<Container>
<Doc>
  <content>stuff</content>
</Doc>
<Doc>
  <content>different stuff</content>
</Doc>
...
</Container>
[download]

  All of the examples for XML::SAX::ByRecord that I've looked at showed how to write *filters* that process as I've described. None that I have seen (probably a problem with my eyesight, not the documentation) have explained how to work with each of these "Documents" in the stream as its own isolated chunk of content so that one can process it independently of the filter.

Below is my the code that I've come up with to handle the task that I've explained above. I can't help thinking that it's a bit of a kludge. What is a more elegant way to deal with this type of stream processing?

Thanks in advance,

  dug

#!/usr/bin/perl

use warnings;
use strict;
$|++;

use XML::SAX::Machines qw( :all );

my $output_handle; # global stream output container

##
# callback for end_document event.
my $write_hook = sub {
  my $self = shift;
  my $current_doc = $output_handle; # get contents of output buffer
  $output_handle = '';              # clear buffer for next doc
  ## process current doc
  process_doc( $current_doc );
};

my $filter = EndDocumentAction->new(end_hook => $write_hook);

my $machine = Pipeline(
  ByRecord( $filter ),
  \$output_handle,
);

$machine->parse_file( \*DATA );

sub process_doc {
  my $content = shift;
  # do something interesting
  print $content, "\n";
}

package EndDocumentAction;
use base qw( XML::SAX::Base );

sub new {
  my ($class, %args) = @_;
  my $self = {};
  $self->{End_Hook}    = $args{end_hook}; # install callback for end_d
+ocument
  $self->{start_counter} = 0;
  bless $self, $class;
  return $self;
}

sub end_document {
  my $self = shift;
  my $callback = $self->{End_Hook};
  $self->$callback();
}

1;

__END__
<Stream>
<Doc>
<foo>hey man</foo>
</Doc>
<Doc>
<bar>hey man, how's it goin'?</bar>
</Doc>
<Doc>
<baz>pretty right on.</baz>
</Doc>
</Stream>
[download]