[patch, fortran] Fix CMPLX documentation, document COMPLEX intrinsic.

[Steve Kargl]

On Sat, May 05, 2007 at 11:08:18AM +0200, Daniel Franke wrote:
> On Friday 04 May 2007 23:14:53 Brooks Moses wrote:
> > The attached patch fixes the CMPLX documentation to describe the result
> > kind, and documents the COMPLEX intrinsic (which was somehow left
> > entirely undocumented).
> 
> Is DCMPLX a specfic name of CMPLX or COMPLEX? If yes, its node should be 
> removed and it should be merged into the appropiate node in a "Specific Name" 
> section?!

dcmplx is not the specific name of either cmplx or cmplex.
It is similar to cmplx in that it returns a double complex
such that it's inputs can be any of integer, real, or complex
with any kind type.  dcmplx also lacks the 3rd optional 
argument of cmplx.  Finally, dcmplx is an extension.

-- 
Steve