Class name mangling

[Per Bothner]

The problem was inconsistent mangling conventions.
One fix is attached - it changes field (and variable)
names to be consistently mangled using the "reversible"
mangling scheme.

Of course it is not obvious this is the best fix.  An
alternative fix would be to use the *non*-reversible
mangling scheme more consistently.  More specifically,
in ClassMemberConstraint when importing a value that
doesn't have a "name" property.  But that solution
doesn't guarantee that we'd get the same name back.
So my inclination is to use the reversible mangling.

See http://www.gnu.org/software/kawa/Mangling.html
for what the manual says about mangling -  that would
of course need to be updated.

Comments, everyone:  What are your thoughts about changing
the mangling for field names (not method names, at least at
this point) to a less natural but reversible mangling?
-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/per/
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: mangle.patch
URL: <http://sourceware.org/pipermail/kawa/attachments/20021130/4950edd9/attachment.ksh>