Thats a good point. Right now it is configured to use only one config file, but it wouldn't be too difficult to add in some added functionality that could account for other config files based on time or something like that. I'm just trying to keep this reasonably simple but flexible enough to be changed in the future as unforeseen needs come up.

If you want any level of maintainability or reusability - go with a config file.

For example if your script will end up running in more than one environment, you'll probably want to scan a different set of directories in each environment. And you don't want to be maintaining multiple versions of the script for that.

Also, if you're working on a team with multiple developers, and use a version control product, separating into 2 parts will improve the granularity.

• the last modified time of the perl source will acurately show the last time you mucked with it's inner workings
• some of the modules can automatically re-read the file if you ever daemonize the thing

Another strategy is to detect the existance of the config file, and if it doesn't exist, use the contents of __DATA__ to write your default config file. That could be used to create the basic config file in such a way that it is easy to text-edit (and thus to modify) by the person maintaining your scripts installation. The following steps would be taken:

• Check for config file.
• If file exists, open and use it.
• If file doesn't exist, read __DATA__ and write out a config file.
• Use the newly written config file.

This will leave the config file as an artifact of the script's first run. The user can then customize it without having to sift through your source code, possibly ruining something important. ;)

muntfish and Solo both make excellent points. To add to what Solo said re: daemonizing your script. It is important to note that once the DATA handle is read, it cannot (easily) be re-read (seeking to the beginging of the file puts you at the top of your source, not the top of the DATA handle). So if you were to convert it to a daemon, you would need to watch out for this. But then if you daemonize it, you would be better off with an external config anyway, that way it your config can change without restarting your daemon.

It is important to note that once the DATA handle is read, it cannot (easily) be re-read (seeking to the beginging of the file puts you at the top of your source, not the top of the DATA handle).

Erm, no it's very easy to do this.

#!/usr/bin/perl
use Fcntl qw( :seek );

my $top_o_data = tell( DATA );

for( 1 .. 2 ) {
  print "read $_\n---\n";
  seek( DATA, $top_o_data, SEEK_SET );
  print while <DATA>;        
}

__DATA__
Wubba.
Zoikes.
Jinkies.
[download]

Update: The problem of course is that you may not see updates (at least not on OS X or Linux that I've checked).