[Patch] Second half of tr1/complex

[Paolo Carlini]

Hi,

the below would be the remaining part of tr1/complex, the additional
overloads.

Actually, these are the most thorny and debated bits (same for the real
counterparts, of course) but I believe we are closely following the
letter and the spirit of the actual TR1. From the latter point of view,
there is a problem with the existing std::pow(const complex<>&, int)
overload which prevails over the new pow overloads and makes impossible
the effective promotion to complex<double> for a pow(complex<float>,
int) call. Well, some may consider this behavior even a virtue ;)
Otherwise, as far as I can see, we gonna need a mechanism to change,
upon user request of strict conformance to TR1, the behavior of existing
std facilities... worth some long thoughts...

Tested x86-linux.

Paolo.

////////////////////
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: CL_tr1_8_1_rest
URL: <http://gcc.gnu.org/pipermail/libstdc++/attachments/20060113/c3d26f1c/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: patch_tr1_8_1_rest
URL: <http://gcc.gnu.org/pipermail/libstdc++/attachments/20060113/c3d26f1c/attachment-0001.ksh>