2 package IkiWiki::Plugin::htmlscrubber;
9 hook(type => "sanitize", id => "htmlscrubber", call => \&sanitize);
12 sub sanitize (@) { #{{{
14 return scrubber()->scrub($params{content});
19 return $_scrubber if defined $_scrubber;
21 # Only known uri schemes are allowed to avoid all the ways of
22 # embedding javascrpt.
23 # List at http://en.wikipedia.org/wiki/URI_scheme
24 my $uri_schemes=join("|",
25 # IANA registered schemes
26 "http", "https", "ftp", "mailto", "file", "telnet", "gopher",
27 "aaa", "aaas", "acap", "cap", "cid", "crid",
28 "dav", "dict", "dns", "fax", "go", "h323", "im", "imap",
29 "ldap", "mid", "news", "nfs", "nntp", "pop", "pres",
30 "sip", "sips", "snmp", "tel", "urn", "wais", "xmpp",
32 # data is a special case. Allow data:text/<image>, but
33 # disallow data:text/javascript and everything else.
34 qr/data:text\/(?:png|gif|jpeg)/,
35 # Selected unofficial schemes
36 "about", "aim", "callto", "cvs", "ed2k", "feed", "fish", "gg",
37 "irc", "ircs", "lastfm", "ldaps", "magnet", "mms",
38 "msnim", "notes", "rsync", "secondlife", "skype", "ssh",
39 "sftp", "sms", "steam", "webcal", "ymsgr",
41 my $link=qr/^(?:$uri_schemes:|[^:]+$)/i;
43 eval q{use HTML::Scrubber};
45 # Lists based on http://feedparser.org/docs/html-sanitization.html
46 # With html 5 video and audio tags added.
47 $_scrubber = HTML::Scrubber->new(
49 a abbr acronym address area b big blockquote br br/
50 button caption center cite code col colgroup dd del
51 dfn dir div dl dt em fieldset font form h1 h2 h3 h4
52 h5 h6 hr hr/ i img input ins kbd label legend li map
53 menu ol optgroup option p p/ pre q s samp select small
54 span strike strong sub sup table tbody td textarea
55 tfoot th thead tr tt u ul var
58 default => [undef, { (
60 abbr accept accept-charset accesskey
61 align alt axis border cellpadding cellspacing
62 char charoff charset checked cite class
63 clear cols colspan color compact coords
64 datetime dir disabled enctype for frame
65 headers height hreflang hspace id ismap
66 label lang longdesc maxlength media method
67 multiple name nohref noshade nowrap prompt
68 readonly rel rev rows rowspan rules scope
69 selected shape size span start summary
70 tabindex target title type usemap valign
72 autoplay loopstart loopend end
75 "/" => 1, # emit proper <hr /> XHTML