On Wed, 2002-05-08 at 09:29, Melvin Smith wrote: > > Thanks for that, Bryan! I knew someone had posted some stuff claiming > to have some float conversion stuff, but I couldn't remember who. > > This is a very good compilation of the issues we have to address. > > I implemented a byteordering strawman which works, I still need > to go back through and make sure we optimize for loading native bytecodes. > My approach for byteordering was a bit different; the goal being > 'simple' rather than 'clever', but its done with no #ifdefs. I'll > extend it to handle 128bit values as you mentioned.
If you can. I've not been successful in coming up with a comprehensive inline method of doing float conversions. > > Floats are where I'm in new ground, since I just don't do any > work with portable floats, so I'd appreciate anything you wish > to contribute. If its code you have and you want it used, you can > shoot it to me and I'll try to integrate, and credit you of course; > otherwise, I'm going to keep moving forward I hope. I'll post non-code first. (I've legal issues that haven't been hammered out yet. -- Bryan C. Warnock bwarnock@(gtemail.net|capita.com)