Hi Owen, attached you'll find the updated version of my Plug-Socket patch. It would be nice if you could re-think your policy about applying it. Pros: - it almost works at least for the testsocket.c case - including Gtk<Plug|Socket> for all major platforms instead of making it optional gives a platform independent API. Otherwise other ports (i.e. PyGtk) need to special case platform dependencies for no good - the win32 Plugs/Sockets won't get much better or at least get review from interested parties if they are only available on my local hard disk. - If someone finds a real use case for Plugs/Sockets on win32 it is a good starting point - you would get more testing for them, even patches (there is a bug in tests/testsocket.c) : --- from-cvs/gtk+/tests/testsocket.c Fri Jul 20 12:35:34 2001 +++ my-gtk/gtk+/tests/testsocket.c Fri Jul 20 21:29:20 2001 @@ -3,7 +3,9 @@ #include <string.h> #include <stdlib.h> #include <stdio.h> +#ifdef HAVE_UNISTS_H #include <unistd.h> +#endif int n_children = 0; @@ -131,7 +133,7 @@ gtk_box_pack_start (GTK_BOX (vbox), socket->box, TRUE, TRUE, 0); gtk_widget_show (socket->box); - gtk_socket_steal (GTK_SOCKET (socket), xid); + gtk_socket_steal (GTK_SOCKET (socket->socket), xid); } void Cons : - gtkplug.c and gtksocket.c will get some 'if defined (GDK_WINDOWING_ ...' - there is one thing more not fully functional included in Gtk+2.0 - you'll find more ... Regards, Hans
Attachment:
Plug-Socket-2001-07-20-hb.zip
Description: Zip archive
-------- Hans "at" Breuer "dot" Org ----------- Tell me what you need, and I'll tell you how to get along without it. -- Dilbert