Hello, I’m parsing Kraken json objects in Rainmeter to have a list of streams for certain games directly on my desktop. A few days ago, the script broke. After investigating I noticed that the ordering of keys changed which caused the problem (I read in another topic that the order of keys isn’t guaranteed, which is a bit annoying, but since it doesn’t change too often I can live with it).
However the bigger problem is that since the ordering change I also seem to have encoding problems. Wheras before Korean letters would be displayed perfectly fine in my skin I now only get unicode codes (like \u041f\u0420\u041e etc). Any idea on how I can fix this?
This is the URL I pull with Rainmeter (json object gets saved to disc): https://api.twitch.tv/kraken/streams?game=League+of+Legends
This is the regex I use to extract the desired information I want to display in Rainmeter (right now I have it set so the Top 8 viewer streams are displayed):
RegExp="(?siU)(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",").*(?(?=.*"channel":).*"display_name":"(.*)",".*status":"(.*)","url":"(.*)",")
(The lookahead checks here are of course unnecessary for LoL, since there are always more than 8 streams, but I use it for other games aswell, eg Dwarf Fortress where there are not always 8 streams and using the regex without lookahead checks would break the entire thing).
Any help would be appreciated
Update: I circumvented the problem by using xml list instead: http://api.justin.tv/api/stream/list.xml?category=gaming&meta_game=League%20of%20Legends
The xml list uses HTML entities instead of unicode codes. Rainmeter can auto-decode the HTML entities to restore Korean and other special chars, so my problem is solved now. Still wondering why it broke to begin with.