Chaim Frenkel wrote:
>
> Tony Olekshy wrote:
> >
> > If you write this:
> >
> > try { my $p = P->new;
> > my $q = Q->new;
> > }
> > finally { $p and $p->Done;
> > $q and $q->Done;
> > }
> >
> > what happens if both constructors succeed, but $p->Done dies?
>
> try { my $p = P->new; }
> try { my $q = Q->new; }
> finally { $p and $p->Done; }
>
> If P barfs then there is no q. If q barfs, then there is only a p.
Please read more carefully. I did not ask what happens when P->new
dies. I asked what happens when $p->Done dies. $q->Done is supposed
to get called anyway.
> And if you argue that there is more complicated stuff going on in
> the try block. Then you'll never get anything worked out.
Speak for yourself. We regularly do stuff like this:
try {
my $p = P->New...
.
. handful of lines here
.
$foo and my $q = Q->New... # might not happen
.
. handful of lines here
.
}
finally { $p and $p->Done; }
finally { $q and $q->Done; }
> Doing a whole series of finally's will not help you without thinking
> about the problem of restoring the invariants.
I'm well aware of that, and perfectly capable of doing so. The problem
is, once I've done so and find I need to make sure each constructor is
paired with a Done, how do I do that in Perl. RFC 88 just makes it
easier, no more than that.
Yours, &c, Tony Olekshy
