Thanks for all your help. I think we're getting closer. I'm using the
Makefile rules supplied with unpaper:
unpaper:
cc unpaper.c -o unpaper -lm
so I was already using no optimization. I was using gcc-3.2.2 which
is the default compiler on my system. I tried building with gcc-2.95.3
and it worked! It also fails with gcc-3.4.3. Oh wait! When I use
gcc-2.95.3, it also fails when I use -O2! Perhaps gcc-3 is doing some
optimazation automatically that gcc-2 only does when -O is used. Wierd.
Now that I can build both a working and non-working version, I should
be able to track down where the two are diverging. Not sure how quickly
I can get to it though.
--
Peter Fales
Peter "at" fales-lorenz.net
On Fri, Mar 11, 2005 at 10:11:15PM +0100, Jens Gulden wrote:
> Peter Fales schrieb:
> >Now that's wierd. I was using a privately built copy of unpaper, but I
> >just unpacked the tarball and get the same results with the prebuilt
> >binary!
>
> I just reproduced the problem on one remote Intel machine I have access to.
>
> Try to compile without any optimization (no -Ox), this was the only way
> for me to get a usable executable for that machine. Does it help for
> you? (I am using gcc 3.3.5 on Debian.)
>
> Jens
>
> >
> >It seems that the same demo file and binary works for you but gives a white
> >page for me. I can't think of any reason why that would be. My first
> >thought was shared libraries, but unpaper doesn't seem to have any unusual
> >shared library dependencies. Do you invoke any external programs that
> >might be different on my system?
> >
> >I'll try to take a look at the code this weekend and see if I see any
> >ideas. If you have any debugging you want me to do (e.g. strategically
> >placed
> >printf's ) just let me know.
> >
>
>
> --
> sane-devel mailing list: sane-devel@lists.alioth.debian.org
> http://lists.alioth.debian.org/mailman/listinfo/sane-devel
> Unsubscribe: Send mail with subject "unsubscribe your_password"
> to sane-devel-requ...@lists.alioth.debian.org
