HTTP/2 200 cache-control: max-age=43200 server: Combust/Plack (Perl) vary: Accept-Encoding content-encoding: gzip content-length: 1873 content-type: text/html; charset=utf-8 last-modified: Mon, 13 Oct 2025 16:28:25 GMT date: Mon, 13 Oct 2025 16:28:25 GMT strict-transport-security: max-age=15768000 Front page | perl.perl5.porters | Postings from January 2015

[perl #123698] 5.20 threaded regression: boolean.pm -truth option breaks string literals in closures

Thread Previous

On Thu Jan 29 07:06:25 2015, dagolden@cpan.org wrote:
> On Perl 5.20.X with threads, when boolean.pm uses the '-truth' option
> on import -- which replaces the interpreter's internal true/false
> values with boolean.pm objects -- we've seen a case where a string
> literal in an anonymous function becomes undefined, but only if a
> string assignment happens in the same function.
> 
> The attached program demonstrates the issue.
> 
> On Perl 5.18.2-threaded and Perl 5.20.0-not-threaded the output is:
>     calling not_broken:
>     what
>     calling broken:
>     what
> 
> On Perl 5.20.0-threaded and 5.20.1-threaded, the output is:
>     calling not_broken:
>     what
>     calling broken:
>     Use of uninitialized value0 in print at boolean-broken.pl line 12.
> 
> Discovered and discussed in these Github issues:
> * https://github.com/Test-More/test-more/issues/546
> * https://github.com/ingydotnet/boolean-pm/issues/8
> 
> While the change in 5.21.5 to prevent unlocking read-only values
> (which is what boolean.pm relies on) prevents boolean.pm from doing
> this trick, that the problem appears unreleated to true/false values,
> only affects threaded 5.20.X and occurs in such a peculiar
> circumstance suggests that there might be a deeper bug worth
> investigating.
> 
> My perl -V output is attached.
I tried rigging the latest bleadperl locally to allow PL_sv_yes and PL_sv_no to be modified, and your example script worked.  Is it worth pursuing this to see when the problem emerged and when it disappeared?
-- 
Father Chrysostomos
---
via perlbug:  queue: perl5 status: open
https://rt.perl.org/Ticket/Display.html?id=123698

Thread Previous

• Re: [perl #123698] 5.20 threaded regression: boolean.pm -truth optionbreaks string literals in closures by David Golden
• [perl #123698] 5.20 threaded regression: boolean.pm -truth option breaks string literals in closures by James E Keenan via RT
• [perl #123698] 5.20 threaded regression: boolean.pm -truth option breaks string literals in closures by Father Chrysostomos via RT