Subject: Tests fail for Maxima (HEAD) with ECL 8.12.0
From: Richard Fateman
Date: Wed, 08 Apr 2009 07:03:29 -0700
If the only differences in the test suite runs are the direct or
indirect result of differences in the floating-point arithmetic
precision and/or accuracy, then
the solution may be to make the passage of the test suite insensitive to
such variations.
I think that the Common Lisp standard could not reasonably be made
prescriptive and was descriptive with respect to floating-point arithmetic.
A language that attempts to prescribe bit-wise results of arithmetic
beyond what is specified in the hardware standard is probably a mistake.
RJF