Bug#482143: planet: Exceptions out given seemingly valid RSS feeds (wordpress.com)

[Chris Lamb]

tags 482143 + patch
thanks

Patch attached.

>         <media:content url="http://example.com/quxx.png"; medium="image">
>             <media:title type="html">baz</media:title>
>         </media:content>

My reading of this is that feedparser can not apply a "type" to this snippet,
which planet requires in order to determine what form of escaping the
content requires.

The attached patch skips "content fields" to which a type has not been
assigned, which is safer than simply emitting the content (which the current
code does without the failed type lookup, eek).

If any content is being lost, we could simply apply HTML sanitation. This
change does not result in any changes to the test suite.


Regards,

-- 
Chris Lamb, UK                                       [EMAIL PROTECTED]
                                                            GPG: 0x634F9A20

diff -urNad planet-2.0.orig/planet/__init__.py planet-2.0/planet/__init__.py
--- planet-2.0.orig/planet/__init__.py  2008-05-21 01:55:46.000000000 +0100
+++ planet-2.0/planet/__init__.py       2008-05-21 01:56:22.000000000 +0100
@@ -880,15 +880,18 @@
                 # Content field: concatenate the values
                 value = ""
                 for item in entry[key]:
-                    if item.type == 'text/html':
-                        item.value = sanitize.HTML(item.value)
-                    elif item.type == 'text/plain':
-                        item.value = escape(item.value)
-                    if item.has_key('language') and item.language and \
-                       (not self._channel.has_key('language') or
-                       item.language != self._channel.language) :
-                        self.set_as_string(key + "_language", item.language)
-                    value += cache.utf8(item.value)
+                    try:
+                        if item.type == 'text/html':
+                            item.value = sanitize.HTML(item.value)
+                        elif item.type == 'text/plain':
+                            item.value = escape(item.value)
+                        if item.has_key('language') and item.language and \
+                           (not self._channel.has_key('language') or
+                           item.language != self._channel.language) :
+                            self.set_as_string(key + "_language", 
item.language)
+                        value += cache.utf8(item.value)
+                    except AttributeError:
+                        pass
                 self.set_as_string(key, value)
             elif isinstance(entry[key], (str, unicode)):
                 # String fields

signature.asc
Description: PGP signature