[Oc-pm] Chain of subs...

[Wilson, Douglas]

> > > I'm writing a log parser for some things at work, and am 
> > > trying to figure
> > > out the best way to run each line of the log through a chain 
> > > of subroutines.
> > > 
> > > Is an array of sub references the best way to go for this?
> > 
> > That seems as good a way to go as any. Then you can easily
> > add or remove any of the parsing actions. But if you
> > are doing some of the same things in each subroutine (e.g.
> > parsing the line into tokens), then you might think about
> > doing some of the work before passing off the rest of the
> > work to the subs.
> 
> Oops, I meant to put that into my initial e-mail.
> 
> I'm parsing out the lines into a hash table first, and then 
> I'll pass a
> reference to the hash table to each of the functions.

I'm all for keeping it simple, but if you wanted to, you
could go OO and write a dispatching class something like:

package MyLogDispatcher;

sub new {
  my $class = shift;
  bless [@_], $class
}

sub process {
  my $self = shift;
  # tokenize @_ to %args or whatever
  $_->process(%args) for @$self;
}

package MyLogProcessor1;

sub process {
  my $self = shift;
  # process (@_)
}

package MyLogProcessor2;

sub process {
  my $self = shift;
  # process (@_)
}

package main;

my $processor = MyLogDispatcher->new(qw(
  MyLogProcessor1
  MyLogProcessor2
));

while (<>) {
  $processor->process($_);
}