axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] --patch-40


From: daly
Subject: [Axiom-developer] --patch-40
Date: Sun, 12 Jun 2005 12:33:56 -0500

*,

patch-40 is up on arch. i've been working on the failures that are
uncovered by the input file tests. i've fixed one failure and found
a workaround for the other. the third wasn't really a failure at all.

elemnum fails because we call lisp functions from within the algebra
DoubleFloat code and these functions do not exist in lisp. i've defined
these functions and they work but i need to write a whole battery of
test cases. plus the definitions are not "double float" but they give
the same answers as the original NAG version. 

which raises the desire for an "infinite float domain" that is similar
in spirit to the infinite precision integer domain. 

stream2.input fails because the interpreter seems to be started in
some sort of badly-initialized state with respect to compiling functions.
the workaround is to do ")set functions compile on" which seems to do
the right initializations. i'll track down the differences shortly.

ode.input failed simply because it did not have ")set break resume"
and the intentionally failing test stopped the input. this is fixed.

====================================================================

Summary: fix input file bugs
Keywords: daly test cases

20050612 tpd --patch-40
20050611 tpd src/input/*.pamphlet \usepackage{axiom}
20050611 tpd src/input/Makefile  add dvi files
20050611 tpd src/input/Makefile  add stream2.output
20050611 tpd src/input/stream2.input add )set functions compile on
20050610 tpd src/interp/vmlisp.lisp define DFLOAT sec, csc, acsc, asec, csch
20050610 tpd src/interp/vmlisp.lisp define DFLOAT coth sech acsch acoth asech
20050609 tpd src/input/Makefile  add elemnum.output
20050609 tpd src/input/elemnum.input add )set break resume
20050609 tpd src/input/Makefile  add ode.output
20050609 tpd src/input/ode.input add )set break resume







reply via email to

[Prev in Thread] Current Thread [Next in Thread]