ComingSoon URL retrieval error

[Omertron]

Original issue 2736 created by Omertron on 2014-04-30T18:04:27.000Z:

I'm using r3888 (thanks for fixing the previous encoding problem so fast). For some films only, the ComingSoon plugin correctly finds the "key" of the film searched, but erroneously compose the URL to fetch the data from.

For example, I have "Being John Malkovich" film. YAMJ searches for it and correctly finds it's associated to key "3814" on the ComingSoon website. However, it tries to open the following address:

http://www.comingsoon.it/Film/Scheda/?key=3814+%22Being+John+Malkovich%22+site%3Acomingsoon.it

It returns an error because all the words after the key number aren't needed. In fact the following trimmed address works fine:

http://www.comingsoon.it/Film/Scheda/?key=3814

To complete my report, this is what shows up in moviejukebox.ERROR.log:

ERROR com.moviejukebox.tools.WebBrowser - WebBrowser: Error getting URL http://www.comingsoon.it/Film/Scheda/?key=3814+%22Being+John+Malkovich%22+site%3Acomingsoon.it, Server returned HTTP response code: 500 for URL: http://www.comingsoon.it/Film/Scheda/?key=3814+%22Being+John+Malkovich%22+site%3Acomingsoon.it

[Omertron]

Comment #1 originally posted by Omertron on 2014-04-30T18:42:59.000Z:

The comingsoon layout has changed, perhaps thats the reason.

Of interest is a snippet of the log file, to see:

• where has been search
• which is the retrieved key
• which search engine has been used

etc.

[Omertron]

Comment #2 originally posted by Omertron on 2014-04-30T19:39:08.000Z:

This error happens only for some films. I couldn't identify which differences determines its appearance. I can only say that it's consistent, not random: always the same films are affected.

The only way I found to avoid this problem is by adding a NFO file in which I manually specify the field key. This should exclude that's a "changed layout problem" even if the logs say so (ComingSoon plugin has been updated in r3880 for that reason). I think the problem resides where the URL is being composed, and since it's being wrongly composed, the resulting page is non-existent.

I'm attaching a snippet of the main log file, hoping it could be useful. Thanks!

[Omertron]

Comment #3 originally posted by Omertron on 2014-04-30T20:47:13.000Z:

Okay ... I think the ID in the XML is wrong due an older scraping, which could have been errorness due site changes or searchenginge changes. So it's needed to rescrape a correct id.

This could be solved by deleting the generated files for "Being John Malkovich" in your jukebox folder.

The genres now will be correctly scraped with latesr release; I've just fixed that.

Please delete the jukebox files, rescrape your library and check what happens.

[Omertron]

Comment #4 originally posted by Omertron on 2014-05-01T00:15:46.000Z:

Done as you said. All went very well, no more errors (used r3895). Thank you for the good work, such frequent updates are very welcome!

[Omertron]

Comment #5 originally posted by Omertron on 2014-05-01T01:47:19.000Z:

<empty>