Make delayed-permission-denied-for-multiple-requests pass

Make delayed-permission-denied-for-multiple-requests pass
Enqueue permission requests until the first setPermission call is received.

BUG=webkit 43480
TEST=fast/dom/Geolocation/delayed-permission-denied-for-multiple-requests.html

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=58631

[joth, 2010-09-03 14:24:56]

As a side effect of my other fixes, I worked this one out.

[steveblock, 2010-09-03 14:36:15]

LGTM

[jorlow, 2010-09-06 10:52:48]

LGTM

[bulach, 2010-09-06 11:51:47]

LGTM

thanks joth!
minor question below, we probably want to do something similar for the
DumpRenderTree bits, entry point:
third_party/WebKit/WebKitTools/DumpRenderTree/chromium/LayoutTestController.cpp

http://codereview.chromium.org/3294007/diff/2001/3002
File webkit/tools/test_shell/test_geolocation_service.cc (right):

http://codereview.chromium.org/3294007/diff/2001/3002#newcode60
webkit/tools/test_shell/test_geolocation_service.cc:60: DCHECK(bridge);
not entirely sure why you're copying the vector, is it because the bridge can be
detached while traversing it? perhaps then remove this DCHECK with an if?

[joth, 2010-09-06 12:08:57]

Just to clarify, let me know if you still had a query about this.

http://codereview.chromium.org/3294007/diff/2001/3002
File webkit/tools/test_shell/test_geolocation_service.cc (right):

http://codereview.chromium.org/3294007/diff/2001/3002#newcode60
webkit/tools/test_shell/test_geolocation_service.cc:60: DCHECK(bridge);
On 2010/09/06 11:51:48, bulach wrote:
> not entirely sure why you're copying the vector, is it because the bridge can
be
> detached while traversing it? perhaps then remove this DCHECK with an if?

I'm not copying the vector, I'm swapping it (important difference)
 I just find it cleaner style and easier to reason about this way.