[m-rev.] for review: disable_warnings scope

[Zoltan Somogyi]

On Tue, 10 Jan 2017 00:38:47 +1100 (AEDT), "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> I just checked, and
> 
> - when it finds a tailcall, ml_tailcall.m marks the MLDS call statement
>    as a tail call for ml_optimize.m, but
> - ml_optimize.m IGNORES this mark on the call statement, and does
>    its OWN test on each call to see whether it is a self-tail-recursive call.
> 
> Given this fact, I am not sure whether ml_tailcall.m does anything useful
> at all, which is in itself a bit worrying.

Apparently, I didn't check hard enough. ml_tailcall.m does ignore
the "is this a tail call" field in ml_stmt_calls, but it calls other code
which pays attention to it. So disregard the above paragraph.

Zoltan.