prosody-modules: mod_xhtmlim/mod

annotate mod_xhtmlim/mod_xhtmlim.lua @ 5173:460f78654864

mod_muc_rtbl: also filter messages This was a bit tricky because we don't want to run the JIDs through SHA256 on each message. Took a while to come up with this simple plan of just caching the SHA256 of the JIDs on the occupants. This will leave some dirt in the occupants after unloading the module, but that should be ok; once they cycle the room, the hashes will be gone. This is direly needed, otherwise, there is a tight race between the moderation activities and the actors joining the room.

author	Jonas Schäfer <jonas@wielicki.name>
date	Tue, 21 Feb 2023 21:37:27 +0100
parents	1f68287138e3
children

rev	line source
2865 f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	1 -- XEP-0071: XHTML-IM sanitizing
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	2
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	3 local assert = assert;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	4
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	5 local st = require "util.stanza";
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	6 local url = require "socket.url";
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	7
3699 1f68287138e3 mod_xhtmlim: Default to stripping @style attribute by default Kim Alvefur <zash@zash.se> parents: 3404 diff changeset	8 local no_styles = module:get_option_boolean("strip_xhtml_style", true);
2865 f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	9
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	10 -- Tables from XEP-0071
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	11 local xeptables = [[
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	12 <body/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	13 <head/> profile
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	14 <html/> version
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	15 <title/>
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	16 <abbr/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	17 <acronym/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	18 <address/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	19 <blockquote/> class, id, title; style; cite
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	20 <br/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	21 <cite/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	22 <code/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	23 <dfn/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	24 <div/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	25 <em/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	26 <h1/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	27 <h2/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	28 <h3/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	29 <h4/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	30 <h5/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	31 <h6/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	32 <kbd/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	33 <p/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	34 <pre/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	35 <q/> class, id, title; style; cite
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	36 <samp/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	37 <span/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	38 <strong/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	39 <var/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	40 <a/> class, id, title; style; accesskey, charset, href, hreflang, rel, rev, tabindex, type
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	41 <dl/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	42 <dt/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	43 <dd/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	44 <ol/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	45 <ul/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	46 <li/> class, id, title; style
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	47 <img/> class, id, title; style; alt, height, longdesc, src, width
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	48 ]];
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	49
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	50 -- map of whitelisted tag names to set of allowed attributes
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	51 local tags = {}; -- { string : { string : boolean } }
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	52
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	53 for tag, attrs in xeptables:gmatch("<(%w+)/>([^\n]*)") do
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	54 tags[tag] = { xmlns = true, ["xml:lang"] = true };
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	55 for attr in attrs:gmatch("%w+") do
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	56 tags[tag][attr] = true;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	57 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	58 if no_styles then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	59 tags[tag]["style"] = nil;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	60 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	61 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	62
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	63 -- module:log("debug", "tags = %s;", require "util.serialization".serialize(tags));
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	64
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	65 -- TODO Decide if disallowed tags should be bounced or silently discarded.
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	66 -- XEP says "ignore" and replace tag with text content, but that would
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	67 -- need a different transform which can't use `maptags`.
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	68 if not module:get_option_boolean("bounce_invalid_xhtml", false) then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	69 assert = function (x) return x end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	70 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	71
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	72 local function sanitize_xhtml(tag)
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	73 -- module:log("debug", "sanitize_xhtml(<{%s}%s>)", tag.attr.xmlns, tag.name);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	74 if tag.attr.xmlns == "http://www.w3.org/1999/xhtml" then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	75 local allowed = assert(tags[tag.name], tag.name);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	76 if allowed then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	77 for attr, value in pairs(tag.attr) do
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	78 if not allowed[attr] then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	79 -- module:log("debug", "Removing disallowed attribute %q from <%s>", attr, tag.name);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	80 tag.attr[attr] = nil;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	81 elseif attr == "src" or attr == "href" then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	82 local urlattr = url.parse(value);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	83 local scheme = urlattr and urlattr.scheme;
2866 276f7af8afd1 mod_xhtmlim: Fix scheme check (thanks wiktor) Kim Alvefur <zash@zash.se> parents: 2865 diff changeset	84 if scheme ~= "http" and scheme ~= "https" and scheme ~= "mailto" and scheme ~= "xmpp" and scheme ~= "cid" then
2865 f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	85 tag.attr[attr] = "https://url.was.invalid/";
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	86 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	87 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	88 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	89 else
3404 6444fb5dbb51 mod_xhtmlim: Turn unknown elements into <span> Kim Alvefur <zash@zash.se> parents: 2866 diff changeset	90 tag.name = "span"
6444fb5dbb51 mod_xhtmlim: Turn unknown elements into <span> Kim Alvefur <zash@zash.se> parents: 2866 diff changeset	91 tag.attr = { xmlns = "http://www.w3.org/1999/xhtml" }
2865 f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	92 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	93 -- Check child tags
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	94 tag:maptags(sanitize_xhtml);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	95 -- This tag is clean!
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	96 return tag;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	97 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	98 -- Not xhtml, probably best to discard it
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	99 return nil;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	100 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	101
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	102 -- Check for xhtml-im, sanitize if exists
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	103 local function message_handler(event)
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	104 local stanza = event.stanza;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	105 if stanza:get_child("html", "http://jabber.org/protocol/xhtml-im") then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	106 stanza = st.clone(stanza);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	107 if pcall(function() -- try
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	108 stanza:get_child("html", "http://jabber.org/protocol/xhtml-im"):maptags(sanitize_xhtml);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	109 end) then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	110 event.stanza = stanza;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	111 else -- catch
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	112 if stanza.attr.type ~= "error" then
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	113 event.origin.send(st.error_reply(stanza, "modify", "not-acceptable", "Stanza contained illegal XHTML-IM tag"));
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	114 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	115 return true;
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	116 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	117 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	118 end
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	119
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	120 -- Stanzas received from clients
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	121 module:hook("pre-message/bare", message_handler, 71);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	122 module:hook("pre-message/full", message_handler, 71);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	123 module:hook("pre-message/host", message_handler, 71);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	124
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	125 -- Stanzas about to be delivered to clients
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	126 module:hook("message/bare", message_handler, 71);
f6ed4421167d mod_xhtmlim: Attempts to sanitize XMTML-IM messages Kim Alvefur <zash@zash.se> parents: diff changeset	127 module:hook("message/full", message_handler, 71);

Mercurial > prosody-modules

annotate mod_xhtmlim/mod_xhtmlim.lua @ 5173:460f78654864