I see there is a fix on Piwik 0.2.33, FIXED #589 Piwik fails to properly decode and store some chinese keywords (eg. from baidu.com).
But I still see some url with chinese keywords are decoded wrong.
take below link for example, the keywords are ?, but in piwik they become "?", see also in attached screenshot file.
Attachment: screenshot about chinese words decode
There's a new featuren in #2761 that allows multiple encodings. We can try adding utf-8 to the baidu configuration (currently expects gb2312) and edward's url to the unit test.
thanks for the quick feedback!
here are 2 more examples, the first one was decoded right, but the second and third one were not.
IT sounds like new logic might need to be introduced for baidu (use UTF-8 when it is found as a parameter value, default to gb2312 otherwise?)
I was wrong, that's good! :)