Retain cache for thetvdb longer

At the moment XBMC will load the thetvdb data from the scraper and just keep it until the scan process is finished. In the past weeks there have been some longer and shorter outages to the thetvdb site which made scraping impossible.

I think most users have a similar usage scenario, where new episodes of tv shows are added about every week. Adding entirely new series is probably less frequent than that.

The proposal is to keep the thetvdb data in the cache even after the scan. On a scan XBMC would try to first load the data from the scraper provided URL but can use the cache data, if the page is offline.

Since currently running shows usually have the data for a few episodes in advance, this would make it possible for XBMC to scrape the data when the page is down (or network is offline). The only thing missing would be the thumbnail but since that can be automatically generated XBMC would stay usable.

What do you guys think - any pros/cons I didn't see?

FYI, spiff committed a persistentcache option to use inside scrapers. TheTVDB is the only one using it right now. Defaults to 00:15.