2 package IkiWiki::Plugin::htmlscrubber;
8 # This regexp matches urls that are in a known safe scheme.
9 # Feel free to use it from other plugins.
13 hook(type => "sanitize", id => "htmlscrubber", call => \&sanitize);
15 # Only known uri schemes are allowed to avoid all the ways of
16 # embedding javascrpt.
17 # List at http://en.wikipedia.org/wiki/URI_scheme
18 my $uri_schemes=join("|",
19 # IANA registered schemes
20 "http", "https", "ftp", "mailto", "file", "telnet", "gopher",
21 "aaa", "aaas", "acap", "cap", "cid", "crid",
22 "dav", "dict", "dns", "fax", "go", "h323", "im", "imap",
23 "ldap", "mid", "news", "nfs", "nntp", "pop", "pres",
24 "sip", "sips", "snmp", "tel", "urn", "wais", "xmpp",
25 "z39\.50r", "z39\.50s",
26 # Selected unofficial schemes
27 "aim", "callto", "cvs", "ed2k", "feed", "fish", "gg",
28 "irc", "ircs", "lastfm", "ldaps", "magnet", "mms",
29 "msnim", "notes", "rsync", "secondlife", "skype", "ssh",
30 "sftp", "smb", "sms", "snews", "webcal", "ymsgr",
32 # data is a special case. Allow data:image/*, but
33 # disallow data:text/javascript and everything else.
34 $safe_url_regexp=qr/^(?:(?:$uri_schemes):|data:image\/|[^:]+$)/i;
37 sub sanitize (@) { #{{{
39 return scrubber()->scrub($params{content});
44 return $_scrubber if defined $_scrubber;
46 eval q{use HTML::Scrubber};
48 # Lists based on http://feedparser.org/docs/html-sanitization.html
49 # With html 5 video and audio tags added.
50 $_scrubber = HTML::Scrubber->new(
52 a abbr acronym address area b big blockquote br br/
53 button caption center cite code col colgroup dd del
54 dfn dir div dl dt em fieldset font form h1 h2 h3 h4
55 h5 h6 hr hr/ i img input ins kbd label legend li map
56 menu ol optgroup option p p/ pre q s samp select small
57 span strike strong sub sup table tbody td textarea
58 tfoot th thead tr tt u ul var
61 default => [undef, { (
63 abbr accept accept-charset accesskey
64 align alt axis border cellpadding cellspacing
65 char charoff charset checked class
66 clear cols colspan color compact coords
67 datetime dir disabled enctype for frame
68 headers height hreflang hspace id ismap
69 label lang maxlength media method
70 multiple name nohref noshade nowrap prompt
71 readonly rel rev rows rowspan rules scope
72 selected shape size span start summary
73 tabindex target title type valign
75 autoplay loopstart loopend end
78 "/" => 1, # emit proper <hr /> XHTML
79 href => $safe_url_regexp,
80 src => $safe_url_regexp,
81 action => $safe_url_regexp,
82 cite => $safe_url_regexp,
83 longdesc => $safe_url_regexp,
84 poster => $safe_url_regexp,
85 usemap => $safe_url_regexp,