Override mechanism for features.gypi

Override mechanism for features.gypi
This will allow us to have a different set of feature_defines for upstream and downstream webkit builds.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=27789

[yaar, 2009-10-01 21:13:52]

This is the chromium side counterpart to
https://bugs.webkit.org/show_bug.cgi?id=29979

[Mark Mentovai, 2009-10-01 21:15:25]

LG otherwise.

http://codereview.chromium.org/242102/diff/6/1001
File DEPS (right):

http://codereview.chromium.org/242102/diff/6/1001#newcode192
Line 192: "-Isrc/build/features_override.gypi"],
Don't do this.  Modify gyp_chromium.

[yaar, 2009-10-01 21:29:37]

But gyp_chromium is also called upstream.

[TVL, 2009-10-01 21:34:25]

driveby - can you use a supplement.gypi?  those get auto collected into the run,
right?

[Mark Mentovai, 2009-10-01 21:43:55]

yaar wrote:
> But gyp_chromium is also called upstream.

I don't think that's correct, then.  gyp_chromium is supposed to be
Chromium-specific.

[yaar, 2009-10-01 22:03:21]

Forking gyp_chromium is a bit too complex for the scope (and urgency) of this
patch. I'd rather do that as a separate (2-sides) patch. Is that ok?

[yaar, 2009-10-01 22:20:52]

Webkit bug for next step (fork gyp_chromium):
https://bugs.webkit.org/show_bug.cgi?id=29986

[tony, 2009-10-05 17:17:44]

It doesn't seem like you got an actual LG for this change and it caused a lot of
people's builds to break this morning (when johnnyg flipped
ENABLE_NOTIFICATIONS).

Why was this submitted without a LG?  What was so urgent?

I see the fix has been committed upstream, so hopefully this will be pulled
downstream soon, but you're wasting lots of other people's time.

[yaar, 2009-10-05 17:31:15]

Sorry for causing the problems.

This review has an LG-except, and the except was resolved, so I committed.
I'm fixing the downstream gyp_chromium issue right now. I forgot that a few
people use gyp_chromium directly and not only via the DEPS hook.

On Mon, Oct 5, 2009 at 10:17 AM, <tony@chromium.org> wrote:

> It doesn't seem like you got an actual LG for this change and it caused a
> lot of
> people's builds to break this morning (when johnnyg flipped
> ENABLE_NOTIFICATIONS).
>
> Why was this submitted without a LG?  What was so urgent?
>
> I see the fix has been committed upstream, so hopefully this will be pulled
> downstream soon, but you're wasting lots of other people's time.
>
>
> http://codereview.chromium.org/242102
>