[webkit-reviews] review denied: [Bug 14750] [gtk] Implement plugin support in GTK backend : [Attachment 18239] Updated patch to build with autotools also, against r29084

[bugzilla-daemon at webkit.org]

Mark Rowe (bdash) <mrowe at apple.com> has denied Rodney Dawes
<dobey at wayofthemonkey.com>'s request for review:
Bug 14750: [gtk] Implement plugin support in GTK backend
http://bugs.webkit.org/show_bug.cgi?id=14750

Attachment 18239: Updated patch to build with autotools also, against r29084
http://bugs.webkit.org/attachment.cgi?id=18239&action=edit

------- Additional Comments from Mark Rowe (bdash) <mrowe at apple.com>
Sorry that this has been relatively untouched for so long.

There are a few relatively serious issues here:  PluginViewGtk has a large
amount of duplicated code with PluginViewWin, some of which could trivially be
factored out (handlePost, parseRFC822HeaderFields and friends).  It also
contains many chunks of commented-out code, unused variables
(kWebPluginViewWindowClassName, kWebPluginViewProperty) and plugin quirks which
are probably not needed on non-Windows platforms.  I'm not sure that
PluginMessageThrottlerGtk would be needed on non-Windows platforms as I believe
the Windows equivalent exists due to peculiarities with paint events and the
Windows message queue, but I suspect Anders would know more about that.  I
suspect the references to XID might need to be conditional on X11 being
available, both in PluginViewGtk.cpp and PluginViewGtk.h.  PluginViewGtk seems
to contain a bunch of references to "wndProc" still which is a Windows concept
that seems out of place in this Gtk code.

PluginStreamGtk seems to have a large amount of duplicate code with
PluginStreamWin, only the Gtk version looks to be broken in the case of
NP_ASFILE/NP_ASFILEONLY streams.  It should be possible to share the majority
of this code with Windows, and eventually share with the Mac when that code is
refactored.

With a bit more polish we should be able to get this landed soon.