RexxMathLib 38.02

Short:        Faster & shorter math support for ARexx.
Author:       thor@einstein.math.tu-berlin.de (Thomas Richter)
Uploader:     thor einstein math tu-berlin de (Thomas Richter)
Type:         util/rexx
Version:      38.02
Requires:     Os 2.04
Architecture: m68k-amigaos

This library provides transcendential functions for the
ARexx - programming language. It is a completely new revision of 
Willy Langeveld's rexxmathlib, although no original code has been used.
The new release has been completely rewritten in assembly language and
is therefore not only faster (approx. 10 times), but provides also
a higher precision of 15.9 digits, thanks to smarter ASCII- to float 
conversion routines.

The rexxmathlib.library uses the system math libraries, namely
the mathieeedoubbas.library and the mathieeedoubtrans.library and will
therefore work fine, regardless of a math-coprocessor.
____________________________________________________________________________

Changes since 38.02:

	- LibInit didn't release all memory in case of failure.
	- LibOpen might have run into a Wait() state and might hence
	  have caused some race conditions. Did not clear DEL_EXPUNGE
	  as well. Fixed.
	- Added two new functions, DEG() and RAD() to convert between
	  degrees and radiants.
	- The precision of atanh() has been increased in case a FPU
	  is available.
	- The library does no longer produce a unary plus sign since some
	  programs seem to have trouble with it.

____________________________________________________________________________

Changes since 38.01:

	- Fixed some typos in the guide, reworked it.
	- Haven't found any bugs in the library, yet.

____________________________________________________________________________

Thomas Richter,
	November 1999