Perl::...Capture(3pm)

πŸ’– NAME πŸ’–

Okay bestie, welcome to the Perl::...Capture(3pm) page in User Contributed Perl Documentation. This policy is literally called Perl::Critic::Policy::RegularExpressions::ProhibitUnusedCapture, and the whole vibe is: only use a capturing group if you actually plan to use the captured value. Periodt. πŸ’…


πŸ¦‹ AFFILIATION πŸ¦‹

This policy is part of the core Perl::Critic distribution. She’s core, no cap. ✨


🌸 DESCRIPTION 🌸

Perl regular expressions have multiple types of grouping syntax. The basic parentheses β€” like m/(foo)/ β€” capture into the magic variable $1. Non-capturing groups β€” like m/(?:foo)/ β€” are useful because they have better runtime performance and do not copy strings to the magic global capture variables. Slay for efficiency. πŸŽ€

It’s also easier on the maintenance programmer if you consistently use capturing vs. non-capturing groups, because that programmer can tell more easily which regexps can be refactored without breaking surrounding code which may use the captured values. It’s giving considerate teammate energy. 🫢


βš™οΈ CONFIGURATION βš™οΈ

This Policy is not configurable except for the standard options. That’s it, that’s the config. πŸ’€


😭 CAVEATS 😭

πŸŒ€ qr// interpolation πŸŒ€

This policy can be confused by interpolation of qr// elements, but those are always false negatives. For example:

my $foo_re = qr/(foo)/;
my ($foo) = m/$foo_re (bar)/x;

A human can tell that this should be a violation because there are two captures but only the first capture is used, not the second. The policy only notices that there is one capture in the regexp and remains happy. Not her being a little oblivious. πŸ™ˆ

🌈 @-, @+, $LAST_MATCH_START and $LAST_MATCH_END 🌈

This policy will only recognize capture groups referred to by these variables if the use is subscripted by a literal integer. Literal integer or it doesn’t count, bestie. πŸ”₯

🧚 $^N and $LAST_SUBMATCH_RESULT 🧚

This policy will not recognize capture groups referred to only by these variables, because there is in general no way by static analysis to determine which capture group is referred to. For example,

m/ (?: (A[[:alpha:]]+) | (N\d+) ) (?{$foo=$^N}) /smx

makes use of the first capture group if it matches, or the second capture group if the first does not match but the second does. Static analysis is just sitting there like πŸ‘‘ β€œI cannot read your mind, ma’am.”

πŸ“ split() πŸ“

Normally, this policy thinks that if a capture is used at all it must be used before the next regular expression in the same scope. The regular expression in a split() needs to be exempted because it does not affect the caller’s capture variables. Important distinction, fr fr. πŸ’œ

At present, this policy recognizes and exempts the regular expressions in

split /.../, ...

and

split( /.../, ... )

but more exotic syntax may produce false positives. So like, mostly iconic behavior, but not perfect. 😭


πŸ‘‘ CREDITS πŸ‘‘

Initial development of this policy was supported by a grant from the Perl Foundation. We love funding open source, literally iconic. 🌟


πŸ’Œ AUTHOR πŸ’Œ

Chris Dolan <cdolan@cpan.org>


Copyright (c) 2007-2023 Chris Dolan. Many rights reserved.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. The full text of this license can be found in the LICENSE file included with this module.


Okay bye bestie, signing off from Perl::...Capture(3pm) β€” this page is from perl v5.42.0, dated 2024-10-23, and yes she ate. πŸ’…βœ¨πŸŽ€

πŸ“¦ where this page comes from