improving the complex number division

[Gabriel Dos Reis]

dewar@gnat.com (Robert Dewar) writes:

| It is indeed surprising to find this entirely naive and quite incorrect

What actually surprised me is that neither Jérôme nor you  noticed
that that implementation isn't being used for anything we have defined
semantics for.

| implementation of complex division in gcc. This is certainly a computation
| for which absolutely standard algorithms exist.

True.  But, as noted in the reply I made to Jérôme's message, that
implementation wasn't never ever used for the cases the C++ standard
defined semantics for (i.e. specializations of std::complex<T> with 
T=float, T=double, T=long double).

| I am not clear whether
| Jerome is quoting one of these standard algorithms

That is standard algorithm.

| (in which case a reference would be helpful),

I believe Goldberg's paper or G. Stewart's book on matrix computations
covers it.

-- Gaby