Re: [MiNT] GCC and Highwire test

[Peter Slegg <p.slegg@scubadivers.co.uk>]

As a further experiment I jut tried to build libTIFF 3.9.4 and got this:

EPDIR=.deps depmode=gcc /bin/sh ../config/depcomp \
/bin/sh ../libtool  --tag=CC   --mode=compile gcc -DHAVE_CONFIG_H -I.     -g -O2 -Wall -W -c -o tif_packbits.lo tif_packbits.c
libtool: compile:  gcc -DHAVE_CONFIG_H -I. -g -O2 -Wall -W -c tif_packbits.c -Wp,-MD,.deps/tif_packbits.TPlo -o tif_packbits.o
\
source='tif_pixarlog.c' object='tif_pixarlog.lo' libtool=yes \
DEPDIR=.deps depmode=gcc /bin/sh ../config/depcomp \
/bin/sh ../libtool  --tag=CC   --mode=compile gcc -DHAVE_CONFIG_H -I.     -g -O2 -Wall -W -c -o tif_pixarlog.lo tif_pixarlog.c
libtool: compile:  gcc -DHAVE_CONFIG_H -I. -g -O2 -Wall -W -c tif_pixarlog.c -Wp,-MD,.deps/tif_pixarlog.TPlo -o tif_pixarlog.o
../mpfr-3.0.0/const_pi.c:82: MPFR assertion failed: (__builtin_constant_p (1) && (1) == 0 ? ((S)->_mpfr_exp < ((mpfr_exp_t)((~((~(unsigned long)0)>>1))+3)) ? (((S)->_mpfr_exp == ((mpfr_exp_t)((~((~(unsigned long)0)>>1))+2))) ? mpfr_set_erangeflag () : (void) 0), 0 : ((S)->_mpfr_sign)) : mpfr_cmp_ui_2exp ((S),(1),0)) < 0
tif_pixarlog.c: In function 'PixarLogMakeTables':
tif_pixarlog.c:501:5: internal compiler error: Abort
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
make[2]: *** [tif_pixarlog.lo] Error 1
make[2]: Leaving directory `/i/usr/src/libtiff-3.9.4/libtiff'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/i/usr/src/libtiff-3.9.4/libtiff'
make: *** [all-recursive] Error 1

I suspect gcc is not very well.


Peter


On Sat, 12 Feb 2011 19:32:24 , Peter Slegg <p.slegg@scubadivers.co.uk> wrote:
> I made another build of HW for 020-060 but it is still unstable.
> I am not sure which of the rpm installs caused it.
>
> It crashes quite easily and left a file with this name
>
> ____0_
>
>
> The download is here:
>
> http://www.p.slegg.scubadivers.co.uk/download/highwire060.app
>
> As I write this, I just checked the stack and it is still 64k.
> I will fix the stack size and have another go.
>
> Peter
>