Re: tail call reduction

[Stefan Monnier <monnier@iro.umontreal.ca>]

> I don't recall the actual cases where I bumped into problems. IIRC I
> got beyond `max-lisp-eval-depth' once or twice with recursive
> functions on a depth-first traversal of medium-sized trees. Since the
> default of said variable is 300, this should be o.k.. Hm, well, I
> guess it was with a loop implemented by recursive calls.

The depth used by Emacs might be significantly smaller if the code is
byte-compiled.  And you can bump up the max-lisp-eval-depth (you can even
bump it locally inside your code with a let-binding).

Or you can change your algorithm to use trees that are better balanced and
will thus not be so deep (though not always if you're using something like
splay trees).


        Stefan