Escaping characters in the array

penguinfuz has asked for the wisdom of the Perl Monks concerning the following question:

While populating an array, (data used later for comparison), I have noticed some entries contain characters that break a regexp, a plus sign "+" for example.

I am questioning the best way to escape these characters. It seems easiest to continue pushing everything "as-is" into the array and do the escaping as needed when parsing through, but I am not sure.

Now for the obligatory coding example...

#! /usr/bin/perl -w
use strict;
my @current     = `rpm -qa`;
my @base        = `cat /tmp/install.log`;
my $cur_cnt     = @current;
my $base_cnt    = @base;

$cur_cnt        -= 1;
$base_cnt       -= 1;

foreach (@current) {
  # Reset to stay out of the negative when decrementing below.
  $base_cnt     = @base;
  # Set these because the first iteration through the
  # loop gripes about unitialized value.
  $cur_cnt      -= 1;
  $base_cnt     -= 1;
foreach (@base) {
    if ($base[$base_cnt] =~ $current[$cur_cnt]) {
      print "Unchanged: $base[$base_cnt]";
    } else {
      #Do Nothing
    }
    $base_cnt -= 1;
  }
}
[download]

This iterates through (while begging for optimization) a listing of installed RPMs, and matches against a listing of RPMs installed at build time, Any deviation, either version update or new software package is printed to the screen.

The trouble arises when the loop hits a package name with special chars, for example: timidity++-2.11.

Comment on Escaping characters in the array Download Code

Replies are listed 'Best First'.
Re: Escaping characters in the array by tadman (Prior) on May 30, 2002 at 21:45 UTC
While particle mentions quotemeta, you can also use the in-line version of same. The real problem seems to be the way you are using your arrays. You're not indexing correctly. Looks like some C-style looping you've got there. How about this? `foreach my $cur_cnt (0..$#current) { foreach my $base_cnt (0..$#base) { if ($base[$base_cnt] =~ /^\Q$current[$cur_cnt]\E/) { # ... } } }` [download] Better still to just iterate and forget the index unless you actually need it: `foreach my $current (@current) { foreach my $base (@base) { if ($base =~ /^\Q$current\E/) { # ... } } }` [download] Less is more, especially with programming. As an explanation, the \Q starts "escaping" special characters, the \E stops, using a quotemeta equivalent.	[reply] [d/l] [select]
Re: Escaping characters in the array by particle (Vicar) on May 30, 2002 at 21:28 UTC
see quotemeta ~Particle accelerates	[reply]
Re: Escaping characters in the array by penguinfuz (Pilgrim) on May 31, 2002 at 13:55 UTC
particle wrote: see quotemeta Cheer Particle. quotemeta is a great tool to add to my growing arsenal. ;) After quoting the meta characters, I found my logic was all wrong so I started rewriting. tadman wrote: The real problem seems to be the way you are using your arrays. You're not indexing correctly.... Better still to just iterate and forget the index unless you actually need it... As an explanation, the \Q starts "escaping" special characters, the \E stops... Very thought provoking stuff, especially the bit about indexing correctly. I am still not positive about what would exactly be the correct way to index with this issue, but I took your advice and did away with it, actually, the code is quite different now, but it _is_ doing what I want. ;) Here's a snippet my @new = `cat /path/to/some/file`; foreach my $new (@new) { my $args = `rpm -qa \| grep $new`; print $new if ($args =~ m/^$/); } [download] The actual application for this code has changed focus since my first posting, This time the definition is truelly what I want to do, where as the last description was simply a secondary function. I take a listing of "new" RPMs and compare them against the installed RPMs. If there is no match, I will assume (at the moment) the package should be downloaded and installed. What I am working towards now is: Sanity checking in relation to packages versions. Older version "new" packages should not be installed Using a hash for the values in @new. This is because when I query an RPM package, I cannot query with the extension in tack, otherwise nothing will match. For example: Search for: glibc-2.2.5-34 Download: glibc-2.2.5-34.i386.rpm Ok, enough talking back to writing. ;) BTW: ++ all around guys for the great ideas!	[reply] [d/l]