Linearness check about optima

[m2ym]

See #45

[fare]

This should probably be done in parse-pattern, which should take and pass around an extra optional argument to store the current list of variables. To scale to large computer-generated patterns, we should use a hash-table rather than a list -- and since there are no built-in pure hash-tables, we should remove entries from the table when we walk back in the tree.

[fare]

Why is this issue closed? Should I reopen? Should I add the comment to issue 45 instead?

[m2ym]

I've just closed this issue as dup #45