See also bug 106760
Created attachment 183672 [details] Patch
Created attachment 184864 [details] Patch
Created attachment 185497 [details] Patch Rebased against ToT
Comment on attachment 185497 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=185497&action=review > Source/WebCore/platform/graphics/gtk/FullscreenVideoControllerGtk.h:82 > + unsigned long m_exitFullcreenActionActivateSignalHandler; Two naming nits about this class: these really aren't handlers but rather ids, so how about s/Handler/ID/? And secondly, keeping Gtk out of this class' name would avoid having an #if PLATFORM() block for each of GStreamer-based ports in the MediaPlayerPrivate part of this patch, wouldn't it?
Hum but how would that work? I expect each port to have its own FullscreenController implementation deriving from the base class. Having a Gtk subsclass kinda makes sense to me :) I guess I'm misunderstanding your point
Carlos clarified this for me, the create method should be moved to the GStreamer base class, avoiding the ugly #define FullscreenVideoControllerClass in the player so I can directly use the base class there.
Committed r141566: <http://trac.webkit.org/changeset/141566>
Comment on attachment 185497 [details] Patch Clearing r+ flag of landed patch