Front page | perl.perl5.porters |
Postings from January 2022
demerphq <demerphq@gmail.com> wrote:
:On Tue, 25 Jan 2022, 23:43 , <hv@crypt.org> wrote:
:> In this case, it's an inline recursive function of which the relevant
:> parts look a bit like:
:>
:> sub do_stuff {
:> my $prod = 1;
:> my $recurse;
:> $recurse = sub {
:> ...
:> for my $new (@interesting_values) {
:> local $prod = $prod * $new; # aspirational
:> $recurse->();
:> }
:> };
:> $recurse->();
:> }
:>
:
:Just a minor nit, maybe not relevant, but that will leak.
[...]
:Solving the leak this causes is the reason we created __SUB__
Ah good point - not hugely relevant in this case (the outer sub is called
at most once), but I had forgotten about __SUB__.
:I think the reason it's not allowed was local $x predates my $x. And it was
:defined to operate on globals. But you could get the same result with a
:global. Just rearrange the clauses a bit. Eg, add an our $prod outside the
:outer sub. Then move and change your my $prod to right before the initial
:call to $recurse as a local $prod.
:
:You could fix the leakage and use globals for $prod by changing the recurse
:to
:
:our $prod;
:local *recurse = sub {};
:local $prod=1;
:recurse();
Just changing the inner call to '__SUB__->()' is enough to fix the leak,
right?
We avoid using globals for good reasons, I wouldn't want to use them
for this unless there really was no alternative.
Hugo
Thread Previous
|
Thread Next