Perl is regularly tested with valgrind by a number of folks. When doing this, Perl should be built and run in a special way, as described in perlhack. Determining whether a specific valgrind-reported leak is serious or not is something of a dark art and I'm not an expert (the experts tend to hang out on the P5P mailing list). This node: Valgrind, Perl, and banishing leaks and its links might be useful to you.

Perl Internals References

• PerlGuts Illustrated
• Extending and Embedding Perl by Tim Jenness and Simon Cozens

• Study the Perl source code itself
• Join the perl5porters@perl.org mailing list (see lists.perl.org)

perldoc and CPAN

• Internals and C Language Interface

• perlintern
• perlhack
• perlhacktut
• perlhacktips
• perlsource
• perlinterp
• perlguts
• perlapi
• perlcall

• perlxstut
• perlxs
• perlembed
• perlxstypemap

PM Nodes

• Hacking perl by robin (2005)
• Perl closures: internal details.(Updated) by BrowserUk (2012)
• Yet Another XS Tutorial by cdybedahl (2013)
• Why not perl have raw/native type by xiaoyafeng (2020)
• XS Error: Segfault with B::HooksAtRuntime by Ovid (2022)

• How to write a test for backticks threads hang by me (2006) - backticks hang, submitted a patch to P5P
• Re: Valgrind, Perl, and banishing leaks by me (2006) - valgrind, P5P and memory leaks
• Re^4: Replacing closures (to work around threads crash) by me (2004) - closure crash when using Perl threads
• Re: How is the default @INC constructed? by me (2004) - building a relocatable perl exe via long build path hack

• Re: Perl, C and deallocation responsibility by etj (2023) - mentions perlxstypemap

See Also

• Re: Perl Memory problem ... (Memory Tools References)

Updated: Many references added long after the original reply was made.