Client on 64-bit Windows. On Win64 we will use a 64-bit executable to load NaCl modules. Since we have only one version of Chrome on Windows, this executable has to be built together with 32-bit Chrome binaries.
For the targets that are required for both 32-bit and 64-bit builds, this CL moves the shared parts of the target definitions into target-specific gypi files (such as base.gypi). New 64-bit only targets are added. Since we would like to port as few code as possible at this point, some 64-bit targets include less source files or less dependencies than the original 32-bit targets. These targets will have nacl_win64 in their name - to emphasize that they should be used for NaCl only. Some 64-bit targets are identical to the 32-bit targets in everything except for the target platform - these will have only win64 in their name. 

TEST=none
BUG=28176
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=35237

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=35305

[bradn, 2009-12-18 18:58:42]

Sigh....
This certainly isn't pretty, but at least it only affects things like chrome_exe
/ chrome_dll that hopefully are fairly isolated.

We've got to sort out the configuration inheritance hierarchy before this lands,
as its gonna do unpredictable stuff on the bots, and will add in extra configs.

http://codereview.chromium.org/384068/diff/42001/42006
File chrome/chrome_common.gypi (right):

http://codereview.chromium.org/384068/diff/42001/42006#newcode136
chrome/chrome_common.gypi:136: # All .cc, .h, and .mm files under chrome/common
except for tests.
This kind of contradicts what's above.
Maybe label these so its clear to someone who's adding a source file, where is
should go.

http://codereview.chromium.org/384068/diff/42001/42006#newcode364
chrome/chrome_common.gypi:364: 'Common': {
All this is potentially gonna clash with ian's change. We need to get that
sorted out first.

http://codereview.chromium.org/384068/diff/42001/42004
File chrome/chrome_exe.gypi (right):

http://codereview.chromium.org/384068/diff/42001/42004#newcode533
chrome/chrome_exe.gypi:533: 'Common': {
Same story here.

http://codereview.chromium.org/384068/diff/42001/42005
File chrome/common_constants.gypi (right):

http://codereview.chromium.org/384068/diff/42001/42005#newcode88
chrome/common_constants.gypi:88: 'Common': {
And here too....

http://codereview.chromium.org/384068/diff/42001/42003
File chrome/nacl.gypi (right):

http://codereview.chromium.org/384068/diff/42001/42003#newcode92
chrome/nacl.gypi:92: 'Common': {
:-)

[bradn, 2009-12-18 19:20:35]

Ok from offline discussion.
LGTM, assuming you fix up the comments.
The configuration inheritance changes can wait for a separate change, ASSUMING
that NO ONE rolls the nacl DEPS version until this has been fixed.

[gregoryd, 2009-12-18 19:46:10]

http://codereview.chromium.org/384068/diff/42001/42006
File chrome/chrome_common.gypi (right):

http://codereview.chromium.org/384068/diff/42001/42006#newcode136
chrome/chrome_common.gypi:136: # All .cc, .h, and .mm files under chrome/common
except for tests.
On 2009/12/18 18:58:42, bradn wrote:
> This kind of contradicts what's above.
> Maybe label these so its clear to someone who's adding a source file, where is
> should go.

Done.

[bradn, 2009-12-18 19:47:22]

LGTM, good luck :-)