C++ PATCH: static_cast regression

[Mark Mitchell]

>>>>> "Martin" == Martin v Loewis <martin@loewis.home.cs.tu-berlin.de> writes:

    >> But I'm uncomfortable with the (widespread in g++) practice of
    >> coding up routines to do things that the standard specifies,
    >> but in a way that bears no resemblance to *what* the standard
    >> specifies.

    Martin> It is really nobodies fault that the final algorithm
    Martin> differs from what g++ does. If there is an opportunity,
    Martin> rewriting such algorithms from scratch is often the right
    Martin> thing, if time permits.

I agree, on both counts.  I by no means meant my message to be
construed so as cast aspersions on the illustrious forefathers of g++!

Rather, I was simply pointing out that in looking at the static_cast
routine in G++, I'm really not convinced it's handling all the various
cases correctly. :-(

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com