Re: [Axiom-developer] A curious algebra failure

[Gabriel Dos Reis]

"Weiss, Juergen" <address@hidden> writes:

| Hi,
| 
| The file patches.lisp in the scratchpad sources from 1990 contained the
| following line:
| 
| (defun |pmatch| (s p) (|pmatchWithSl| s p '(ok))) ;assoc with atom at
| eol
| 
| to override the definition in c-util.boot.
| 
| Seems that pmatchWithSl returns nil when a match
| is not possible, empty assoc list != nil when no substitutions are
| necessary
| for a match and an assoc list with substitutions otherwise. 

Thanks for clarifying further the intended semantics.

| 
| For example
|   v:= ASSOC(p,al) => s=rest v or al
| returns 't if (p,s) is in al.
| 
| So this function can never typecheck in ansi common lisp. The
| usage seems to come from a time when alists could contain
| atomic properties and associations as well.

Indeed.

| By the way, I think the line must read
| 
|   v:= ASSOC(p,al) => s=rest v and al
|                               ====
| 
| this makes much more sense logically and should typecheck with 
| the   pmatch(s,p) == pmatchWithSl(s,p,[nil])


Yes, it typechecks; I just started a new build; I will report back.
I like your semantics description and suggested fix better.

-- Gaby