PERL587DELTA(1)

πŸ’– perl587delta β€” the tea on what's new in perl v5.8.7 πŸ’–

okay bestie, welcome to perl587delta, straight out of the Perl Programmers Reference Guide (section 1, obviously). this is the glow-up recap. sit down, hydrate, let's go 🫢


✨ DESCRIPTION ✨

so basically? this doc is the receipts on the differences between the 5.8.6 release and the 5.8.7 release. that's it, that's the plot πŸ’…


πŸ’€ Incompatible Changes πŸ’€

ngl there are no changes incompatible with 5.8.6. zero. none. your code is safe, bestie, breathe 😌


πŸ‘‘ Core Enhancements πŸ‘‘

🌈 Unicode Character Database 4.1.0 🌈

the copy of the Unicode Character Database that ships inside Perl 5 has been updated, and it went from 4.0.1 all the way up to 4.1.0. if you wanna see the notable changes, they're over at http://www.unicode.org/versions/Unicode4.1.0/#NotableChanges. obsessed with an upgrade fr fr ✨

πŸ”’ suidperl less insecure πŸ”’

a pair of exploits in suidperl involving debugging code have been closed. bye πŸ‘‹

⚠️ for new projects, the core perl team strongly recommends that you use dedicated, single purpose security tools such as sudo instead of suidperl. like, please. for me.

πŸŽ€ Optional site customization script πŸŽ€

the perl interpreter can be built so it allows the use of a site customization script. by default this is not enabled, to stay consistent with previous perl releases (respecting the old ways, we love that). to use it, add -Dusesitecustomize to the command line flags when you run the Configure script. also go peek at -f in perlrun. iykyk 🧚

πŸ¦‹ Config.pm is now much smaller πŸ¦‹

Config.pm is now about 3K rather than 32K, because the infrequently used code and %Config values get loaded on demand. this is completely transparent to the programmer, but it means most code saves parsing and loading 29K of script (for example, code that uses File::Find). she lost weight and she's thriving πŸ’…


πŸ“¦ Modules and Pragmata πŸ“¦

glow-ups incoming, one per bullet, no skips, no cuts:

  • B upgraded to version 1.09
  • base upgraded to version 2.07
  • bignum upgraded to version 0.17
  • bytes upgraded to version 1.02
  • Carp upgraded to version 1.04
  • CGI upgraded to version 3.10
  • Class::ISA upgraded to version 0.33
  • Data::Dumper upgraded to version 2.121_02
  • DB_File upgraded to version 1.811
  • Devel::PPPort upgraded to version 3.06
  • Digest upgraded to version 1.10
  • Encode upgraded to version 2.10
  • FileCache upgraded to version 1.05
  • File::Path upgraded to version 1.07
  • File::Temp upgraded to version 0.16
  • IO::File upgraded to version 1.11
  • IO::Socket upgraded to version 1.28
  • Math::BigInt upgraded to version 1.77
  • Math::BigRat upgraded to version 0.15
  • overload upgraded to version 1.03
  • PathTools upgraded to version 3.05
  • Pod::HTML upgraded to version 1.0503
  • Pod::Perldoc upgraded to version 3.14
  • Pod::LaTeX upgraded to version 0.58
  • Pod::Parser upgraded to version 1.30
  • Symbol upgraded to version 1.06
  • Term::ANSIColor upgraded to version 1.09
  • Test::Harness upgraded to version 2.48
  • Test::Simple upgraded to version 0.54
  • Text::Wrap upgraded to version 2001.09293, and that one was to fix a bug when wrap() was called with a non-space separator πŸ›
  • threads::shared upgraded to version 0.93
  • Time::HiRes upgraded to version 1.66
  • Time::Local upgraded to version 1.11
  • Unicode::Normalize upgraded to version 0.32
  • utf8 upgraded to version 1.05
  • Win32 upgraded to version 0.24, which provides Win32::GetFileVersion

πŸ› οΈ Utility Changes πŸ› οΈ

πŸ” find2perl enhancements πŸ”

find2perl got itself some new options: -iname, -path and -ipath. three new flags, main character energy, periodt πŸ’…


⚑ Performance Enhancements ⚑

the internal pointer mapping hash used during ithreads cloning now uses an arena for memory allocation. in tests this reduced ithreads cloning time by about 10%. faster is faster, bestie πŸ”₯


🧰 Installation and Configuration Improvements 🧰

  • the Win32 dmake makefile.mk has been updated so it's compatible with the latest versions of dmake.
  • PERL_MALLOC, DEBUG_MSTATS, PERL_HASH_SEED_EXPLICIT and NO_HASH_SEED should now work in Win32 makefiles. they were being ignored and now they're not, we love a functioning env var 🫢

πŸ› Selected Bug Fixes πŸ›

okay this section is a whole redemption arc, buckle up:

  • the socket() function on Win32 has been fixed so that it is able to use transport providers which specify a protocol of 0 (meaning any protocol is allowed) once more. (this was broken in 5.8.6, and typically caused the use of ICMP sockets to fail.)
  • another obscure bug involving substr and UTF-8, caused by bad internal offset caching, has been identified and fixed. sneaky little thing, caught.
  • a bug involving the loading of UTF-8 tables by the regexp engine has been fixed β€” code such as "\x{100}" =~ /[[:print:]]/ will no longer give corrupt results.
  • case conversion operations such as uc on a long Unicode string could exhaust memory. this has been fixed. she was eating ALL the RAM, no more 😭
  • index/rindex were buggy for some combinations of Unicode and non-Unicode data. this has been fixed.
  • read (and presumably sysread) would expose the UTF-8 internals when reading from a byte oriented file handle into a UTF-8 scalar. this has been fixed.
  • several pack/unpack bug fixes, a whole mini-arc:
    • checksums with b or B formats were broken.
    • unpack checksums could overflow with the C format.
    • U0 and C0 are now scoped to () pack sub-templates.
    • counted length prefixes now don't change C0/U0 mode.
    • pack Z0 used to destroy the preceding character. (rude.)
    • P/p pack formats used to only recognise literal undef
  • using closures with ithreads could cause perl to crash. this was due to failure to correctly lock internal OP structures, and has been fixed.
  • the return value of close now correctly reflects any file errors that occur while flushing the handle's data, instead of just giving failure if the actual underlying file close operation failed.
  • not() || 1 used to segfault. not() now behaves like not(0), which was the pre 5.6.0 behaviour.
  • h2ph has various enhancements to cope with constructs in header files that used to result in incorrect or invalid output.

🚨 New or Changed Diagnostics 🚨

there's a new taint error: "%ENV is aliased to %s". this error gets thrown when taint checks are enabled and when *ENV has been aliased, so that %ENV has no env-magic anymore and hence the environment cannot be verified as taint-free. the vibes are unsafe, so perl says so πŸ’€

the internals of pack and unpack have been updated. all legitimate templates should work as before, but there may be some changes in the error reported for complex failure cases. any behaviour changes for non-error cases are bugs, and should be reported.


πŸ”§ Changed Internals πŸ”§

there has been a fair amount of refactoring of the C source code, partly to make it tidier and more maintainable. the resulting object code and the perl binary may well be smaller than 5.8.6, and hopefully faster in some cases, but apart from this there should be no user-detectable changes. (glow-up that only you can tell happened, iconic πŸ’…)

${^UTF8LOCALE} has been added to give perl space access to PL_utf8locale.

the size of the arenas used to allocate SV heads and most SV bodies can now be changed at compile time. the old size was 1008 bytes, the new default size is 4080 bytes.


😬 Known Problems 😬

Unicode strings returned from overloaded operators can be buggy. this is a long standing bug reported since 5.8.6 was released, but we do not yet have a suitable fix for it. still sitting there. still open. we're aware πŸ™ˆ


🌍 Platform Specific Problems 🌍

on UNICOS, lib/Math/BigInt/t/bigintc.t hangs burning CPU. ext/B/t/bytecode.t and ext/Socket/t/socketpair.t both fail tests. these are unlikely to be resolved, as our valiant UNICOS porter's last Cray is being decommissioned. rest in peace to a real one πŸ«‘πŸ’œ


πŸ“’ Reporting Bugs πŸ“’

if you find what you think is a bug, you might check the articles recently posted to the comp.lang.perl.misc newsgroup and the perl bug database at http://bugs.perl.org. there may also be information at http://www.perl.org, the Perl Home Page.

if you believe you have an unreported bug, please run the perlbug program included with your release. be sure to trim your bug down to a tiny but sufficient test case β€” the tiny test case is the whole assignment, bestie, do not skip it ✨ your bug report, along with the output of perl -V, will be sent off to perlbug@perl.org to be analysed by the Perl porting team. you can browse and search the Perl 5 bugs at http://bugs.perl.org/


πŸ‘€ SEE ALSO πŸ‘€

  • the Changes file, for exhaustive details on what changed.
  • the INSTALL file, for how to build Perl.
  • the README file, for general stuff.
  • the Artistic and Copying files, for copyright information.

and that's the whole thing, bestie β€” thanks for reading, love u, go write some perl πŸ’–

this lil doc came out of perl v5.42.3, dated 2026-08-03. slay, bye πŸŽ€πŸ¦‹βœ¨

πŸ“¦ where this page comes from