test/html-webhacc/cc.cgi

#!/usr/bin/perl
use strict;
use utf8;

use lib qw[/home/httpd/html/www/markup/html/whatpm
           /home/wakaba/work/manakai2/lib];
use CGI::Carp qw[fatalsToBrowser];
use Scalar::Util qw[refaddr];

  require WebHACC::Input;
  require WebHACC::Result;
  require WebHACC::Output;

my $out;

  require Message::DOM::DOMImplementation;
  my $dom = Message::DOM::DOMImplementation->new;
{
  use Message::CGI::HTTP;
  my $http = Message::CGI::HTTP->new;

  if ($http->get_meta_variable ('PATH_INFO') ne '/') {
    print STDOUT "Status: 404 Not Found\nContent-Type: text/plain; charset=us-ascii\n\n400";
    exit;
  }
  
  load_text_catalog ('en'); ## TODO: conneg

  $out = WebHACC::Output->new;
  $out->handle (*STDOUT);
  $out->set_utf8;
  $out->set_flush;
  $out->html (qq[Content-Type: text/html; charset=utf-8

<!DOCTYPE html>
<html lang="en">
<head>
<title>Web Document Conformance Checker (BETA)</title>
<link rel="stylesheet" href="../cc-style.css" type="text/css">
</head>
<body>
<h1><a href="../cc-interface">Web Document Conformance Checker</a> 
(<em>beta</em>)</h1>
]);

  my $input = get_input_document ($http, $dom);

  $out->input ($input);
  $out->unset_flush;

  my $result = WebHACC::Result->new;
  $result->output ($out);
  $result->{conforming_min} = 1;
  $result->{conforming_max} = 1;

  $out->html ('<script src="../cc-script.js"></script>');

  check_and_print ($input => $result => $out);
  
  $result->generate_result_section;

  $out->nav_list;

  exit;
}

sub check_and_print ($$$) {
  my ($input, $result, $out) = @_;
  my $original_input = $out->input;
  $out->input ($input);

  $input->generate_info_section ($result);

  $input->generate_transfer_sections ($result);

  unless (defined $input->{s}) {
    $result->{conforming_min} = 0;
    return;
  }

  my $checker_class = {
    'text/cache-manifest' => 'WebHACC::Language::CacheManifest',
    'text/css' => 'WebHACC::Language::CSS',
    'text/html' => 'WebHACC::Language::HTML',
    'text/x-webidl' => 'WebHACC::Language::WebIDL',

    'text/xml' => 'WebHACC::Language::XML',
    'application/atom+xml' => 'WebHACC::Language::XML',
    'application/rss+xml' => 'WebHACC::Language::XML',
    'image/svg+xml' => 'WebHACC::Language::XML',
    'application/xhtml+xml' => 'WebHACC::Language::XML',
    'application/xml' => 'WebHACC::Language::XML',
    ## TODO: Should we make all XML MIME Types fall
    ## into this category?

    ## NOTE: This type has different model from normal XML types.
    'application/rdf+xml' => 'WebHACC::Language::XML',
  }->{$input->{media_type}} || 'WebHACC::Language::Default';

  eval qq{ require $checker_class } or die "$0: Loading $checker_class: $@";
  my $checker = $checker_class->new;
  $checker->input ($input);
  $checker->output ($out);
  $checker->result ($result);

  ## TODO: A cache manifest MUST be text/cache-manifest
  ## TODO: WebIDL media type "text/x-webidl"

  $checker->generate_syntax_error_section;
  $checker->generate_source_string_section;

  my @subdoc;
  $checker->onsubdoc (sub {
    push @subdoc, shift;
  });

  $checker->generate_structure_dump_section;
  $checker->generate_structure_error_section;
  $checker->generate_additional_sections;

=pod

  if (defined $doc or defined $el) {

    print_listing_section ({
      id => 'identifiers', label => 'IDs', heading => 'Identifiers',
    }, $input, $elements->{id}) if keys %{$elements->{id}};
    print_listing_section ({
      id => 'terms', label => 'Terms', heading => 'Terms',
    }, $input, $elements->{term}) if keys %{$elements->{term}};
    print_listing_section ({
      id => 'classes', label => 'Classes', heading => 'Classes',
    }, $input, $elements->{class}) if keys %{$elements->{class}};
   
    print_rdf_section ($input, $elements->{rdf}) if @{$elements->{rdf}};
  }

=cut

  my $id_prefix = 0;
  for my $_subinput (@subdoc) {
    my $subinput = WebHACC::Input::Subdocument->new (++$id_prefix);
    $subinput->{$_} = $_subinput->{$_} for keys %$_subinput;
    $subinput->{base_uri} = $subinput->{container_node}->base_uri
        unless defined $subinput->{base_uri};
    $subinput->{parent_input} = $input;

    $subinput->start_section ($result);
    check_and_print ($subinput => $result => $out);
    $subinput->end_section ($result);
  }

  $out->input ($original_input);
} # check_and_print


{
  my $Msg = {};

sub load_text_catalog ($) {
#  my $self = shift;
  my $lang = shift; # MUST be a canonical lang name
  open my $file, '<:utf8', "cc-msg.$lang.txt"
      or die "$0: cc-msg.$lang.txt: $!";
  while (<$file>) {
    if (s/^([^;]+);([^;]*);//) {
      my ($type, $cls, $msg) = ($1, $2, $_);
      $msg =~ tr/\x0D\x0A//d;
      $Msg->{$type} = [$cls, $msg];
    }
  }
} # load_text_catalog

sub get_text ($;$$) {
#  my $self = shift;
  my ($type, $level, $node) = @_;
  $type = $level . ':' . $type if defined $level;
  $level = 'm' unless defined $level;
  my @arg;
  {
    if (defined $Msg->{$type}) {
      my $msg = $Msg->{$type}->[1];
      $msg =~ s{<var>\$([0-9]+)</var>}{
        defined $arg[$1] ? ($arg[$1]) : '(undef)';
      }ge;                 ##BUG: ^ must be escaped
      $msg =~ s{<var>{\@([A-Za-z0-9:_.-]+)}</var>}{
        UNIVERSAL::can ($node, 'get_attribute_ns')
            ?  ($node->get_attribute_ns (undef, $1)) : ''
      }ge; ## BUG: ^ must be escaped
      $msg =~ s{<var>{\@}</var>}{        ## BUG: v must be escaped
        UNIVERSAL::can ($node, 'value') ? ($node->value) : ''
      }ge;
      $msg =~ s{<var>{local-name}</var>}{
        UNIVERSAL::can ($node, 'manakai_local_name')
          ? ($node->manakai_local_name) : ''
      }ge;  ## BUG: ^ must be escaped
      $msg =~ s{<var>{element-local-name}</var>}{
        (UNIVERSAL::can ($node, 'owner_element') and
         $node->owner_element)
          ?  ($node->owner_element->manakai_local_name)
          : '' ## BUG: ^ must be escaped
      }ge;
      return ($type, 'level-' . $level . ' ' . $Msg->{$type}->[0], $msg);
    } elsif ($type =~ s/:([^:]*)$//) {
      unshift @arg, $1;
      redo;
    }
  }
  return ($type, 'level-'.$level, ($_[0]));
                                 ## BUG: ^ must be escaped
} # get_text

}

sub get_input_document ($$) {
  my ($http, $dom) = @_;

  require Encode;
  my $request_uri = Encode::decode ('utf-8', $http->get_parameter ('uri'));
  my $r = WebHACC::Input->new;
  if (defined $request_uri and length $request_uri) {
    my $uri = $dom->create_uri_reference ($request_uri);
    unless ({
             http => 1,
            }->{lc $uri->uri_scheme}) {
      $r = WebHACC::Input::Error->new;
      $r->{uri} = $request_uri;
      $r->{request_uri} = $request_uri;
      $r->{error_status_text} = 'URL scheme not allowed';
    }

    require Message::Util::HostPermit;
    my $host_permit = new Message::Util::HostPermit;
    $host_permit->add_rule (<<EOH);
Allow host=suika port=80
Deny host=suika
Allow host=suika.fam.cx port=80
Deny host=suika.fam.cx
Deny host=localhost
Deny host=*.localdomain
Deny ipv4=0.0.0.0/8
Deny ipv4=10.0.0.0/8
Deny ipv4=127.0.0.0/8
Deny ipv4=169.254.0.0/16
Deny ipv4=172.0.0.0/11
Deny ipv4=192.0.2.0/24
Deny ipv4=192.88.99.0/24
Deny ipv4=192.168.0.0/16
Deny ipv4=198.18.0.0/15
Deny ipv4=224.0.0.0/4
Deny ipv4=255.255.255.255/32
Deny ipv6=0::0/0
Allow host=*
EOH
    unless ($host_permit->check ($uri->uri_host, $uri->uri_port || 80)) {
      my $r = WebHACC::Input::Error->new;
      $r->{uri} = $request_uri;
      $r->{request_uri} = $request_uri;
      $r->{error_status_text} = 'Connection to the host is forbidden';
      return $r;
    }

    require LWP::UserAgent;
    my $ua = WDCC::LWPUA->new;
    $ua->{wdcc_dom} = $dom;
    $ua->{wdcc_host_permit} = $host_permit;
    $ua->agent ('Mozilla'); ## TODO: for now.
    $ua->parse_head (0);
    $ua->protocols_allowed ([qw/http/]);
    $ua->max_size (1000_000);
    my $req = HTTP::Request->new (GET => $request_uri);
    $req->header ('Accept-Encoding' => 'identity, *; q=0');
    my $res = $ua->request ($req);
    ## TODO: 401 sets |is_success| true.
    if ($res->is_success or $http->get_parameter ('error-page')) {
      $r->{base_uri} = $res->base; ## NOTE: It does check |Content-Base|, |Content-Location|, and <base>. ## TODO: Use our own code!
      $r->{uri} = $res->request->uri;
      $r->{request_uri} = $request_uri;

      ## TODO: More strict parsing...
      my $ct = $res->header ('Content-Type');
      if (defined $ct and $ct =~ /;\s*charset\s*=\s*"?([^\s;"]+)"?/i) {
        $r->{charset} = lc $1;
        $r->{charset} =~ tr/\\//d;
        $r->{official_charset} = $r->{charset};
      }

      my $input_charset = $http->get_parameter ('charset');
      if (defined $input_charset and length $input_charset) {
        $r->{charset_overridden}
            = (not defined $r->{charset} or $r->{charset} ne $input_charset);
        $r->{charset} = $input_charset;
      }

      ## TODO: Support for HTTP Content-Encoding

      $r->{s} = ''.$res->content;

      require Whatpm::ContentType;
      ($r->{official_type}, $r->{media_type})
          = Whatpm::ContentType->get_sniffed_type
              (get_file_head => sub {
                 return substr $r->{s}, 0, shift;
               },
               http_content_type_byte => $ct,
               has_http_content_encoding =>
                   defined $res->header ('Content-Encoding'),
               supported_image_types => {});
    } else {
      $r->{uri} = $res->request->uri;
      $r->{request_uri} = $request_uri;
      $r->{error_status_text} = $res->status_line;
    }

    $r->{header_field} = [];
    $res->scan (sub {
      push @{$r->{header_field}}, [$_[0], $_[1]];
    });
    $r->{header_status_code} = $res->code;
    $r->{header_status_text} = $res->message;
  } else {
    $r->{s} = ''.$http->get_parameter ('s');
    $r->{uri} = q<thismessage:/>;
    $r->{request_uri} = q<thismessage:/>;
    $r->{base_uri} = q<thismessage:/>;
    $r->{charset} = ''.$http->get_parameter ('_charset_');
    $r->{charset} =~ s/\s+//g;
    $r->{charset} = 'utf-8' if $r->{charset} eq '';
    $r->{official_charset} = $r->{charset};
    $r->{header_field} = [];

    require Whatpm::ContentType;
    ($r->{official_type}, $r->{media_type})
        = Whatpm::ContentType->get_sniffed_type
            (get_file_head => sub {
               return substr $r->{s}, 0, shift;
             },
             http_content_type_byte => undef,
             has_http_content_encoding => 0,
             supported_image_types => {});
  }

  my $input_format = $http->get_parameter ('i');
  if (defined $input_format and length $input_format) {
    $r->{media_type_overridden}
        = (not defined $r->{media_type} or $input_format ne $r->{media_type});
    $r->{media_type} = $input_format;
  }
  if (defined $r->{s} and not defined $r->{media_type}) {
    $r->{media_type} = 'text/html';
    $r->{media_type_overridden} = 1;
  }

  if ($r->{media_type} eq 'text/xml') {
    unless (defined $r->{charset}) {
      $r->{charset} = 'us-ascii';
      $r->{official_charset} = $r->{charset};
    } elsif ($r->{charset_overridden} and $r->{charset} eq 'us-ascii') {
      $r->{charset_overridden} = 0;
    }
  }

  if (length $r->{s} > 1000_000) {
    $r->{error_status_text} = 'Entity-body too large';
    delete $r->{s};
    return $r;
  }

  $r->{inner_html_element} = $http->get_parameter ('e');

  return $r;
} # get_input_document

package WDCC::LWPUA;
BEGIN { push our @ISA, 'LWP::UserAgent'; }

sub redirect_ok {
  my $ua = shift;
  unless ($ua->SUPER::redirect_ok (@_)) {
    return 0;
  }

  my $uris = $_[1]->header ('Location');
  return 0 unless $uris;
  my $uri = $ua->{wdcc_dom}->create_uri_reference ($uris);
  unless ({
           http => 1,
          }->{lc $uri->uri_scheme}) {
    return 0;
  }
  unless ($ua->{wdcc_host_permit}->check ($uri->uri_host, $uri->uri_port || 80)) {
    return 0;
  }
  return 1;
} # redirect_ok

=head1 AUTHOR

Wakaba <w@suika.fam.cx>.

=head1 LICENSE

Copyright 2007-2008 Wakaba <w@suika.fam.cx>

This library is free software; you can redistribute it
and/or modify it under the same terms as Perl itself.

=cut

## $Date: 2008/07/21 05:24:32 $
1	wakaba	1.1	#!/usr/bin/perl
2			use strict;
3	wakaba	1.23	use utf8;
4	wakaba	1.1
5			use lib qw[/home/httpd/html/www/markup/html/whatpm
6	wakaba	1.16	/home/wakaba/work/manakai2/lib];
7	wakaba	1.1	use CGI::Carp qw[fatalsToBrowser];
8	wakaba	1.2	use Scalar::Util qw[refaddr];
9	wakaba	1.1
10	wakaba	1.53	require WebHACC::Input;
11			require WebHACC::Result;
12			require WebHACC::Output;
13
14			my $out;
15	wakaba	1.2
16	wakaba	1.35	require Message::DOM::DOMImplementation;
17			my $dom = Message::DOM::DOMImplementation->new;
18			{
19	wakaba	1.16	use Message::CGI::HTTP;
20			my $http = Message::CGI::HTTP->new;
21	wakaba	1.1
22	wakaba	1.16	if ($http->get_meta_variable ('PATH_INFO') ne '/') {
23	wakaba	1.8	print STDOUT "Status: 404 Not Found\nContent-Type: text/plain; charset=us-ascii\n\n400";
24			exit;
25			}
26	wakaba	1.53
27	wakaba	1.7	load_text_catalog ('en'); ## TODO: conneg
28
29	wakaba	1.53	$out = WebHACC::Output->new;
30			$out->handle (*STDOUT);
31			$out->set_utf8;
32			$out->set_flush;
33			$out->html (qq[Content-Type: text/html; charset=utf-8
34	wakaba	1.2
35			<!DOCTYPE html>
36			<html lang="en">
37			<head>
38			<title>Web Document Conformance Checker (BETA)</title>
39	wakaba	1.3	<link rel="stylesheet" href="../cc-style.css" type="text/css">
40	wakaba	1.2	</head>
41			<body>
42	wakaba	1.13	<h1><a href="../cc-interface">Web Document Conformance Checker</a>
43			(<em>beta</em>)</h1>
44	wakaba	1.53	]);
45	wakaba	1.2
46	wakaba	1.14	my $input = get_input_document ($http, $dom);
47	wakaba	1.55
48	wakaba	1.53	$out->input ($input);
49			$out->unset_flush;
50
51	wakaba	1.55	my $result = WebHACC::Result->new;
52			$result->output ($out);
53			$result->{conforming_min} = 1;
54			$result->{conforming_max} = 1;
55	wakaba	1.14
56	wakaba	1.55	$out->html ('<script src="../cc-script.js"></script>');
57	wakaba	1.54
58	wakaba	1.55	check_and_print ($input => $result => $out);
59
60			$result->generate_result_section;
61	wakaba	1.1
62	wakaba	1.53	$out->nav_list;
63	wakaba	1.16
64	wakaba	1.53	exit;
65	wakaba	1.35	}
66	wakaba	1.1
67	wakaba	1.53	sub check_and_print ($$$) {
68			my ($input, $result, $out) = @_;
69			my $original_input = $out->input;
70			$out->input ($input);
71	wakaba	1.31
72	wakaba	1.55	$input->generate_info_section ($result);
73
74	wakaba	1.54	$input->generate_transfer_sections ($result);
75	wakaba	1.31
76	wakaba	1.55	unless (defined $input->{s}) {
77			$result->{conforming_min} = 0;
78			return;
79			}
80	wakaba	1.31
81	wakaba	1.53	my $checker_class = {
82			'text/cache-manifest' => 'WebHACC::Language::CacheManifest',
83			'text/css' => 'WebHACC::Language::CSS',
84			'text/html' => 'WebHACC::Language::HTML',
85			'text/x-webidl' => 'WebHACC::Language::WebIDL',
86
87			'text/xml' => 'WebHACC::Language::XML',
88			'application/atom+xml' => 'WebHACC::Language::XML',
89			'application/rss+xml' => 'WebHACC::Language::XML',
90			'image/svg+xml' => 'WebHACC::Language::XML',
91			'application/xhtml+xml' => 'WebHACC::Language::XML',
92			'application/xml' => 'WebHACC::Language::XML',
93			## TODO: Should we make all XML MIME Types fall
94			## into this category?
95
96			## NOTE: This type has different model from normal XML types.
97			'application/rdf+xml' => 'WebHACC::Language::XML',
98			}->{$input->{media_type}} \|\| 'WebHACC::Language::Default';
99
100			eval qq{ require $checker_class } or die "$0: Loading $checker_class: $@";
101			my $checker = $checker_class->new;
102			$checker->input ($input);
103			$checker->output ($out);
104			$checker->result ($result);
105
106			## TODO: A cache manifest MUST be text/cache-manifest
107			## TODO: WebIDL media type "text/x-webidl"
108
109			$checker->generate_syntax_error_section;
110			$checker->generate_source_string_section;
111
112	wakaba	1.55	my @subdoc;
113	wakaba	1.53	$checker->onsubdoc (sub {
114			push @subdoc, shift;
115			});
116
117			$checker->generate_structure_dump_section;
118			$checker->generate_structure_error_section;
119			$checker->generate_additional_sections;
120
121			=pod
122	wakaba	1.31
123			if (defined $doc or defined $el) {
124	wakaba	1.53
125	wakaba	1.33	print_listing_section ({
126			id => 'identifiers', label => 'IDs', heading => 'Identifiers',
127			}, $input, $elements->{id}) if keys %{$elements->{id}};
128			print_listing_section ({
129			id => 'terms', label => 'Terms', heading => 'Terms',
130			}, $input, $elements->{term}) if keys %{$elements->{term}};
131			print_listing_section ({
132			id => 'classes', label => 'Classes', heading => 'Classes',
133			}, $input, $elements->{class}) if keys %{$elements->{class}};
134	wakaba	1.53
135	wakaba	1.45	print_rdf_section ($input, $elements->{rdf}) if @{$elements->{rdf}};
136	wakaba	1.31	}
137	wakaba	1.34
138	wakaba	1.53	=cut
139
140	wakaba	1.34	my $id_prefix = 0;
141	wakaba	1.53	for my $_subinput (@subdoc) {
142	wakaba	1.55	my $subinput = WebHACC::Input::Subdocument->new (++$id_prefix);
143	wakaba	1.53	$subinput->{$_} = $_subinput->{$_} for keys %$_subinput;
144	wakaba	1.34	$subinput->{base_uri} = $subinput->{container_node}->base_uri
145			unless defined $subinput->{base_uri};
146	wakaba	1.55	$subinput->{parent_input} = $input;
147	wakaba	1.34
148	wakaba	1.55	$subinput->start_section ($result);
149	wakaba	1.53	check_and_print ($subinput => $result => $out);
150	wakaba	1.55	$subinput->end_section ($result);
151	wakaba	1.34	}
152	wakaba	1.53
153			$out->input ($original_input);
154	wakaba	1.31	} # check_and_print
155
156	wakaba	1.45
157	wakaba	1.7	{
158			my $Msg = {};
159
160			sub load_text_catalog ($) {
161	wakaba	1.53	# my $self = shift;
162	wakaba	1.7	my $lang = shift; # MUST be a canonical lang name
163	wakaba	1.26	open my $file, '<:utf8', "cc-msg.$lang.txt"
164			or die "$0: cc-msg.$lang.txt: $!";
165	wakaba	1.7	while (<$file>) {
166			if (s/^([^;]+);([^;]*);//) {
167			my ($type, $cls, $msg) = ($1, $2, $_);
168			$msg =~ tr/\x0D\x0A//d;
169			$Msg->{$type} = [$cls, $msg];
170			}
171			}
172			} # load_text_catalog
173
174	wakaba	1.53	sub get_text ($;$$) {
175			# my $self = shift;
176	wakaba	1.15	my ($type, $level, $node) = @_;
177	wakaba	1.7	$type = $level . ':' . $type if defined $level;
178	wakaba	1.29	$level = 'm' unless defined $level;
179	wakaba	1.7	my @arg;
180			{
181			if (defined $Msg->{$type}) {
182			my $msg = $Msg->{$type}->[1];
183	wakaba	1.10	$msg =~ s{<var>\$([0-9]+)</var>}{
184	wakaba	1.53	defined $arg[$1] ? ($arg[$1]) : '(undef)';
185			}ge; ##BUG: ^ must be escaped
186	wakaba	1.15	$msg =~ s{<var>{\@([A-Za-z0-9:_.-]+)}</var>}{
187			UNIVERSAL::can ($node, 'get_attribute_ns')
188	wakaba	1.53	? ($node->get_attribute_ns (undef, $1)) : ''
189			}ge; ## BUG: ^ must be escaped
190			$msg =~ s{<var>{\@}</var>}{ ## BUG: v must be escaped
191			UNIVERSAL::can ($node, 'value') ? ($node->value) : ''
192	wakaba	1.15	}ge;
193	wakaba	1.17	$msg =~ s{<var>{local-name}</var>}{
194			UNIVERSAL::can ($node, 'manakai_local_name')
195	wakaba	1.53	? ($node->manakai_local_name) : ''
196			}ge; ## BUG: ^ must be escaped
197	wakaba	1.17	$msg =~ s{<var>{element-local-name}</var>}{
198			(UNIVERSAL::can ($node, 'owner_element') and
199			$node->owner_element)
200	wakaba	1.53	? ($node->owner_element->manakai_local_name)
201			: '' ## BUG: ^ must be escaped
202	wakaba	1.17	}ge;
203	wakaba	1.29	return ($type, 'level-' . $level . ' ' . $Msg->{$type}->[0], $msg);
204	wakaba	1.7	} elsif ($type =~ s/:([^:]*)$//) {
205			unshift @arg, $1;
206			redo;
207			}
208			}
209	wakaba	1.53	return ($type, 'level-'.$level, ($_[0]));
210			## BUG: ^ must be escaped
211	wakaba	1.7	} # get_text
212
213			}
214
215	wakaba	1.9	sub get_input_document ($$) {
216			my ($http, $dom) = @_;
217
218	wakaba	1.54	require Encode;
219			my $request_uri = Encode::decode ('utf-8', $http->get_parameter ('uri'));
220	wakaba	1.53	my $r = WebHACC::Input->new;
221	wakaba	1.9	if (defined $request_uri and length $request_uri) {
222			my $uri = $dom->create_uri_reference ($request_uri);
223			unless ({
224			http => 1,
225			}->{lc $uri->uri_scheme}) {
226	wakaba	1.54	$r = WebHACC::Input::Error->new;
227			$r->{uri} = $request_uri;
228			$r->{request_uri} = $request_uri;
229			$r->{error_status_text} = 'URL scheme not allowed';
230	wakaba	1.9	}
231
232			require Message::Util::HostPermit;
233			my $host_permit = new Message::Util::HostPermit;
234			$host_permit->add_rule (<<EOH);
235			Allow host=suika port=80
236			Deny host=suika
237			Allow host=suika.fam.cx port=80
238			Deny host=suika.fam.cx
239			Deny host=localhost
240			Deny host=*.localdomain
241			Deny ipv4=0.0.0.0/8
242			Deny ipv4=10.0.0.0/8
243			Deny ipv4=127.0.0.0/8
244			Deny ipv4=169.254.0.0/16
245			Deny ipv4=172.0.0.0/11
246			Deny ipv4=192.0.2.0/24
247			Deny ipv4=192.88.99.0/24
248			Deny ipv4=192.168.0.0/16
249			Deny ipv4=198.18.0.0/15
250			Deny ipv4=224.0.0.0/4
251			Deny ipv4=255.255.255.255/32
252			Deny ipv6=0::0/0
253			Allow host=*
254			EOH
255			unless ($host_permit->check ($uri->uri_host, $uri->uri_port \|\| 80)) {
256	wakaba	1.54	my $r = WebHACC::Input::Error->new;
257			$r->{uri} = $request_uri;
258			$r->{request_uri} = $request_uri;
259			$r->{error_status_text} = 'Connection to the host is forbidden';
260			return $r;
261	wakaba	1.9	}
262
263			require LWP::UserAgent;
264			my $ua = WDCC::LWPUA->new;
265			$ua->{wdcc_dom} = $dom;
266			$ua->{wdcc_host_permit} = $host_permit;
267			$ua->agent ('Mozilla'); ## TODO: for now.
268			$ua->parse_head (0);
269			$ua->protocols_allowed ([qw/http/]);
270			$ua->max_size (1000_000);
271			my $req = HTTP::Request->new (GET => $request_uri);
272	wakaba	1.28	$req->header ('Accept-Encoding' => 'identity, *; q=0');
273	wakaba	1.9	my $res = $ua->request ($req);
274	wakaba	1.16	## TODO: 401 sets \|is_success\| true.
275			if ($res->is_success or $http->get_parameter ('error-page')) {
276	wakaba	1.9	$r->{base_uri} = $res->base; ## NOTE: It does check \|Content-Base\|, \|Content-Location\|, and <base>. ## TODO: Use our own code!
277			$r->{uri} = $res->request->uri;
278			$r->{request_uri} = $request_uri;
279
280			## TODO: More strict parsing...
281			my $ct = $res->header ('Content-Type');
282	wakaba	1.22	if (defined $ct and $ct =~ /;\scharset\s=\s*"?([^\s;"]+)"?/i) {
283	wakaba	1.9	$r->{charset} = lc $1;
284			$r->{charset} =~ tr/\\//d;
285	wakaba	1.26	$r->{official_charset} = $r->{charset};
286	wakaba	1.9	}
287
288	wakaba	1.16	my $input_charset = $http->get_parameter ('charset');
289	wakaba	1.9	if (defined $input_charset and length $input_charset) {
290			$r->{charset_overridden}
291			= (not defined $r->{charset} or $r->{charset} ne $input_charset);
292			$r->{charset} = $input_charset;
293	wakaba	1.25	}
294
295			## TODO: Support for HTTP Content-Encoding
296	wakaba	1.9
297			$r->{s} = ''.$res->content;
298	wakaba	1.25
299			require Whatpm::ContentType;
300			($r->{official_type}, $r->{media_type})
301			= Whatpm::ContentType->get_sniffed_type
302			(get_file_head => sub {
303			return substr $r->{s}, 0, shift;
304			},
305			http_content_type_byte => $ct,
306			has_http_content_encoding =>
307			defined $res->header ('Content-Encoding'),
308			supported_image_types => {});
309	wakaba	1.9	} else {
310			$r->{uri} = $res->request->uri;
311			$r->{request_uri} = $request_uri;
312			$r->{error_status_text} = $res->status_line;
313			}
314
315			$r->{header_field} = [];
316			$res->scan (sub {
317			push @{$r->{header_field}}, [$_[0], $_[1]];
318			});
319			$r->{header_status_code} = $res->code;
320			$r->{header_status_text} = $res->message;
321			} else {
322	wakaba	1.16	$r->{s} = ''.$http->get_parameter ('s');
323	wakaba	1.9	$r->{uri} = q<thismessage:/>;
324			$r->{request_uri} = q<thismessage:/>;
325			$r->{base_uri} = q<thismessage:/>;
326	wakaba	1.16	$r->{charset} = ''.$http->get_parameter ('_charset_');
327	wakaba	1.9	$r->{charset} =~ s/\s+//g;
328			$r->{charset} = 'utf-8' if $r->{charset} eq '';
329	wakaba	1.26	$r->{official_charset} = $r->{charset};
330	wakaba	1.9	$r->{header_field} = [];
331	wakaba	1.25
332			require Whatpm::ContentType;
333			($r->{official_type}, $r->{media_type})
334			= Whatpm::ContentType->get_sniffed_type
335			(get_file_head => sub {
336			return substr $r->{s}, 0, shift;
337			},
338			http_content_type_byte => undef,
339			has_http_content_encoding => 0,
340			supported_image_types => {});
341	wakaba	1.9	}
342
343	wakaba	1.16	my $input_format = $http->get_parameter ('i');
344	wakaba	1.9	if (defined $input_format and length $input_format) {
345			$r->{media_type_overridden}
346			= (not defined $r->{media_type} or $input_format ne $r->{media_type});
347			$r->{media_type} = $input_format;
348			}
349			if (defined $r->{s} and not defined $r->{media_type}) {
350			$r->{media_type} = 'text/html';
351			$r->{media_type_overridden} = 1;
352			}
353
354			if ($r->{media_type} eq 'text/xml') {
355			unless (defined $r->{charset}) {
356			$r->{charset} = 'us-ascii';
357	wakaba	1.26	$r->{official_charset} = $r->{charset};
358	wakaba	1.9	} elsif ($r->{charset_overridden} and $r->{charset} eq 'us-ascii') {
359			$r->{charset_overridden} = 0;
360			}
361			}
362
363			if (length $r->{s} > 1000_000) {
364			$r->{error_status_text} = 'Entity-body too large';
365			delete $r->{s};
366			return $r;
367			}
368
369	wakaba	1.35	$r->{inner_html_element} = $http->get_parameter ('e');
370
371	wakaba	1.9	return $r;
372			} # get_input_document
373
374			package WDCC::LWPUA;
375			BEGIN { push our @ISA, 'LWP::UserAgent'; }
376
377			sub redirect_ok {
378			my $ua = shift;
379			unless ($ua->SUPER::redirect_ok (@_)) {
380			return 0;
381			}
382
383			my $uris = $_[1]->header ('Location');
384			return 0 unless $uris;
385			my $uri = $ua->{wdcc_dom}->create_uri_reference ($uris);
386			unless ({
387			http => 1,
388			}->{lc $uri->uri_scheme}) {
389			return 0;
390			}
391			unless ($ua->{wdcc_host_permit}->check ($uri->uri_host, $uri->uri_port \|\| 80)) {
392			return 0;
393			}
394			return 1;
395			} # redirect_ok
396
397	wakaba	1.1	=head1 AUTHOR
398
399			Wakaba <w@suika.fam.cx>.
400
401			=head1 LICENSE
402
403	wakaba	1.35	Copyright 2007-2008 Wakaba <w@suika.fam.cx>
404	wakaba	1.1
405			This library is free software; you can redistribute it
406			and/or modify it under the same terms as Perl itself.
407
408			=cut
409
410	wakaba	1.56	## $Date: 2008/07/21 05:24:32 $