IkiWiki/Plugin/htmlscrubber.pm

   1 #!/usr/bin/perl
   2 package IkiWiki::Plugin::htmlscrubber;
   3
   4 use warnings;
   5 use strict;
   6 use IkiWiki 3.00;
   7
   8 # This regexp matches urls that are in a known safe scheme.
   9 # Feel free to use it from other plugins.
  10 our $safe_url_regexp;
  11
  12 sub import {
  13         hook(type => "getsetup", id => "htmlscrubber", call => \&getsetup);
  14         hook(type => "sanitize", id => "htmlscrubber", call => \&sanitize);
  15
  16         # Only known uri schemes are allowed to avoid all the ways of
  17         # embedding javascrpt.
  18         # List at http://en.wikipedia.org/wiki/URI_scheme
  19         my $uri_schemes=join("|", map quotemeta,
  20                 # IANA registered schemes
  21                 "http", "https", "ftp", "mailto", "file", "telnet", "gopher",
  22                 "aaa", "aaas", "acap",  "cap", "cid", "crid",
  23                 "dav", "dict", "dns", "fax", "go", "h323", "im", "imap",
  24                 "ldap", "mid", "news", "nfs", "nntp", "pop", "pres",
  25                 "sip", "sips", "snmp", "tel", "urn", "wais", "xmpp",
  26                 "z39.50r", "z39.50s",
  27                 # Selected unofficial schemes
  28                 "aim", "callto", "cvs", "ed2k", "feed", "fish", "gg",
  29                 "irc", "ircs", "lastfm", "ldaps", "magnet", "mms",
  30                 "msnim", "notes", "rsync", "secondlife", "skype", "ssh",
  31                 "sftp", "smb", "sms", "snews", "webcal", "ymsgr",
  32         );
  33         # data is a special case. Allow a few data:image/ types,
  34         # but disallow data:text/javascript and everything else.
  35         $safe_url_regexp=qr/^(?:(?:$uri_schemes):|data:image\/(?:png|jpeg|gif)|[^:]+(?:$|[\/\?]))/i;
  36 }
  37
  38 sub getsetup () {
  39         return
  40                 plugin => {
  41                         safe => 1,
  42                         rebuild => undef,
  43                         section => "core",
  44                 },
  45                 htmlscrubber_skip => {
  46                         type => "pagespec",
  47                         example => "!*/Discussion",
  48                         description => "PageSpec specifying pages not to scrub",
  49                         link => "ikiwiki/PageSpec",
  50                         safe => 1,
  51                         rebuild => undef,
  52                 },
  53 }
  54
  55 sub sanitize (@) {
  56         my %params=@_;
  57
  58         if (exists $config{htmlscrubber_skip} &&
  59             length $config{htmlscrubber_skip} &&
  60             exists $params{destpage} &&
  61             pagespec_match($params{destpage}, $config{htmlscrubber_skip})) {
  62                 return $params{content};
  63         }
  64
  65         return scrubber()->scrub($params{content});
  66 }
  67
  68 my $_scrubber;
  69 sub scrubber {
  70         return $_scrubber if defined $_scrubber;
  71
  72         eval q{use HTML::Scrubber};
  73         error($@) if $@;
  74         # Lists based on http://feedparser.org/docs/html-sanitization.html
  75         # With html5 tags added.
  76         $_scrubber = HTML::Scrubber->new(
  77                 allow => [qw{
  78                         a abbr acronym address area b big blockquote br br/
  79                         button caption center cite code col colgroup dd del
  80                         dfn dir div dl dt em fieldset font form h1 h2 h3 h4
  81                         h5 h6 hr hr/ i img input ins kbd label legend li map
  82                         menu ol optgroup option p p/ pre q s samp select small
  83                         span strike strong sub sup table tbody td textarea
  84                         tfoot th thead tr tt u ul var
  85
  86                         video audio source section nav article aside hgroup
  87                         header footer time mark canvas
  88                 }],
  89                 default => [undef, { (
  90                         map { $_ => 1 } qw{
  91                                 abbr accept accept-charset accesskey
  92                                 align alt axis border cellpadding cellspacing
  93                                 char charoff charset checked class
  94                                 clear cols colspan color compact coords
  95                                 datetime dir disabled enctype for frame
  96                                 headers height hreflang hspace id ismap
  97                                 label lang maxlength media method
  98                                 multiple name nohref noshade nowrap prompt
  99                                 readonly rel rev rows rowspan rules scope
 100                                 selected shape size span start summary
 101                                 tabindex target title type valign
 102                                 value vspace width
 103
 104                                 autofocus autoplay preload loopstart
 105                                 loopend end playcount controls pubdate
 106                                 placeholder min max step
 107                         } ),
 108                         "/" => 1, # emit proper <hr /> XHTML
 109                         href => $safe_url_regexp,
 110                         src => $safe_url_regexp,
 111                         action => $safe_url_regexp,
 112                         cite => $safe_url_regexp,
 113                         longdesc => $safe_url_regexp,
 114                         poster => $safe_url_regexp,
 115                         usemap => $safe_url_regexp,
 116                 }],
 117         );
 118         return $_scrubber;
 119 }
 120
 121 1