[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#150755: gcc-3.1: -fprofile-arcs option is broken



reassign 150755 ccache
thanks

Michel LESPINASSE writes:
> On Sun, Jul 14, 2002 at 04:57:49PM +0200, Matthias Klose wrote:
> > Ok, I didn't work with profiling before. Please could you send me a
> > small example program and a Makefile to show the problem?
> 
> Yes. However in the mean time I have figured out that the problem was
> actually due to ccache (if I disable ccache everything is fine). Here
> is my small test case:
> 
> ------------ foo.c ------------
> #include <stdio.h>
> 
> int main (void)
> {
>     int i;
> 
>     for (i = 0; i < 10; i++)
>         printf ("i = %d\n", i);
> 
>     return 0;
> }
> -------------------------------
> 
> ---------- build script -------
> #! /bin/sh
> 
> # build with -fprofile-arcs
> gcc-3.1 -fprofile-arcs foo.c -c -o foo.o
> gcc-3.1 -fprofile-arcs foo.o -o foo
> 
> # create arcs information
> ./foo
> 
> # rebuild with -fbranch-probabilities
> # (the build takes the arcs information into account)
> gcc-3.1 -fbranch-probabilities foo.c -c -o foo.o
> gcc-3.1 -fbranch-probabilities foo.o -o foo
> -------------------------------
> 
> I have ccache installed, if I run ./build I get the warning message,
> if I run CCACHE_DISABLE=1 ./build everything works fine.
> 
> Can you reassign the bug to ccache ? I'm not sure I can do it from my
> side.  It'd be easy enough to workaround, ccache.c already has a check
> for bad options at line 564, it should be extended to also check for
> -fprofile-arcs and -fbranch-probabilities there. And the bug priority
> should probably be normal or maybe even only feature-request.
> 
> -- 
> Michel "Walken" LESPINASSE
> Is this the best that god can do ? Then I'm not impressed.


-- 
To UNSUBSCRIBE, email to debian-gcc-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org



Reply to: