Ensure we check the page pointer before using it after we come out of NPP_Han...

Ensure we check the page pointer before using it after we come out of NPP_HandleEvent, as it might have gone away depending on JavaScript that was executed by the plugin.

BUG=9955
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=13858

[jam, 2009-04-16 02:35:40]

BTW I tried very hard to make the html be embedded in the test case, but it was
impossible since if iframe loaded the data from a data: uri, then the origins
were different and it couldn't script the parent.

[ananta, 2009-04-16 16:37:24]

LGTM. Nice test

http://codereview.chromium.org/75026/diff/1017/25
File webkit/glue/plugins/test/plugin_windowless_test.cc (right):

http://codereview.chromium.org/75026/diff/1017/25#newcode78
Line 78: NPError result = browser->evaluate(id(), window_obj,
For consistency should we call SignalTestCompleted?

http://codereview.chromium.org/75026/diff/1017/26
File webkit/glue/plugins/webplugin_delegate_impl.cc (right):

http://codereview.chromium.org/75026/diff/1017/26#newcode223
Line 223: if (handle_event_depth_) {
Would we run into the same issue with windowed plugins?

[jam, 2009-04-16 17:56:13]

http://codereview.chromium.org/75026/diff/1017/25
File webkit/glue/plugins/test/plugin_windowless_test.cc (right):

http://codereview.chromium.org/75026/diff/1017/25#newcode78
Line 78: NPError result = browser->evaluate(id(), window_obj,
On 2009/04/16 16:37:24, ananta wrote:
> For consistency should we call SignalTestCompleted?

I think it would confuse things since this isn't really signalling completion,
just deleting the frame.

http://codereview.chromium.org/75026/diff/1017/26
File webkit/glue/plugins/webplugin_delegate_impl.cc (right):

http://codereview.chromium.org/75026/diff/1017/26#newcode223
Line 223: if (handle_event_depth_) {
On 2009/04/16 16:37:24, ananta wrote:
> Would we run into the same issue with windowed plugins?

no since only windowless plugins would use HandleEvent