> > Hmm, I fear that will cause more problems than it solves... Probably right. But still not encounter on my tests.
> > "myi->other_xid" is subsequently used in several places without always > checking to see if it is NULL or not, so I worry that this will throw > other issues. I hope not in to much places because Fl_Double_Window::resize(...) already put other_xid to 0 (marker for deletion) I quickly check and see nowhere unchecked used in the code > I think it would be better if we contrived to ensure that > the offscreen (which is used to simulate double-buffering) > was never created with zero size, which is why I suggested > we try setting MyWin to a nonzero value in my earlier post. My code was just a sample shorter than the tile one to reproduce the same issue present on tile (when resizing to 0 a "tile" with a double sub window) > Creating 0 sized windows is a Bad Thing, in general. > does the tile demo do that? It looks like ... same symptoms at least. > I'm not at all familiar with the way in which the Fl_Tile logic > works... Neither I with fltk. > I assume, from the codepath, that your X-server is not reporting > that it can support XDBE, and that is why you are seeing this > problem. I guess that these days, most testing happens > on servers that do claim to support XDBE so we go down > the alternate codepath and never hit this issue? May be. On thing I see is the code under WIN32 don't seem to be reach at all because draw() is not called for a 0 size window. _______________________________________________ fltk mailing list [email protected] http://lists.easysw.com/mailman/listinfo/fltk
