> I would like to propose an additional, perhaps controversial, use.
> 5 Breaking the flow of control:
> #ifdef K5BAIL(x) do {
> ret = x;
> if (ret) {
> /* format error message and whatnot
> * in terms of #x.
> */
>
> goto bail;
> }
> } while (0)
> You can't do this with an inline function, because it can't use goto
> or return.
If you're willing to commit to gcc - which I gather is the case, if
we're even discussing gccisms like __attribute__((__always_inline__)) -
then you can do it really easily and relatively elegantly (relative to
the other ways, that is) with a nested function and a nonlocal goto:
void something_hairy(...)
{
__label__ bail_to;
...variable declarations...
auto void bail(const char *, ...)
__attribute__((__format__(__printf__,1,2),__noreturn__));
auto void bail(const char *fmt, ...)
{ va_list ap;
va_start(ap,fmt);
...
va_end(ap);
goto bail_to;
}
if (0)
{
bail_to:;
if (thing1) cleanup_thing1();
if (thing2) cleanup_thing2();
...
return;
}
thing1 = 0;
thing2 = 0;
...
thing1 = new_thing1();
if (...failure...) bail("can't do it: %d < %d (from %s)",...);
...
}
(Okay, that's not quite true. Using always_inline doesn't necessarily
mean committing to gcc. But _depending on_ always_inline does, ISTM.)
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mo...@rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
