I was recently reading through Perl Regular Expression Tutorial (perlretut) when I found the section on using the ‘r’ option. I’d seen this before but rarely had any reason to use it. Then it occurred to me that you could chain these together.

I’ve been programming Perl for 30+ years and don’t think I’ve ever seen this before.

Here’s an example that shows how I would usually do this, i.e. a series of var =~ s/// lines, and how to do the same thing in one line while initializing a variable.

#!/usr/bin/env perl
use strict;
use warnings;
use feature 'say';

my $T = "one two trash THREE random";
my $T2 = $T;
$T2 =~ s/trash|random//g;
$T2 =~ s/(THREE)/lc($1)/e;
$T2 =~ s/\s+/ /g;
$T2 =~ s/^/Count with me: /;

my $T3 = $T =~ s/trash|random//gr =~ s/(THREE)/lc($1)/er =~ s/\s+/ /gr =~ s/^/Count with me: /r;

say "T  := $T\nT2 := $T2\nT3 := $T3";

Output:

$ ./perl-subst-test.pl 
T  := one two trash THREE random
T2 := Count with me: one two three 
T3 := Count with me: one two three 

Anyways, I had to tell someone …

  • @zikal@programming.dev
    link
    fedilink
    English
    31 year ago

    I really like how easy it is to use regex in Perl with =~. I’ve always wondered why that’s not the case in Go and Rust where I need to compile the regex using non-std library.

    • umbraklatOP
      link
      English
      21 year ago

      Agreed. Though I’m very comfortable using RE, I’ve still never used it with backtracking or recursion. I always feel like I can go much deeper in Perl, though 99% of what I need to do doesn’t require it.