Re: GCC 4.1 problem?

[Enrico Sersale]

On 2006-05-28 17:14:25 +0300 Fred Kiefer <[EMAIL PROTECTED]> wrote:ver happen? 
I put in a few NSLog statements into

[NSAffineTransform scaleXBy:YBy:] and the problem was gone.

For me this proves that there is a gcc problem (gcc (GCC) 4.1.0 (SUSE
Linux)), but how to deal with it? The simplest thing is surely to switch
off some optimisation for this file. As far as I remember, there is even
a switch for this in GNUmake. The danger is that the same problem may
show up in other places as well. How could we prepare for that?

What I don't understand is why this problem doesn't show up for other
backend and why I am the first one affected? Many of you should already
use gcc 4.1, what is so special about the SuSE version? Perhaps nobody
ever uses the xlib backend any more? Could those using gcc 4.1 on a 32
bit Intel machine try at least once?

Fred


I don't know if the problem is related to gcc 4.1 because I've two Intel 
machines, the first with linux (Fedora 4) and gcc 4.1 and the second with 
darwin 8.0.1 and gcc 4.0.2. The problem has appeared on both the machines the 
last week, after updating from svn.

On the linux box gcc is:

Target: i686-pc-linux-gnu
Configured with: ../gcc-4.1.0/configure --prefix=/usr/local --enable-shared 
--enable-threads --enable-languages=c,c++,objc,obj-c++
Thread model: posix
gcc version 4.1.0

and, on the Darwin one:

Target: i686-apple-darwin8.0.1
Configured with: ../gcc-4.0.2/configure --prefix=/opt --enable-threads 
--enable-languages=c,c++,objc
Thread model: posix
gcc version 4.0.2



___
Gnustep-dev mailing list
Gnustep-dev@gnu.org
http://lists.gnu.org/mailman/listinfo/gnustep-dev

GCC 4.1 problem?

[Fred Kiefer]

While trying to reproduce a problem with the xlib backend on my newly
installed SuSE 10.1 system, I found that menus get only partially
displayed, also the knob in horizontal sliders was missing. With this
few hints I set off to drag down the bug behind it. It took my quite
some time, but now I think I found it. The affine transform used to
compute the state transformation suddenly had a NaN component as TY. How
could this ever happen? I put in a few NSLog statements into
[NSAffineTransform scaleXBy:YBy:] and the problem was gone.

For me this proves that there is a gcc problem (gcc (GCC) 4.1.0 (SUSE
Linux)), but how to deal with it? The simplest thing is surely to switch
off some optimisation for this file. As far as I remember, there is even
a switch for this in GNUmake. The danger is that the same problem may
show up in other places as well. How could we prepare for that?

What I don't understand is why this problem doesn't show up for other
backend and why I am the first one affected? Many of you should already
use gcc 4.1, what is so special about the SuSE version? Perhaps nobody
ever uses the xlib backend any more? Could those using gcc 4.1 on a 32
bit Intel machine try at least once?

Fred


___
Gnustep-dev mailing list
Gnustep-dev@gnu.org
http://lists.gnu.org/mailman/listinfo/gnustep-dev