http://qs321.pair.com?node_id=281668
Category: Text Processing
Author/Contact Info Adam Monsen <adamm@wazamatta.com>
Description: Edit files in-place using a single regex (regular expression) that will be applied to the entire file as a string. Useful for applying multiline regexes and avoiding shell redirection and backup files. Also, quite dangerous!

Call like so:  ./edit.pl <regex> <files>...

example:

./edit.pl 's/your([\n\s]*?)friend/joe${1}${2}/ms' foo.txt

Please let me know if there is are easier/more elegant ways to do this, especially as a one-liner. When making suggestions, please keep in mind that my goals were to provide:

  • complete power of search and replace with Perl regexes
  • in-place file editing on multiple files
This script is mainly for Perl-savvy users or those familiar with Perl-compatible regular expressions.
#!/usr/bin/perl -w
use strict;                                                           
+           
use Getopt::Declare;

# $Id: multiline_inplace_edit.pl,v 1.5 2003/08/07 00:44:02 adamm Exp $

our $VERSION = 0.01;

our $usage_spec = <<'UsageSpec';
  [strict]

  Edit files in-place using a single regex that will be applied to
  the entire file as a string.

  <regex>               regex to be applied to files
                          [required]
  <files>...            one or more files to slurp in and run regex ag
+ainst
                          [required]

  WARNING: the first argument is added to an eval block!

  Example:
    ./edit.pl 's/your([\n\s]*?)friend/joe${1}${2}/ms' foo.txt
UsageSpec

our $args = new Getopt::Declare ($usage_spec) or die;

local $/;

for (@{$args->{'<files>'}}) {
  open (FH, "< $_") or die "error opening $_ for reading: $!";
  my $whole_file = <FH>;
  my $matching_code = '{ $whole_file =~ ' . $args->{'<regex>'} . ' }';
  eval $matching_code;
  die $@ if $@;
  close(FH) or die "error closing $!";

  open (FH, "> $_") or die "error opening $_ for writing: $!";
  print FH $whole_file;
  close(FH) or warn "error closing $!";
}