A gmp flavored gcc?
Sylvain.Pion at sophia.inria.fr
Fri Oct 27 21:28:47 CEST 2006
Paul Zimmermann wrote:
>>I recall from some time ago (perhaps on another gmp list) that there was=20
>>talk of creating a gmp flavored gcc. The basic idea being that having a=20
>>gcc with builtin big numbers would be cool, and combining gcc and gmp in th=
>>way would save lots of compiler woes. Of course, "porting gmp" to=20
>>"porting gcc" is not a reduction. Anyway, I found myself daydreaming=20
>>about this the other day and I was wondering if anyone had given that=20
>>idea more thought?
> I don't know about gcc, but I'm aware the C++ committee is seriously
> considering providing an arbitrary precision "integer" type, which would
> be implemented using gmp of course.
> More details can be obtained from Sylvain Pion (in cc).
The current status is that the C++ committee appears to be demanding for
such a functionality, but it's mostly a volunteer organization, and the
current proposals for this functionality have been rejected (on the C++
API aspects). So it's "only" missing a volunteer to draft something
with a higher acceptance probability. If nothing acceptable gets
proposed, nothing will happen. At least this is my understanding of
the current status.
Of course if something like this was standardized, it's highly probable
that the GCC implementation would use GMP/MPFR underneath.
As far as C++ is concerned, such a type would belong to the library, it
would not be a new language built-in type.
More information about the gmp-discuss