Searchresult Episode Butler Error

[tedrock]

Started getting this today, I don't know if a change was made to TVrage or Newzbin to cause this but every time my EB does a search now I'm given the following error in red:
Mar 17 '10 03:42:21 AM
    Found 0.
Mar 17 '10 03:42:21 AM
    Searchresult doesn't have a csv contenttype.
Mar 17 '10 03:39:12 AM
    Searching for episodes.

Also seems that when I do get a result the NZB is not downloaded correctly because it doesn't get passed to SAB and I can't find the NZB anywhere on my drive. This may have started after it preformed "Updating episode information." but I can't tell because the log won't go back any further then that.

[clony3k]

Same problem here, woot in the hell is this problem???

[Eejit]

I'm not getting this error.
I suppose it could be associated with a particular show which you two download and I dont.

[tedrock]

since the problem started yesterday it's matched two shows (NSCI and Biggest Loser) and both times the NZB was either not downloaded or just not pasted to SAB and lost some where (cannot be located on my drive). The error does not appear 100% of the time and when it does appear it's unrelated to an episode being found or not.

Log

Mar 17 '10 19:41:23 PM
    Found 0.
Mar 17 '10 19:41:16 PM
    Searching for episodes.
Mar 17 '10 17:41:16 PM
    Found 0.
Mar 17 '10 17:41:09 PM
    Searching for episodes.
Mar 17 '10 15:41:09 PM
    Found 0.
Mar 17 '10 15:41:09 PM
    Searchresult doesn't have a csv contenttype.
Mar 17 '10 15:38:02 PM
    Found The Biggest Loser - 9x09 - Season 9, Episode 9.
Mar 17 '10 15:38:01 PM
    Searching for episodes.
Mar 17 '10 13:43:01 PM
    Found 0.
Mar 17 '10 13:43:01 PM
    Searchresult doesn't have a csv contenttype.
Mar 17 '10 13:39:52 PM
    Searching for episodes.
Mar 17 '10 11:44:52 AM
    Found 0.
Mar 17 '10 11:44:52 AM
    Searchresult doesn't have a csv contenttype.
Mar 17 '10 11:41:46 AM
    Searching for episodes.
Mar 17 '10 09:41:46 AM
    Found 0.

[Janoz]

This problem is caused by a change in the newzbin feed urls. I opened a support ticket and they are working on it. There is a problem with the authentication. The 'no cvs content type' is because a html login page is returned instead of a csv file containing the search results. A quick fix for now is to force basic authentication on the feeds (which previously was the default) by adding hauth = 1 to the search parameters. This can be done this way:

Go to "Config" -> "Newzbin"

and add "hauth = 1" to "Search parameters"

The content should look something like

Code: Select all

fpn = p
sort = ps_edit_date
order = desc
u_completions = 1
hauth = 1

This fix workt for me yesterday, but some other users still reported problems so I can not guarantee this will work for everyone. I hope I'll hear soon from the newzbin developers.

[pspgimp]

Fix does not work for me, hope someone fixes this for good soon....

[pspgimp]

Just logged a ticket of my own with newzbin and they said that Janoz has informed them everything is now fixed ??

maybe there is some confusion ?


pspgimp

[clony3k]

the fix tid not fix the problem for me also. Hope the problem will be solved, i love episodebutler !!!

[alan]

I can also confirm that this doesn't fix the problem.

What is also worse, is that Episode Butler actually flags the episodes as being downloaded, so you've got to go and manually un-flag them again - it's a PITA when this happened whilst I was on holiday, as I had to go through every folder and work out which ones they were.

My 'search' parameters include the hauth=1 line (see below)

order=desc
u_nfo_posts_only=0
emu_subcat=-1
sort=ps_edit_date
fpn=p
u_completions=1
emu_subcat_done=-1
u_comment_posts_only=0
u_url_posts_only=0
hauth=1

I've done a clean install of Episode Butler too, and still the problem persists - a couple of episodes downloaded correctly, but it soon came back.

[tedrock]

Janoz fix DID work but they changed something on newzbin again and now the fix doesn't work and i'm pretty sure removing the hauth doesn't help either.

[pspgimp]

No it doesn't work if you remove hauth again.

I really need to get this working as I am away in a week for a month and the wife will want it to be working so her episodes are downloaded correctly.

Newzbin support are just ignoring me now !!!!

[pspgimp]

Apparently a Newzbin coder is still looking at this according to an update on my support ticket with them.

[Janoz]

Sorry everyone for the late response. I was off the grid this weekend because of a bachelor party and have been very busy at work (Last week of an iteration).

Based on the responses I get and the logfiles I received the error occurs when a lot of searches are done. There are in fact two problems.
1- when a lot of feed requests are done, newzbin presents a login page and wont return a feed result anymore. The Episode Butler sees this as a fatal error and aborts the search (with the now famous error message)
2- When the search is aborted all the already found episodes won't be send to Sabnzbd or other downloaders, although they are marked as downloaded.

This evening I'll work on both problems. First I'll make sure that found episodes are really send to sabnzbd when they are marked as downloaded. This way, even if the search aborts, all the already found episodes will still be downloaded. The Episode Butler will than at least be a little usable again.

After that I will focus on the login error when requesting feeds. I think I'll make a test program to trigger the bug and supply this to the developers at newzbin so they can reproduce the problem.

I hope I'll get a more usable Episode Butler version out by the end of this day. Sorry for the inconvenience.

[alan]

Hi Janoz, hope you had a great weekend 

I'm wondering how much success you'll have with the guys at NewzBin, as it appears that you're not actually using one of their API's - perhaps it would be better to switch to one of them instead.

It does appear that the API in question http://www.newzbin.com/api/reportfind/  (documentation http://docs.newzbin.com/index.php/Newzb ... ReportFind) - whilst it doesn't list all of the parameters, you can still use things like 'group', etc however some things have changed like file size limits are measured in bytes rather than MB.

I've done a little testing with the reportfind api, and it does appear to work reliably 100% of the time.

I know it's probably more work, but might be worth it in the long run.

Alan

[Janoz]

Hmm, this API is rather new (went public last Januari) and I wasn't aware of it. Until now I have used the feed urls. I will take a look at this tonight. I think, since the parameters are the same and a tab separated file doesn't realy differ from a comma separated file, the implementation wont really be that much work.

Thanks for pointing that out to me!