Chinese-related problems introduced in 0.2.4

[RobinQu]

1. Decoding mail headers with Chinese may throw exceptions. And little we can do to reproduce the problem, as it seemingly occurs at random.

2. Fail to decode quoted strings which were correctly handled. e.g.

   "=?gb2312?B?ob67otDh1OexqKG/?="

Last working version we used is v0.1.8.

[andris9]

mimelib 0.2.x is a complete rewrite so there is no single change that could have broken the implementation. If you could provide me with some more information about the issue (which headers throw the exceptions etc.) then I could look at it but currently I have no clue what might cause it. As far as I tested, mimelib converts "=?gb2312?B?ob67otDh1OexqKG/?=" to '【虎嗅早报】' with no errors.