[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Gcl-devel] Floating point overflow
From: |
Camm Maguire |
Subject: |
[Gcl-devel] Floating point overflow |
Date: |
18 Oct 2005 21:29:43 -0400 |
User-agent: |
Gnus/5.09 (Gnus v5.9.0) Emacs/21.2 |
Greetings! Looking at the ansi tests which point out GCL's current
failure to signal a floating point overflow error on certain
arithmetic operations. We've separately discussed the possible
utility in pushing inf and nan forward to the user, which is a good
idea I feel regardless of signaling behavior. If memory serves, one
can instruct gcc to compile code which sends the process a signal on
such floating point exceptions, though I don't know how portable this
is. Short of that, manually checking for inf and nan in all
arithmetic functions appears quite cumbersome, though we may be able
to boil it down to a few primitives. Just thought I'd solicit ideas
on the best behavior during overflow from those expressing related
interests in the past.
Take care,
--
Camm Maguire address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah