This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: optimization/10393: gcc does not generate code for loop
- From: Dag Agren <dagren at abo dot fi>
- To: Eric Botcazou <ebotcazou at libertysurf dot fr>
- Cc: gcc-bugs at gcc dot gnu dot org, <nobody at gcc dot gnu dot org>, <gcc-gnats at gcc dot gnu dot org>
- Date: Wed, 16 Apr 2003 04:00:35 +0300 (EET DST)
- Subject: Re: optimization/10393: gcc does not generate code for loop
On Tue, 15 Apr 2003, Eric Botcazou wrote:
> > Hmm, trying to reproduce it myself from the .ii file fails. Also, the
> > generated assembler output does not look familiar to what I got in my
> > earlier tests. This is most confusing, as I did not save the orignal
> > source file. Would gcc ever produce different output from a .cpp and the
> > corresponding .ii file?
>
> Certainly not, otherwise we would not be able to debug GCC at all.
>
> Maybe you looked at a different but similar part of the assembly file the
> first time? Or you used different compile options?
I tried to recreate the compile options I used, but to no avail.
I did look at the same part - in the earlier compile, the glEnd() and
glDisable() calls that are inside and outside the loops, respectively,
were right after each other, with no branch in between. In the new
version, the glDisable() is much higher up in the code than the glEnd(),
because the loop structure is different.
Most mystifying.
I did compile it with gcc 3.2 at first, and when that didn't work I
installed 3.2.2 on top of it, and it produced the same results. I did not
reboot the system in-between, but I don't see how that could have affected
anything.
--
Dag Agren <> d at c3 dot cx <> http://www.abo.fi/~dagren/ <> Legalize oregano