[RFC] What to do for complex::pow(0, 0)

[Paolo Carlini]

Mark Mitchell wrote:

> FWIW, my preference would be to match cpow, in the absence of clear 
> guidance from the ISO C++ committee.  I think consistency between the 
> C and C++ runtimes is desirable, unless explicitly disallowed.

Indeed. Now, for c99 target we are actually calling __builtin_cpow (was 
not the case one week ago due to a trivial oversight) and consistency 
with cpow is basically automatic. In case of non-c99 platforms (thus 
missing, in general, cpow) we open code the complex::pow code (and 
return consistently (0, 0) for zero^zero). Seems a satisfactory solution.

Paolo.