The thing I sent is a gcc error that is preventing it rev 1661 from
building on Mac, it's not a warning...
see here: http://thorbrian.com/pygame/builds.php
so I guess what I'm asking is this something that needs fixing?
On Wed, Oct 15, 2008 at 4:33 PM, Lenard Lindstrom <len-l@xxxxxxxxx
<mailto:len-l@xxxxxxxxx>> wrote:
Hi Brian,
It means that the OS X gcc compiler is unhappy with a function
alias I created. So I made it the weak alias it wants for its ELF
object file. The 'weak' option generates a warning in MinGW, but
is otherwise ignored. Since Windows Pygame builds use a
precompiled scale_mmx32 it is not a problem.
Lenard
Brian Fisher wrote:
From a mac OS X 10.5 (Intel) build 12 minutes ago from my
build farm:
scale_mmx32.c:451 last rev: 1658:lenard
ERROR:only weak aliases are supported in this configuration
So what does that tell you?
On Wed, Oct 15, 2008 at 3:36 PM, Lenard Lindstrom
<len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>
<mailto:len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>>> wrote:
Great! That means the smoothscale filter routines with the
inline
assembly code compiles properly on non-Windows systems. It also
means I didn't mess up the cut-and-paste in recovering
Richard's
original code from SVN. Now to see the Visual C 2003
toolkit will
link a MinGW compiled object file. If that works then Brian
Fisher's Pygame snapshots will also support MMX/SSE. After that
there's Visual Studio 2008.
Lenard.
René Dudfield wrote:
hi,
I get 'SSE'.
I'm on a core 2 duo... so it includes SSE.
The example seems to run ok.
cu,
On Thu, Oct 16, 2008 at 9:05 AM, Lenard Lindstrom
<len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>
<mailto:len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>>> wrote:
Hi again,
Could you try out this from a python session:
import pygame
pygame.transform.get_smoothscale_backend()
This should return 'SSE'. If so, then it means Richard
Goedeken's original
smoothscale assembly code is available for SSE capable
machines. The MMX
version is still there for older processors. If you
could
also try out
examples\scaletest.py that would be great.
Lenard
René Dudfield wrote:
cool. It's working here on linux.
cheers,
On Thu, Oct 16, 2008 at 8:44 AM, Lenard Lindstrom
<len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>
<mailto:len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>>> wrote:
Hi René,
Thanks, I tested it and made a repair that
should
only affect Windows. I
also removed the problem transform_test.py
test.
It was to confirm that
MMX/SSE support was only available on x86
processors. Unfortunately
platform.machine() can return something
other than
the documented 'i386'.
Now to waiting for the "The Spectacularly
Adequate
Automated Pygame Build
Page".
Lenard
René Dudfield wrote:
hi,
Committed revision 1659.
Only tested on linux.
cheers,
On Thu, Oct 16, 2008 at 7:34 AM, René
Dudfield
<renesd@xxxxxxxxx
<mailto:renesd@xxxxxxxxx> <mailto:renesd@xxxxxxxxx
<mailto:renesd@xxxxxxxxx>>>
wrote:
... I see you're adding scale_mmx.c
in the
setup.py ... However I
think it needs src/scale_mmx.c ?
Maybe instead of adding the scale_mmx.c
there, you could make it so
that the scale_mmx.c is added in the
normal way(to Setup.in), and then
replaced with the .o by your mod?
cheers,
On Thu, Oct 16, 2008 at 7:29 AM, René
Dudfield <renesd@xxxxxxxxx
<mailto:renesd@xxxxxxxxx>
<mailto:renesd@xxxxxxxxx
<mailto:renesd@xxxxxxxxx>>>
wrote:
hi,
I get this error:
gcc -pthread -fno-strict-aliasing
-DNDEBUG -g -fwrapv -O2 -Wall
-Wstrict-prototypes -fPIC
-D_REENTRANT
-I/usr/X11R6/include
-I/usr/include/SDL
-I/usr/include/python2.5 -c
src/rotozoom.c -o
build/temp.linux-i686-2.5/src/rotozoom.o
gcc -pthread -fno-strict-aliasing
-DNDEBUG -g -fwrapv -O2 -Wall
-Wstrict-prototypes -fPIC
-D_REENTRANT
-I/usr/X11R6/include
-I/usr/include/SDL
-I/usr/include/python2.5 -c
src/scale2x.c -o
build/temp.linux-i686-2.5/src/scale2x.o
gcc -pthread -fno-strict-aliasing
-DNDEBUG -g -fwrapv -O2 -Wall
-Wstrict-prototypes -fPIC
-D_REENTRANT
-I/usr/X11R6/include
-I/usr/include/SDL
-I/usr/include/python2.5 -c
scale_mmx.c -o
build/temp.linux-i686-2.5/scale_mmx.o
gcc: scale_mmx.c: No such file or
directory
gcc: no input files
error: command 'gcc' failed
with exit
status 1
On Thu, Oct 16, 2008 at 7:13 AM,
Lenard Lindstrom
<len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>
<mailto:len-l@xxxxxxxxx
<mailto:len-l@xxxxxxxxx>>>
wrote:
Hi everyone,
I have just changed how
pygame.transform is built (SVN
rev. 1658). I
can
only test this with
Windows. So if
anyone has problems with
Linux or
OS
X
let me know and I will try
and fix
it or back out the changes.