Fedor Sergeev <Fedor.Sergeev at sun.com> writes: > Nor I have seen any requests from customers flying around about that.
That's at least a data point. > As decimal floating point support is not likely to be included into the > upcoming > C and C++ standards (at least not sought out as that right now) it is > "categorized" > for SunStudio compiler project as a useful gcc feature. It's obviously not only a GCC feature: from http://www.google.de/url?sa=t&source=web&ct=res&cd=1&ved=0CAcQFjAA&url=http%3A%2F%2Fwww.ac.usc.es%2Farith19%2Fsites%2Fdefault%2Ffiles%2F3670a225-spec-session-DFP-paper2.pdf&ei=aw0gS_q0OonE-Qadv62BDA&usg=AFQjCNH0q579_YEVFUOhmap8wfCgTapkpg it seems that at least HP cc on IA-64 and Intel C++ implement it as well. And Power6 even supports it in hardware, and IBM's xlc supports that, of course. Obviously other vendors see the need, and it would be good for Sun to at least follow, if it cannot lead the pack. > We do work on "useful gcc features", but honestly this one does not have a > chance to hit the first > half of my own priority list. > > Would it move from a technical report into the standard, that would change > alot. There seems to be a misunderstanding here: pending the next revision of the C standard, type 1/2 ISO TRs are as close to standardization as you can get. Rainer -- ----------------------------------------------------------------------------- Rainer Orth, Center for Biotechnology, Bielefeld University
