NAME
Perl::Critic::Policy::RegularExpressions::ProhibitRegexToSplitAtDelimiter - Find the first literal separator with index and substr, not a regex.
VERSION
version 0.001
Perl::Critic::Policy::RegularExpressions::ProhibitRegexToSplitAtDelimiter
Using regexes to do simple text manipulation in a loop is a common cause of performance problems. This can be prevented from happening by refraining from doing so everywhere. One pattern in that class is splitting at a delimiter without using split. This policy prevents doing just that with regexes.
Do index + substr to pick out what you are looking for, or split when you expect many matches, and can afford loading the entire string.
# Reading buffer chunks till we get a full line to process w/ take()
while ( $buffer =~ s/\A([^\n]*)\n// ) { take($1) } # reported
while ( ( my $at = index $buffer, "\n" ) >= 0 ) { # OK!
take( substr $buffer, 0, $at );
substr $buffer, 0, $at + 1, q{};
}
# Parsing a key from a key=value config file
my ($key) = $line =~ m/\A(.*?)=/s; # reported
my $at = index $line, '='; # OK!
my $key = $at >= 0 ? substr( $line, 0, $at ) : undef;
my ($key,$value) = split(/=/, $line, 2); # OK!The substitution takes the part before the separator off the string, together with the separator. Four-argument substr with an empty replacement does that.
Without a capture, the substitution only takes that much off, and the match only asks whether the separator is there at all:
$buffer =~ s/\A[^\n]*\n//; # reported
my $at = index $buffer, "\n";
substr $buffer, 0, $at + 1, q{} if $at >= 0;
next unless $line =~ m/\A[^=]*=/; # reported
next unless index( $line, '=' ) >= 0;Under /r the substitution leaves the string alone and returns what follows the separator, which is substr( $buffer, $at + 1 ).
PROHIBITED
s/\A([^\n]*)\n// # the negated class of the separator
m/\A([^=]+)=/ # one or more, see CAVEATS
m/^([^:]*):/ # ^ without /m is the start of the string
m/\A([^:]*):/m # /m changes ^ and $, not \A
m/\A(.*?)--/s # a lazy dot under /s, before a string
m/\A(.+?), /s # one or more, see CAVEATS
m/\A(.*?)\n/ # without /s, . is anything but a newline
m/\A(.*)\n/ # so before a newline it can be greedy too
m/\A(\N*)\n/ # and so can \N
s/\A([^\n]*)\n//r # /r: a copy
m/\A(?<key>[^:]*):/ # a named capture
s/\A([^\n]*)\x0a// # any spelling of the character
s/\A([^\n]*)[\n]// # a class of one character is that character
m/\A([^.]*)\./ # an escaped metacharacter is a character
m/\A(.*?)\Q-.\E/s # and so is anything between \Q and \E
m/\A ( [^:]* ) : /x # whitespace under /x is not text
s/\A[^\n]*\n// # no capture: the first line off
s/\A.*?--//s # and every other form above, so
m/\A[^:]*:/ # a match without a capture is index >= 0ALLOWED
m/\A(.*?):/ # without /s, . refuses a newline before the :
m/\A(.*):/s # greedy: the last separator, which is rindex
m/^([^:]*):/m # /m: ^ is the start of any line
m/([^:]*):/ # unanchored
m/\G([^:]*):/gc # \G is where the last match stopped
m/\A([^:]*):/g # /g
s/\A([^\n]*)\n//g # /g
m/\A([^x]*)x/i # /i: index compares case
m/\A([^:;]*)[:;]/ # a class of several characters
m/\A([^a]*)ab/ # the first a, which must be followed by b
m/\A(.*?)\s/s # \s, \R and the like are not literal text
m/\A(.*?)(?:,|;)/s # alternation
m/\A(.*?)$sep/s # interpolation
s/\A([^\n]*)\n+// # a quantified separator
m/\A([^:]{2,}):/ # a counted quantifier
m/\A(\w*):/ # a capture that checks what it captures
m/\A([^:]*):(.*)\z/s # two captures
s/\A([^:]*):\s*// # more pattern after the separator
m/\A([^\n]*)\n\z/ # an anchor at the end as well
s/\A([^\n]*)\n/$1 / # a replacement
qr/\A([^\n]*)\n/ # a compiled regex, which may be used elsewhere
s/\A(?:[^\n]*)\n// # a group that captures nothingCAVEATS
[^X]+ and .+? refuse a string that starts with the separator, and index finds it at position 0, with or without a capture. When that check is the point, keep the regex and say ## no critic (ProhibitRegexToSplitAtDelimiter) and why, or test $at > 0 instead of $at >= 0.
A match without a capture also sets $&, @- and @+. When the code goes on to read $+[0], that is index plus the length of the separator.
index returns -1 when the separator is not there. Test the position before you take substr of it: substr( $line, 0, -1 ) is the whole string less its last character, not an error.
Modifiers turned on by use re in the lexical scope of the regex count as if they were written on it. So under use re '/s', m/\A(.*?):/ is reported, and under use re '/m', m/^([^:]*):/ is not.
METHODS
supported_parameters
default_severity
default_themes
applies_to
violates
BUGS
Please report any bugs or feature requests on the bugtracker website https://github.com/teodesian/perl-critic-policy-prohibitregextosplitatdelimiter/issues
When submitting a bug or request, please include a test-file or a patch to an existing test-file that illustrates the bug or desired feature.
AUTHORS
Current Maintainers:
George S. Baugh <george@troglodyne.net>
CONTRIBUTOR
troglodyne-bot <morpheus@troglodyne.net>
COPYRIGHT AND LICENSE
Copyright (c) 2026 Troglodyne LLC
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.