Fix #2447 (OTR & HTML) -- partial WONTFIX

- Guess-parse the OTR messages in search for xhtml upon arrival
- add a configurable option to decode it or not
- We have XHTML-IM for a reason, and therefore we will *not* implement a
  full html parser for clients that dump whatever formatting inside the
  OTR payload (looking at you, pidgin)

2 files changed, 22 insertions, 8 deletions

diff --git a/plugins/otr.py b/plugins/otr.py
index 3f22687f..96792a4d 100644
--- a/plugins/otr.py
+++ b/plugins/otr.py
@@ -121,6 +121,11 @@ Configuration
 .. glossary::
     :sorted:
 
+    decode_xhtml
+        **Default:** ``true``
+
+        Decode embedded XHTML.
+
     keys_dir
         **Default:** ``$XDG_DATA_HOME/poezio/otr``
 
@@ -141,8 +146,8 @@ Configuration
 
         Log conversations (OTR start/end marker, and messages).
 
-The :term:`allow_v1`, :term:`allow_v2` and :term:`log` configuration
-parameters are tab-specific.
+The :term:`allow_v1`, :term:`allow_v2`, :term:`decode_html`
+and :term:`log` configuration parameters are tab-specific.
 
 Important details
 -----------------
@@ -166,6 +171,7 @@ import curses
 from potr.context import NotEncryptedError, UnencryptedMessage, ErrorReceived, NotOTRMessage,\
         STATE_ENCRYPTED, STATE_PLAINTEXT, STATE_FINISHED, Context, Account, crypt
 
+import xhtml
 from plugin import BasePlugin
 from tabs import ConversationTab, DynamicConversationTab, PrivateTab
 from common import safeJID
@@ -463,8 +469,14 @@ class Plugin(BasePlugin):
             user = None
 
         body = txt.decode()
+        if self.config.get_by_tabname('decode_xhtml', True, msg['from'].bare):
+            try:
+                body = xhtml.xhtml_to_poezio_colors(body, force=True)
+            except:
+                pass
         tab.add_message(body, nickname=tab.nick, jid=msg['from'],
-                forced_user=user, typ=ctx.log, nick_color=theming.get_theme().COLOR_REMOTE_USER)
+                        forced_user=user, typ=ctx.log,
+                        nick_color=theming.get_theme().COLOR_REMOTE_USER)
         hl(tab)
         self.core.refresh_window()
         del msg['body']
diff --git a/src/xhtml.py b/src/xhtml.py
index 33a89b5e..48664311 100644
--- a/src/xhtml.py
+++ b/src/xhtml.py
@@ -281,13 +281,15 @@ def trim(string):
     return re.sub(whitespace_re, ' ', string)
 
 class XHTMLHandler(sax.ContentHandler):
-    def __init__(self):
+    def __init__(self, force_ns=False):
         self.builder = []
         self.formatting = []
         self.attrs = []
         self.list_state = []
         self.is_pre = False
         self.a_start = 0
+        # do not care about xhtml-in namespace
+        self.force_ns = force_ns
 
     @property
     def result(self):
@@ -305,7 +307,7 @@ class XHTMLHandler(sax.ContentHandler):
         self.builder.append(characters if self.is_pre else trim(characters))
 
     def startElementNS(self, name, _, attrs):
-        if name[0] != XHTML_NS:
+        if name[0] != XHTML_NS and not self.force_ns:
             return
 
         builder = self.builder
@@ -356,7 +358,7 @@ class XHTMLHandler(sax.ContentHandler):
             self.append_formatting('\x19b')
 
     def endElementNS(self, name, _):
-        if name[0] != XHTML_NS:
+        if name[0] != XHTML_NS and not self.force_ns:
             return
 
         builder = self.builder
@@ -387,13 +389,13 @@ class XHTMLHandler(sax.ContentHandler):
         if 'title' in attrs:
             builder.append(' [' + attrs['title'] + ']')
 
-def xhtml_to_poezio_colors(xml):
+def xhtml_to_poezio_colors(xml, force=False):
     if isinstance(xml, str):
         xml = xml.encode('utf8')
     elif not isinstance(xml, bytes):
         xml = ET.tostring(xml)
 
-    handler = XHTMLHandler()
+    handler = XHTMLHandler(force_ns=force)
     parser = sax.make_parser()
     parser.setFeature(sax.handler.feature_namespaces, True)
     parser.setContentHandler(handler)