This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: failed to compile gcc-4.3-20061209/gcc/varasm.c on OSX 10.3
- From: Steve Ellcey <sje at cup dot hp dot com>
- To: toa at pop dot agri dot ch
- Cc: gcc at gcc dot gnu dot org
- Date: Mon, 11 Dec 2006 10:33:34 -0800 (PST)
- Subject: Re: failed to compile gcc-4.3-20061209/gcc/varasm.c on OSX 10.3
Andreas Tobler wrote:
> Dominique Dhumieres wrote:
>>...
>>cc1: warnings being treated as errors
>>../../gcc-4.3-20061209/gcc/varasm.c: In function 'elf_record_gcc_switches':
>>../../gcc-4.3-20061209/gcc/varasm.c:6268: warning: format '%llu' expects type 'long long unsigned int', but argument 3 has type 'long int'
>>../../gcc-4.3-20061209/gcc/varasm.c:6275: warning: format '%llu' expects type 'long long unsigned int', but argument 3 has type 'long int'
>>../../gcc-4.3-20061209/gcc/varasm.c:6283: warning: format '%llu' expects type 'long long unsigned int', but argument 3 has type 'long int'
>>../../gcc-4.3-20061209/gcc/varasm.c:6302: warning: format '%llu' expects type 'long long unsigned int', but argument 3 has type 'long int'
>>make[3]: *** [varasm.o] Error 1
>>
>> Any idea around about the cause and/or the way to fix it?
>
> This is known to break on all 32-bit targets. (afaik) On 64-bit targets
> it works.
>
> You can either wait until the patch is reverted or the correct fix is done.
Do you know if there a GCC bug report for this defect? I couldn't find
one in bugzilla. I am seeing this problem with IA64 HP-UX on ToT. I
tried the workaround you gave and that makes IA64 HP-UX work but causes
other platforms to fail so I am wondering when there will be a real fix
for this bootstrap problem.
Steve Ellcey
sje@cup.hp.com