Re: 1.4.10rc1 vs. OS X 10.6

2009-09-02 Thread Robert J. Hansen
I tried compiling 1.4.10rc1 on Mac OS X 10.6 without success. I can recreate this bug on 1.4.9 and 1.4.10rc1 on a MacBook Pro running Snow Leopard. I can also confirm that John's fix of passing --disable-asm to the configure script works. Can we get an #ifdef for Darwin to replace the ASM

Re: 1.4.10rc1 vs. OS X 10.6

2009-09-02 Thread David Shaw
On Sep 2, 2009, at 12:47 AM, Joseph Oreste Bruni wrote: I tried compiling 1.4.10rc1 on Mac OS X 10.6 without success. 10.6 ships with a newer version of the compiler toolchain that is giving a few headaches here and there. Until we work out the issue, just compile with --disable-asm.

Re: 1.4.10rc1 vs. OS X 10.6

2009-09-02 Thread Benjamin Donnachie
2009/9/2 Robert J. Hansen r...@sixdemonbag.org: Can we get an #ifdef for Darwin to replace the ASM blocks with compiled code? I tested it as part of my Universal Binary build, and the ASM blocks make little difference with fast processors. Ben ___

Re: 1.4.10rc1 vs. OS X 10.6

2009-09-01 Thread Joseph Oreste Bruni
On Sep 1, 2009, at 9:47 PM, Joseph Oreste Bruni wrote: I tried compiling 1.4.10rc1 on Mac OS X 10.6 without success. During make the compile bombed here: ... mv -f .deps/mpih-mul.Tpo .deps/mpih-mul.Po gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include-g -O2 -Wall -Wno- pointer-sign -MT

Re: 1.4.10rc1 vs. OS X 10.6

2009-09-01 Thread John Clizbe
Joseph Oreste Bruni wrote: I tried compiling 1.4.10rc1 on Mac OS X 10.6 without success. During make the compile bombed here: ... mv -f .deps/mpih-mul.Tpo .deps/mpih-mul.Po gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include-g -O2 -Wall -Wno- pointer-sign -MT mpiutil.o -MD -MP -MF