Re: Gtk 1.2.5

[Owen Taylor <otaylor redhat com>]

Erik Mouw <J.A.K.Mouw@its.tudelft.nl> writes:

> On Fri, 10 Sep 1999 19:32:01 +0200 (CEST), mikeh@bahnhof.se wrote:
> > Is there someone who have tried to compile Gtk+ 1.2.5 with
> > hard optimizing (CFLAGS="-O6 -mpentiumpro"?
> > 
> > This will NOT work on my system...
> > 
> > (using egcs 2.91.66)
> 
> Hmm, IIRC, using "-O6" is the same as "-O3" in egcs-1.1.2 (= 2.91.66), so
> that's not very useful. It can be a problem with the extreme
> optimizations, but it might as well be a problem with "-mpentiumpro".
> 
> Anyway, it is not a bug in GTK+, but in the compiler. It is possible that
> this version of GTK+ *triggers* the bug in the compiler. Don't use those
> extreme optimizations, or try with a new version of the compiler
> (gcc-2.95.1 is the latest, see http://egcs.cygnus.com/).

Actually, it _is_ (was) a bug in GTK+. I just fixed this
in CVS, once we finally figured out that it was reproducible
with -O6. It just happens to work with default optimization.

Regards,
                                        Owen

Index: gtkwindow.c
===================================================================
RCS file: /cvs/gnome/gtk+/gtk/gtkwindow.c,v
retrieving revision 1.75.2.17
diff -u -r1.75.2.17 gtkwindow.c
--- gtkwindow.c	1999/09/10 14:00:45	1.75.2.17
+++ gtkwindow.c	1999/09/10 18:57:51
@@ -1090,7 +1090,7 @@
 	child_allocation.width = allocation->width - child_allocation.x * 2;
       else
 	child_allocation.width = 1;
-      if (child_allocation.height > child_allocation.y * 2)
+      if (allocation->height > child_allocation.y * 2)
 	child_allocation.height = allocation->height - child_allocation.y * 2;
       else
 	child_allocation.height = 1;