Dear Maude Abusers,

A new alpha release can be downloaded from

http://www.csl.sri.com/~eker/Maude/Alpha63a/

Currently only linux is supported. This fixes numerous bugs is the
last alpha release caused by the new treatment of conditions, variables
and left->right sharing of common subexpressions. The fact that _only_
Mark-Oliver reported these bugs leads me to suspect that noone else
tried the last alpha release.

Here is a summary of changes (other than bug fixes) from the last
alpha release:

(1)	Various things are renamed as agreed in previous discussions
	(see prelude.maude).

(2)	The leastSort function now takes membership axioms into
	account (I agreed this change of semantics with Jose yesterday).

(3)	Instead of returning error* for badly formed arguments, the
	descent functions first try any user equations, and failing that
	return an unreduced function call at the kind level; for example:

select META-LEVEL .

red metaReduce(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op 'a : nil -> 'Bar [none] .
  none
  none
endfm, '_+_['1.MachineInt,'12.NzMachineInt]) .


(4)	metaParse now takes an extra argument of sort Type? which
	gives the component to which the result should parse; the
	constant anyType allows any component - the old behavior.
	In the case that metaParse fails it now returns

	noParse(n) if there was no parse where n in the index of the
	first bad token (counting from 0), or the number of tokens in
	the case of unexpected end of input; or

	ambiguity(r1, r2) if there were multiple parses, where r1 and
	r2 are the result pairs correspond to two distinct parses.

	Examples:

red metaParse(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op '1 : nil -> 'Foo  [none] .
  none
  none
endfm, '1 '+ '12, anyType) .  *** OK

red metaParse(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op '1 : nil -> 'Foo  [none] .
  none
  none
endfm, '1 '+ '+ '12, anyType) .  *** bad token

red metaParse(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op '1 : nil -> 'Foo  [none] .
  none
  none
endfm, '1 '+ '2 '+, anyType) .  *** unexpected end of input

red metaParse(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op '1 : nil -> 'Foo  [none] .
  none
  none
endfm, '1, anyType) .  *** ambiguity

red metaParse(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op '1 : nil -> 'Foo  [none] .
  none
  none
endfm, '1, 'Foo) .  *** OK

red metaParse(
fmod 'FOO is
  protecting 'MACHINE-INT .
  sorts 'Foo .
  none
  op '1 : nil -> 'Foo  [none] .
  none
  none
endfm, '1, '`[MachineInt`]) .  *** OK

Please try this version and remember to report bugs!

Steven