Lubos Lunak <l lunak suse cz> writes: > On Tuesday 11 April 2006 02:29, Matthias Clasen wrote: > > On Tue, 2006-04-11 at 09:02 +0900, Carsten Haitzler wrote: > > > so my point - it's not a pi break at all adding extra enumerations. the > > > code should have already handled the case of an enumeration it doesn't > > > know about to start with in some "sane" fashion. > > > > This is not about trusting values from properties (which of course, you > > must not do), it is about GDK apis which have a C enum for the possible > > values of the window type. Adding new values does indeed force us to > > break the api by adding new values to the enumeration. I do agree that > > the api break is minor and that we should probably do it, but it is an > > api break, unless the enumeration is explicitly documented as "may grow > > new values", which is isn't. > > Then you have a serious problem I'm afraid - _NET_WM_WINDOW_TYPE is designed > to grow and it has already done so a couple of times. Here is a new patch against wm-spec.xml that goes back to extending _NET_WM_WINDOW_TYPE instead of adding a new property. I think we can probably deal with the GTK+ problem. If there are no further comments, I'll commit soonish. Soren
Attachment:
specdiff2
Description: The patch