gyp: unify toolkit_views usage

gyp: unify toolkit_views usage

Now that we can rely on toolkit_views to be set when appropriate,
we can simplify a bunch of logic in gyp.

TEST=compiles

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=45727

[Evan Martin, 2010-04-27 00:57:50]

The trybots are inconclusive on this one.  They all seem to compile, at least...
but some tests fail, maybe flakiness.

I think I'll run it through 'em again after this mail just for good measure.

[Evan Martin, 2010-04-27 01:11:51]

The trybots are inconclusive on this one.  They all seem to compile, at least...
but some tests fail, maybe flakiness.

I think I'll run it through 'em again after this mail just for good measure.

[Daniel Erat, 2010-04-27 01:12:24]

LGTM with one question.

Much cleaner; thanks!

http://codereview.chromium.org/1741010/diff/1/11
File third_party/ffmpeg/ffmpeg.gyp (right):

http://codereview.chromium.org/1741010/diff/1/11#newcode42
third_party/ffmpeg/ffmpeg.gyp:42: ['chromeos==1', {
Am I misreading this, or does it change the logic?  Did this match e.g. Windows
before?

[Evan Martin, 2010-04-27 01:16:20]

http://codereview.chromium.org/1741010/diff/1/11
File third_party/ffmpeg/ffmpeg.gyp (right):

http://codereview.chromium.org/1741010/diff/1/11#newcode42
third_party/ffmpeg/ffmpeg.gyp:42: ['chromeos==1', {
On 2010/04/27 01:12:25, Daniel Erat wrote:
> Am I misreading this, or does it change the logic?  Did this match e.g.
Windows
> before?

You are right.  But I think it was wrong before.

(That is, with the old we'd call ourselves ChromeOS on a toolkit_views
non-ChromeOS build.)

[sky, 2010-04-27 14:50:13]

http://codereview.chromium.org/1741010/diff/1/6
File chrome/chrome_exe.gypi (left):

http://codereview.chromium.org/1741010/diff/1/6#oldcode217
chrome/chrome_exe.gypi:217: ['OS=="linux" and (toolkit_views==1 or
chromeos==1)', {
Where is this dependency expressed now?

http://codereview.chromium.org/1741010/diff/1/9
File chrome/test/interactive_ui/interactive_ui_tests.gypi (left):

http://codereview.chromium.org/1741010/diff/1/9#oldcode70
chrome/test/interactive_ui/interactive_ui_tests.gypi:70:
'<(DEPTH)/build/linux/system.gyp:gtk',
Where are these dependencies expressed now?

[Evan Martin, 2010-04-27 16:17:53]

Running through trybots again, last patch didn't apply for some reason.

http://codereview.chromium.org/1741010/diff/1/6
File chrome/chrome_exe.gypi (left):

http://codereview.chromium.org/1741010/diff/1/6#oldcode217
chrome/chrome_exe.gypi:217: ['OS=="linux" and (toolkit_views==1 or
chromeos==1)', {
On 2010/04/27 14:50:13, sky wrote:
> Where is this dependency expressed now?

Since the dependency isn't expressed in the win branch, I figured it wasn't
needed.  (The trybots will confirm it.)  I assume all of the views-using code is
in the "browser" project etc. which has the dep already.

http://codereview.chromium.org/1741010/diff/1/9
File chrome/test/interactive_ui/interactive_ui_tests.gypi (left):

http://codereview.chromium.org/1741010/diff/1/9#oldcode70
chrome/test/interactive_ui/interactive_ui_tests.gypi:70:
'<(DEPTH)/build/linux/system.gyp:gtk',
On 2010/04/27 14:50:13, sky wrote:
> Where are these dependencies expressed now?

system/xdisplaycheck: now on L53
views.gyp: now on L100

The idea is that shared linux stuff should be in a linux branch, and shared
views stuff in a views branch, rather than duplicating the dependency lists. 
(Ideally these branches, that test both linux+views, will eventually go away;
they're currently here just for files that haven't yet been ported.)

[sky, 2010-04-27 16:53:28]

Ok, as long as it all passes the try bots, LGTM.