Perl Monk, Perl Meditation | |
PerlMonks |
comment on |
( [id://3333]=superdoc: print w/replies, xml ) | Need Help?? |
you are missing gcc's 32-bit dependencies
Yes - and I'm not really interested in building such a perl if I need additional downloads to do it. I've implemented USE_QUADMATH capability on Windows. There's nothing public yet, but I should be in a position to submit a PR in the next couple of days. I've been testing all of the various nvtype/ivtype permutations with 32-bit and 64-bit compilers, and they're all fine .... except for one. On Windows, with a 32-bit gcc, I can build blead with ivsize=4 and nvtype=__float128 - and it's failing a few Math::BigInt/BigFloat tests. There's no problem with any of the other available nvtype/ivtype permutations - it's just this one (very rare) configuration that's being a bit of a nuisance. I suspect it's a Math::BigInt/BigFloat bug, but it could also be a bug in my amendments to the perl source. For reference, it would be nice to see how a __float128 build with an ivsize of 4, fares on Ubuntu. The 2 systems I have that do permit the -m32 switch don't provide libquadmath. Cheers, Rob In reply to Re^2: How to build perl-5.32.0 (or later) with ivsize of 4 on Ubuntu-20.04 ?
by syphilis
|
|