bug#28189: 26.0.50; Emacs uses deprecated function gtk_window_parse_geometry

[martin rudalics <rudalics@gmx.at>]

 >> Sure.  We can either remove x_gtk_use_window_move and call
 >> gtk_window_move unconditionally
 >>
 >
 > Would there be any downsides of that?

Currently, x_gtk_use_window_move is a backdoor to get the earlier
behavior that was based on calling gtk_window_parse_geometry.  I'never
been able to understand why Jan did the latter.  I suppose it's because
he wanted to position a window at an offset from the bottom right corner
of the screen.  However, for showing a window for the first time, this
never could have worked because gtk was not able to determine the size
of a yet unfinished window.  But maybe I'm all wrong.

In either case, Emacs doesn't use gtk_window_parse_geometry by default
for a couple of months already and I did not receive any complaints so
far.  Hence there should not be any downsides calling gtk_window_move
unconditionally.

 > I've attached the compilation log including all GTK-related error messages.

Thanks.  I'm still puzzled by the fact that they apparently never
deprecated gtk_window_set_geometry_hints and yet removed it in 3.91.2.
Also, gdk_window_set_override_redirect seems to have been removed
without former deprecation.  That's annoying, at least.  And I have no
idea yet whether gdk_window_process_all_updates is essential and, if so,
how to replace it.

martin