Bug 40730 – allow ports with windowed plugins to support windowless plugin tests

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[Robert Hogan]

Ports that implement plugins windowed by default need to have a way of supporting plugin tests that assume the plugin is windowless. add this feature to the tests and support it in the webkit test plugin. Also add mouse and keyboard event logging to the webkit-test plugin on Unix.

[Robert Hogan]

Created attachment 59027 [details]
Patch

[WebKit Review Bot]

Attachment 59027 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style', '--no-squash']" exit_code: 1
WebKitTools/DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:244:  A case label should not be indented, but line up with its switch statement.  [whitespace/indent] [4]
Total errors found: 1 in 9 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Robert Hogan]

(In reply to comment #2)
> Attachment 59027 [details] did not pass style-queue:
> 
> Failed to run "['WebKitTools/Scripts/check-webkit-style', '--no-squash']" exit_code: 1
> WebKitTools/DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:244:  A case label should not be indented, but line up with its switch statement.  [whitespace/indent] [4]
> Total errors found: 1 in 9 files
> 
> 
> If any of these errors are false positives, please file a bug against check-webkit-style.

This is the style of the other switch statements in the file. Opened 40796 to cover this.

[Kenneth Rohde Christiansen]

Looks good to me. Anders what do you think?

[Robert Hogan]

Alexey, would you be able to look at this?

[Robert Hogan]

This is rotting on the vine. Anyone care to take a look?

[Simon Hausmann]

Comment on attachment 59027 [details]
Patch

WebCore/plugins/qt/PluginViewQt.cpp:297
 +      if (!xEvent->xkey.keycode) {
Perhaps this should be guarded with a drt run-time guard?

WebCore/plugins/qt/PluginViewQt.cpp:@
 +  void PluginView::setFocus(bool focused)
The changes in this function worry me a bit. Should sending the X focus event perhaps be guarded with !m_isWindowed or simply be moved into the else() block, in order to avoid that the event receives the X focus event twice, once through X itself (xembed) and once through the direct dispatchNPEvent() call.

The rest of the patch looks excellent to me! I'm sorry that it's been rotting away :(

I'm going to say r+ with the condition that the X focus event code should be moved so that it's not sent twice in case of windowed plugins.

[Robert Hogan]

Committed r66408: <http://trac.webkit.org/changeset/66408>