CONTEXT@WASHINGTON.ARPA.UUCP (06/12/86)
The CP/M-68K v1.2 upgrade I received from TriSoft included the floating point support for the C compiler (greatly appreciated), but there was no MATH.H file, and the tan() and atan() functions were missing from both the Motorola and IEEE libraries. Has anyone else noticed the lack, or is this an isolated incident that I should resolve with TriSoft? -- Ron -------
HELLER%cs.umass.edu@CSNET-RELAY.ARPA.UUCP (06/13/86)
There is no MATH.H. atan() does not exist (the manual is wrong). tan() is there (it is bundled with sin() and cos()), but the C hook is missing. I have written a short as68 file (i disassembled cos() and sin()) for tan(). I can post it to this mailling list if desired. It will actually work for either library - all it does is save registers, and move the argument and a flag to registers and then jumps to entry point in the tan()/sin()/cos() shared code. Robert
mknox@NGP.UTEXAS.EDU (mknox) (06/14/86)
Ron, I just double checked the TriSoft release, and also a release from IQ Software. Neither had _tan() or _atan() in the library. Nobody every noticed that before. [Proving what I always suspected. Most of us try to avoid floating point numbers like the plague.] I also double checked the master list of OEM files and see no reference to a MATH.H include file. However it would not be the first time DRI managed to mess up a distribution. Meanwhile, still awaiting some kind of answer from DRI on 1.3 floating point. Anyone else find atan or tan??? mknox@ngp.utexas.edu