zs

commit 18454a1fe6f19aa9cdd780128a713066ee23ba9d
parent 0626d417d52845d03223244faa8238210ff87229
Author: Christoph Lohmann <20h@r-36.net>
Date:   Wed, 19 Mar 2014 18:14:03 +0100

Add handling for xml syntax errors.

If XML fails, try HTML. If that fails, bail.

Diffstat:
feed.py
 | 
15
+++++++++++++--
zs.py
 | 
3
+++

2 files changed, 16 insertions(+), 2 deletions(-)
diff --git a/feed.py b/feed.py
@@ -6,6 +6,7 @@
 #
 
 from lxml import objectify 
+from lxml import etree
 from datetime import datetime
 import dateutil.parser
 import urllib.request, urllib.parse, urllib.error
@@ -23,12 +24,22 @@ def removenamespaces(xml):
 				elem.tag = elem.tag[nsl:]
 
 def parsexml(astr):
-	xml = objectify.fromstring(astr)
-	removenamespaces(xml)
+	try:
+		xml = objectify.fromstring(astr)
+		removenamespaces(xml)
+	except etree.XMLSyntaxError:
+		try:
+			parser = etree.HTMLParser()
+			xml = objectify.fromstring(astr)
+			removenamespaces(xml)
+		except etree.XMLSyntaxError:
+			return None
 	return xml
 
 def parse(astr):
 	xml = parsexml(astr)
+	if xml == None:
+		return None
 
 	feed = {}
 	articles = []
diff --git a/zs.py b/zs.py
@@ -46,6 +46,9 @@ def run(db, selfeed=None, dryrun=False):
 			estr = "incompleteread"
 			continue
 
+		if curfeed == None:
+			continue
+
 		# retry handling
 		if estr != None:
 			if retries > 2: