Re: Type of max_specpdl_size and max_lisp_eval_depth

[Stefan Monnier]

> > On principle, I have no objection.  I don't know if there is some
> > technical problem hiding in the change.
> 
> It's a long-standing issue, which causes subtle bugs on 64-bit
> machines and when values become large.

Right, my patch is meant to fix some of those problems.

> I'd say let's install this and see how bad can that be.  (I'm guessing
> Stefan is already running the patched version for at least a day.)

Problem is: on 32bit systems (like the ones I'm using), my patch is really
a noop since it just changes some `int's to `EMACS_INT' but EMACS_INT is
#defined to `int' anyway on those systems.
So if someone could try it out on a 64bit system... (a comparison of
`gcc -Wall' output before and after the patch might be helpful,
although I doubt it).


        Stefan