Register or log in - lost password?

Username Password

Forums
Feed Informer Support

1. azdrew
   Member

   Hi, I noticed a new problem with one of my feeds this morning, and I'm not sure what's happening.

   I have a Feed Digest feed with 3 source feeds. One comes from a Simple Machines Forum (SMF) bulletin board.

   The source feed is at

   http://www.aplaceofgrace.org/newsfeed_forum

   it appears that the feed code for SMF wrappers all text with CDATA tags.

   There is a new post with an apostrophe in the text. The source feed closes on CDATA, inserts the numeric code for the apostrophe, and then starts a new CDATA for the rest of the description. The title of the item in the feed is "New Member Sunday" for your reference.

   This is getting mangled in the Feed Digest Feed. It looks like Feed Digest strips all of the CDATA, except its doing very strange things with this second one around the apostrophe. I get lots of html lt and gt codes, and a wayward CDATA in the Feed Digest stream. I get an "Invalid EOF" from feedvalidator on the feed.

   My Feed Digest feed is here

   http://app.feeddigest.com/digest3/FMY1LDF52G.rss

   Any ideas? This seems like an SMF oddity perhaps (I have a Joomla feed that doesn't split the CDATA like this), but my SMF feed does validate according to feedvalidator.

   I've been searching high and low for a good feed merging service, and I really like Feed Digest so far during my trial, so I'm hoping this is something that can get worked out. Thanks

   Posted 17 years ago #

2. admin
   Key Master

   That's interesting. I've never seen that before.

   This is caused by our new parser which went into service a couple of weeks ago. We used to use a 'real' XML parser to parse feeds, but since so many feeds are totally wrecked/broken/invalid.. we've had to move to a more liberal system just so we can parse '100%' (or as close to) of feeds.

   One thing that was never anticipated in all of this is that multiple CDATA sections would exist within a single section in this way.. so they're not caught and managed as single CDATA sections are. It does appear, however, that as soon as this feature is added, we'll be looking good, since that's the only problem arising in your file.

   I'll try to implement an update to the parser tomorrow to deal with multiple CDATA sections and then, if it's okay with you, wipe the contents of that source feed and recrawl it immediately so that you get the correct output. Let me know if this is okay with you.

   Posted 17 years ago #

3. azdrew
   Member

   Sure, that would be great - thanks!

   Posted 17 years ago #

4. admin
   Key Master

   Have a look now and see what you think :)

   Posted 17 years ago #

5. azdrew
   Member

   Awesome! Thanks very much.

   I just upgraded my account ;-)

   Posted 17 years ago #

RSS feed for this topic

Topic Closed

This topic has been closed to new replies.