t-constant fail i686-darwin10

David Fang fang at csl.cornell.edu
Thu Jan 14 19:46:37 CET 2010


> Jack Howarth <howarth at bromo.med.uc.edu> writes:
>
>      I also see this failure on 32-bit intel darwin9. The failing section of
>  t-constants.c appears to be the line...
>
>  CHECK_LIMB (PP_INVERTED, pp_inverted_calc);

Hello again,
 	Thanks, Jack, for checking there.  (I don't have darwin9 anymore.)
Can this problem be confirmed on i686-apple-darwin8 as well?

> Curious.  Does your build have a file mpn/invert_limb.asm?
> To where does it point (ls -l mpn/invert_limb.asm)?

lrwxr-xr-x  1 fang  admin    28 Jan  9 12:53 mpn/invert_limb.asm@ -> 
.././mpn/x86/invert_limb.asm

> The various invert_limb.asm files read a starting value from a table,
> and the correct references to such tables from assembly is an trouble
> area for MacOS X.
>
> I will have to ignore the various Mac OS issues until I or somebody else
> with the right qualifications can debug this on actual hardware.
> Speculating about the various problems people are seeing is just a time
> waster.

At some point, I might just pass it along to Apple's bugreporter, if no 
one else has done so.  They actually responded to the last problem I 
filed within days.

Fang

David Fang
http://www.csl.cornell.edu/~fang/
http://www.achronix.com/



More information about the gmp-bugs mailing list