Hi,
> * Aleksej Saushev [2009-08-30 01:42:40 +0400]:
>
> cc -DHAVE_CONFIG_H -I. -I/tmp/wip/clisp/work/clisp/modules/syscalls/gllib
> -I..-I/usr/pkg/include -I/usr/include
> -I/tmp/wip/clisp/work/clisp/src/gllib -O2 -mfpmath=sse -msse3 -march=nocona
> -finline-functions -fomit-frame-pointe
Hi Simon,
> > Btw, if someone can offer me a login on a real Solaris/x86 box, I would
> > appreciate it.
>
> Try the opencsw.org people, see http://wiki.opencsw.org/buildfarm
Thanks for the tip. With this access, I could verify that various floating-
point related test failures (from test-isfini
I noticed that coreutils' recent .xz tarball was 440MB larger (that >10%)
than the preceding one, and tracked the difference to an error in
maint.mk. I'd set XZ_OPT=9ev rather than XZ_OPT=-9ev.
This fixes it, along with a few other corrections and clean-up.
>From 53181ad5a87facb5d424a04d039606a86
Jim Meyering wrote:
> Also, you've made it so upon failure it gives an
> inferior diagnostic, using perror, rather than "error".
Yes, the error message ": memory exhausted" is more understandable
for an end user than "realloc: Cannot allocate memory". End users don't
even know what 'realloc' is.
