in reply to HTTP::Proxy filter for dishing out own files

Here's a snippet that will hopefully get you started: 

#!/usr/bin/perl

use HTTP::Proxy;
use HTTP::Proxy::HeaderFilter::simple;
use Digest::MD5 qw(md5_hex);

my $proxy = HTTP::Proxy->new( port => 3128 );

my $filter = HTTP::Proxy::HeaderFilter::simple->new(
    sub {
        my ($self, $headers, $request) = @_;

        my $uri = $request->uri();

        my $cache_path = md5_hex($uri);

        # (depending on the file system being used, you might want to
        # have the cache files spread across several directory levels;
        # you'd of course need to have stored them that way in the fir
+st place...)
        substr($cache_path, $_, 0) = "/" for (2,5,8);
        # prepend base path
        $cache_path = "/path/to/cache/$cache_path";

        # debug
        printf "URI        = %s\n", $uri;
        printf "host       = %s\n", $uri->host();
        printf "path       = %s\n", $uri->path();
        printf "query      = %s\n", $uri->query();
        printf "cache_path = %s\n", $cache_path;
        
        if (-s $cache_path) {  # is it in the cache?

            # this would of course be the content read from the cache 
+file...
            my $content = "<html><body>...yadda yadda...</body></html>
+";

            # create response
            my $res = HTTP::Response->new(200);
            $res->content_type('text/html');
            $res->content($content);

            # send back (short-circuit normal content fetching)
            $self->proxy()->response($res);
        }
    }
);

$proxy->push_filter( request => $filter );

$proxy->start;

[download]

For example, when fetching your PM node via that proxy, the debug output showing the request URI and cache path would look like: 

URI        = http://perlmonks.org/?node_id=799308
host       = perlmonks.org
path       = /
query      = node_id=799308
cache_path = /path/to/cache/0c/12/7f/026c96a135e45706194ba5b1f8

[download]

And if you had stored associated content under the cache path 0c/12/7f/026c96a135e45706194ba5b1f8, it would be served instead of the real remote content...

Replies are listed 'Best First'.
Re^2: HTTP::Proxy filter for dishing out own files
by r1n0 (Beadle) on Oct 06, 2009 at 15:09 UTC
    Thank you very much for the response! :-)

    I have a follow on question before I try what you sent. For the response code:
            if (-s $cache_path) {  # is it in the cache?

            # this would of course be the content read from the cache 
+file...
            my $content = "<html><body>...yadda yadda...</body></html>
+";

            # create response
            my $res = HTTP::Response->new(200);
            $res->content_type('text/html');
            $res->content($content);

            # send back (short-circuit normal content fetching)
            $self->proxy()->response($res);
        }

    [download]

    I would like to know how I can assign the contents of an entire file to the HTTP::Reponse. The response will have the necessary header fields in it if I can just throw the entire file back to the client.
    I might need to through my own response code, but that will be easy enough.
    I would prefer not to have to read in all of the original headers and assign to each component of HTTP::Response, if I can avoid it.

    Thanks in advance for your help.
[reply]
[d/l]

In Section Seekers of Perl Wisdom