ChangeSet r38116 sets ENABLE_NETSCAPE_PLUGIN_API to 0 for Symbian builds; however the ChangeSet fails to turn on the plugin stubs for Symbian (in qt/TemporaryLinkStubs.cpp), so QtLauncher for Symbian fails to link with missing symbols.
In addition ChangeSet r41528 introduced the ability to set ENABLE_NETSCAPE_PLUGIN_API to 0 for all the platforms (including Mac, Win and Linux), but those platforms that enable NETSCAPE_PLUGIN_API support by default will have the same linking problem as the Symbian build.
A patch will follow.
Created attachment 28740 [details]
Clean ups in addition to fixing the bugs mentioned in the report)
- All WebKit ports should now share Netscape plugin stubs (introduced by ChangeSet r41786)
- The decision about Netscape plugin support for Qt is now in one place only (WebCore.pro) - it use to be in the source code as well, and the two were out-of-sync
Lazlo, I just tried to build with the patch applied (before landing) and I noticed that it doesn't work entirely:
tmp/PluginViewQt.o: In function `WebCore::PluginView::userAgentStatic()':
PluginViewQt.cpp:(.text+0x60): multiple definition of `WebCore::PluginView::userAgentStatic()'
tmp/TemporaryLinkStubs.o:TemporaryLinkStubs.cpp:(.text+0x0): first defined here
tmp/PluginViewQt.o: In function `WebCore::PluginView::getValueStatic(NPNVariable, void*)':
PluginViewQt.cpp:(.text+0x70): multiple definition of `WebCore::PluginView::getValueStatic(NPNVariable, void*)'
tmp/TemporaryLinkStubs.o:TemporaryLinkStubs.cpp:(.text+0x10): first defined here
I think we may still need the old defines around the two remaining functions in TemporaryLinkStubs.cpp. What do you think?
Created attachment 29014 [details]
Apologies, I do not know how could I miss that.
If I'm not mistaken we can just remove those 2 functions from TemporaryLinkStubs and make the code even simpler.
Comment on attachment 28740 [details]
bugzilla doesn't seem smart enough to remove things from the commit queue when obsolete. or at least our commit queue query isn't smart enough.
Simon, Can you please review this or assign it to someone else to review ? Thanks, Laszlo.
Comment on attachment 29014 [details]
Looks sane to me.
Landed in r43493.