On Mon, 30 Sep 2002, Brent Dax wrote:
> Andy Dougherty:
> # More generally, though, rather than sprinkling the sources
> # with INTVAL_FMT and other ugly (but correct and portable
> # things), should we be trying to
> # funnel everything through a central printf-like engine and
> # have it automatically provide the correct formats?
> # perl-5.8.0 somewhat belatedly introduced something like that,
> # and it seems like a good idea.
>
> *elbows him in the side and points at /Parrot_v?sn?printf(_[sc])?/ in
> misc.c*
Interesting, yes, that's mostly what I had in mind, but I'm unsure just
how it's intended to be used.[*] That is, if I have
INTVAL iv = 7;
what format do I use to print it? Is it the intent that
Parrot_sprintf_c(interpreter, "i=%d\n", i);
will work? Or do I need some special format character, like
Parrot_sprintf_c(interpreter, "i=%Vd\n", i);
(to borrow perl5's sv.c:Perl_sv_vcatpvfn notation).
As a side note, I see that misc.c currently seems to assume that
sizeof(long) >= sizeof(INTVAL), so I'll need to patch that up anyway, but
at the moment, I'm unclear even as to intention.
[*]As another curious side note, the Parrot_sprintf functions require a
Parrot interpreter. Whether it's safe to pass that interpreter object
around while inside the trace function (which is where I got started) or
inside sundry debug functions is another possibly interesting question,
but largely irrelevant for the bigger issue at hand.
Andy Dougherty [EMAIL PROTECTED]
