I've been thinking about how variables work within LoGS. Vijay and I have
had a couple of offline conversations about this.
I think that the way that variables work right now is broken and needs to
be addressed, re-thought, re-worked, etc.
Right now, for a variable to be set in a match function, it must be
returned as a list of lists of variable and value like:
((FOO 42)(BAR "something")) to set FOO to 42 and BAR to "something".
These variables are then bound with a progv (creating a dynamic variable
binding, which, I believe is the problem) before the action function is
How else could we, essentially, maintain a stack of these variables (so
that a ruleset could bind a variable, and a rule within that ruleset bind
the variable to a different value) ?