<simon.hausmann <at> nokia.com> writes: > > > > Currently the plugin architecture (npapi details) is defined by the platform (x11, win, mac). I don't think we should aim for different plugin host processes based on the available lighthouse backends.
I just ask for things to compile and properly fail (e.g. disabling plugins). Today: To fix build of WebKit2 for Qt/E I do: diff --git a/Source/WebKit2/config.h b/Source/WebKit2/config.h index 39ec8d9..68d5430 100644 --- a/Source/WebKit2/config.h +++ b/Source/WebKit2/config.h @@ -146,7 +146,13 @@ static const type& name() \ #define PLUGIN_ARCHITECTURE_MAC 1 #elif PLATFORM(WIN) #define PLUGIN_ARCHITECTURE_WIN 1 -#elif (PLATFORM(QT) || (PLATFORM(GTK))) && (OS(UNIX) && !OS(MAC_OS_X)) +#elif PLATFORM(QT) + #if defined(Q_WS_X11) + #define PLUGIN_ARCHITECTURE_X11 1 + #else + #define PLUGIN_ARCHITECTURE_UNSUPPORTED 1 + #endif +#elif PLATFORM(GTK) && (OS(UNIX) && !OS(MAC_OS_X)) #define PLUGIN_ARCHITECTURE_X11 1 #else #define PLUGIN_ARCHITECTURE_UNSUPPORTED 1 Future: Q_WS_ will not be there anymore, OS(UNIX) is a bad indication to see if X11 headers are installed. Do we need to add detection for X11 headers to our buildprocess then? To me it looks like the PLUGIN_ARCHITECTURE check needs to become a runtime check, because only at runtime we know if we are running on Xlib/Xcb or on something like Wayland? How do we detect the used QPA platform at runtime? _______________________________________________ webkit-qt mailing list webkit-qt@lists.webkit.org http://lists.webkit.org/mailman/listinfo.cgi/webkit-qt
