[Python-Dev] snprintf breaks build

[Tim Peters]

[Fred]
>   Hmm.  I think the issue with known existing snprintf()
> implementations with Open Source licenses was that they were at
> least somewhat contanimating.  I'll switch back to sprintf()
> until we have a portable snprintf() implementation.

Please don't bother!  Clearly, I've already fixed it on my machine so I
could make progress.  I'll simply check it in.  I didn't like the excessive
cleverness with the fmt vrbl anyway (your compiler may not complain that you
can end up passing more s[n]printf args than the format has specifiers to
convert, but it's a no-no anyway) ....