I used to have the same error message on the previous release 2.8.
I confirm I have the same on 2.9.

I can suggest that "fix" that works on both 2.8 and 2.9 (works for me) :
on file helpers/ContentLoader.php, line 128

change that line :
$icon = $item->getIcon();

to this:
try {
$icon = $item->getIcon();
} catch(\exception $e) {
return;
}

I'm not a developer, and that "fix" didn't really fix the issue.

Sorry for the late reply, was away from the computer for a few days.

It looks like its some kind of general apache problem rather then a selfoss problem - enabling debug in config.ini on selfoss has this at the end of the update

• 03-28-13 9:34:17 Debug Memory usage: 7830112
  
• 03-28-13 9:34:17 Debug Memory peak usage: 21744664
  
• 03-28-13 9:34:17 Debug start insertion of new item "Giving a CNC knitting machine a new brain"
  
• 03-28-13 9:34:17 Debug item content sanitized
  
• 03-28-13 9:34:17 Debug use last icon: http://s1.wp.com/wp-content/themes/vip/hackaday2/images/favicon.ico?m=1353160633g
  
• 03-28-13 9:34:17 Debug item inserted
  
• 03-28-13 9:34:17 Debug Memory usage: 7848952
  
• 03-28-13 9:34:17 Debug Memory peak usage: 21744664
  
• 03-28-13 9:34:17 Debug destroy spout object
  
• 03-28-13 9:34:17 Debug ---
  
• 03-28-13 9:34:17 Debug start fetching source "Hanging Rock Comics (id: 209)
  
• 03-28-13 9:34:17 Debug spout successfully loaded: spouts\rss\feed
  
• 03-28-13 9:34:17 Debug fetch content
  
• 03-28-13 9:34:18 Debug minimum date: 2013-02-26 09:34:18
  
• 03-28-13 9:34:18 Debug start item fetching
  
• 03-28-13 9:34:18 Debug start insertion of new item "blahblah1"
  
• 03-28-13 9:34:18 Debug item content sanitized

which all looks normal.

Apaches log file has this:

[Thu Mar 28 09:34:20 2013] [error] [client 127.0.0.1] Undefined offset: 2

The previous line is from 6 minutes before that, and just about me using update with lower case rather then update with upper case:

[Thu Mar 28 09:28:35 2013] [error] [client 127.0.0.1] Negotiation: discovered file(s) matching request: /var/www/selfoss/update (None could be negotiated).

But that still doesn't give me any more clue as to what the error is.

Does no one else have this update issue?

This time, the update ran for 4 or 5 minutes and pulled around 270 posts - but looking at the comparative time in Google Reader, boing boing (which is only one of the subscriptions) appears to have had roughly this many alone over that time period.

Could it be working and have some other criteria google reader doesn't have for collecting posts? Or is the update process crapping out early? Any clues or anywhere else I should look?

It looks like its some kind of general apache problem rather then a selfoss problem - enabling debug in config.ini on selfoss has this at the end of the update

• 03-28-13 9:34:17 Debug Memory usage: 7830112
  
• 03-28-13 9:34:17 Debug Memory peak usage: 21744664
  
• 03-28-13 9:34:17 Debug start insertion of new item "Giving a CNC knitting machine a new brain"
  
• 03-28-13 9:34:17 Debug item content sanitized
  
• 03-28-13 9:34:17 Debug use last icon: http://s1.wp.com/wp-content/themes/vip/hackaday2/images/favicon.ico?m=1353160633g
  
• 03-28-13 9:34:17 Debug item inserted
  
• 03-28-13 9:34:17 Debug Memory usage: 7848952
  
• 03-28-13 9:34:17 Debug Memory peak usage: 21744664
  
• 03-28-13 9:34:17 Debug destroy spout object
  
• 03-28-13 9:34:17 Debug ---
  
• 03-28-13 9:34:17 Debug start fetching source "Hanging Rock Comics (id: 209)
  
• 03-28-13 9:34:17 Debug spout successfully loaded: spouts\rss\feed
  
• 03-28-13 9:34:17 Debug fetch content
  
• 03-28-13 9:34:18 Debug minimum date: 2013-02-26 09:34:18
  
• 03-28-13 9:34:18 Debug start item fetching
  
• 03-28-13 9:34:18 Debug start insertion of new item "blahblah1"
  
• 03-28-13 9:34:18 Debug item content sanitized

which all looks normal.

Apaches log file has this:

[Thu Mar 28 09:34:20 2013] [error] [client 127.0.0.1] Undefined offset: 2

The previous line is from 6 minutes before that, and just about me using update with lower case rather then update with upper case:

[Thu Mar 28 09:28:35 2013] [error] [client 127.0.0.1] Negotiation: discovered file(s) matching request: /var/www/selfoss/update (None could be negotiated).

But that still doesn't give me any more clue as to what the error is.

Does no one else have this update issue?

This time, the update ran for 4 or 5 minutes and pulled around 270 posts - but looking at the comparative time in Google Reader, boing boing (which is only one of the subscriptions) appears to have had roughly this many alone over that time period.

Could it be working and have some other criteria google reader doesn't have for collecting posts? Or is the update process crapping out early? Any clues or anywhere else I should look?

Selfoss is excellent! At first I was a little hesitant about having everything in one feed, but actually since you parse the text and format things in three columns it looks great!

I'm having a problem with updating however - I'm hosting it on a raspberry pi and whenever I update the update fails with the following error message:

Internal Server Error 500

Internal Server Error
Undefined offset: 2.

Reading about it it sounds like it's some kind of python error? This happens if I run it via the website (mywebsite/selfoss/Update with a capital U, which I gather is a change from 2.2) or via wget or curl on the terminal

curl:

/usr/bin/curl -s -S -m 3300 http://127.0.0.1/selfoss/Update

wget:

wget 127.0.0.1/selfoss/Update

In case it's relevant I have 250 or so feeds and have tried both the mysql and sqlite databases with it - I get the same error. It's 256mb model b pi, and memory usage is only coming up to 200 meg or so, both sqlite and mysql.

Is anyone else having the same issue, and how would I go about fixing it?