HTTP/2 200 cache-control: max-age=43200 server: Combust/Plack (Perl) vary: Accept-Encoding content-encoding: gzip content-length: 3105 content-type: text/html; charset=utf-8 last-modified: Sun, 12 Oct 2025 02:42:50 GMT date: Sun, 12 Oct 2025 02:42:50 GMT strict-transport-security: max-age=15768000 Front page | perl.perl5.porters | Postings from June 2008

Re: my $@ - a proposal

Thread Previous | Thread Next

Aristotle Pagaltzis wrote:
> * Abigail <abigail@abigail.be> [2008-06-28 04:35]:
>   
>>> Well it's documented that DESTROY isn't guaranteed at end of
>>> scope.
>>>       
>> It is? I wasn't aware of that. I always thought, and taught
>> others, that variables go out of scope when their scope ends,
>> and that if a variable goes out of scope, its DESTROY method is
>> called then.
>>     
>
> You have taught correctly. What isn’t guaranteed, as has been
> mentioned elsewhere in the thread, is the order in which DESTROY
> gets called on objects that are garbage-collected during the same
> scope exit. (Or in global destruction, of course.) However, I’ve
> always understood Perl to guarantee timely destruction – garbage
> is always collected as soon as possible.
If the reference count truly hits zero, DESTROY is guaranteed in Perl 5 
to execute at the point that it hits zero. The guarantee might not be in 
the documentation specifically, but it is definitely in the code.
There are cases, though, where the reference count may not hit zero, and 
the user is unaware. I am thinking along the lines of implicit return 
with sv2mortal, or even simpler reference leaks. These are not language 
problems, as they are probably more related to invalid expectations from 
the caller. Until the actual ';' or end of the containing block is 
executed, there is a chance that the reference doesn't hit zero. Here is 
a case which might surprise some people:
> $ perl -e 'sub DESTROY { print "DESTROY\n" }
> sub new { bless {} }                
> sub true { 1 }                         
> if (main->new->true) { print "TRUE\n" }'
> TRUE
> DESTROY
Even though the return value of new() is only used within the condition 
to call ->true, it seems to exist until after the entire if block is 
executed. TRUE gets executed before DESTROY. This may surprise some people.
Cheers,
mark
-- 
Mark Mielke <mark@mielke.cc>

Thread Previous | Thread Next

• my $@ - a proposal by Yuval Kogman
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by David Nicol
• Re: my $@ - a proposal by Graham Barr
• Re: my $@ - a proposal by David Nicol
• Re: my $@ - a proposal by Mark Mielke
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by Graham Barr
• Re: my $@ - a proposal by Graham Barr
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by Ben Morrow
• Re: my $@ - a proposal by Graham Barr
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by David Nicol
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by David Nicol
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Johan Vromans
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Rafael Garcia-Suarez
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Ben Morrow
• Re: my $@ - a proposal by Elliot Shank
• Re: my $@ - a proposal by Abigail
• Re: my $@ - a proposal by Juerd Waalboer
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Graham Barr
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by Graham Barr
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by David Nicol
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Graham Barr
• Re: my $@ - a proposal by David Nicol
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by Aristotle Pagaltzis
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Raphael_Manfredi
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Raphael_Manfredi
• Re: my $@ - a proposal by Abigail
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Raphael_Manfredi
• Re: my $@ - a proposal by Dr.Ruud
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by Abigail
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by Sean O'Rourke
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by Steffen Mueller
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by David Golden
• Re: my $@ - a proposal by Abigail
• Re: my $@ - a proposal by Aristotle Pagaltzis
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Yuval Kogman
• Re: my $@ - a proposal by Dr.Ruud
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by David Cantrell
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by Moritz Lenz
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by Moritz Lenz
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by Moritz Lenz
• Re: my $@ - a proposal by Sam Vilain
• Re: my $@ - a proposal by Abigail
• Re: my $@ - a proposal by Mark Mielke
• Re: my $@ - a proposal by Aristotle Pagaltzis
• Re: my $@ - a proposal by David Nicol
• Re: my $@ - a proposal by Nicholas Clark
• Re: my $@ - a proposal by David Nicol
• RE: my $@ - a proposal by Jan Dubois
• Re: my $@ - a proposal by David Nicol
• Re: my $@ - a proposal by Reini Urban
• Re: my $@ - a proposal by Dr.Ruud
• Re: my $@ - a proposal by Sam Vilain