develooper Front page | perl.perl5.porters | Postings from September 2014

Re: new long doubles option: gcc libquadmath

Thread Previous | Thread Next
Jarkko Hietaniemi
September 22, 2014 11:12
Re: new long doubles option: gcc libquadmath
Message ID:
On Monday-201409-22, 4:30, demerphq wrote:
>     Moi?  Facetious?  I don't even know that word.
> Josh said that, not me.

I know.  Doesn't matter.

 > "ld" meaning "larger than double" instead of a specific size, and 
having to use $CONFG{longdblkind} to find out what kind it is makes me a 
bit concerned about things like Storable and Sereal. We will make Sereal 
"work" for some reasonable definition of work, but Storable seems a 
bigger concern.

I think this is the case.  We may have just been saved, "de facto" by
the combination of two things: (1) uselongdouble is rare (2) given the 
overwhelming x86 population the one long double flavor is overwhelming

Though this theory suffers somewhat from the apparent rarity of 12 vs
16 byte problems... (the most common long double format in 32 vs 64 bit
platforms)... but maybe so many people (using Perl and long double)
are already using 64-bit<handwave/>

But the fact remains: if someone has been pumping out the 16-byte
IEEE 128-bit quad prec long doubles, the x86 80-bit readers (be they
32 or 64 bit) won't be able to decipher them.  And vice versa.  Or
the PPC/MIPS double-doubles.

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About