Added URL decoding to fix #249

[rrkpp]

Added URL decoding before file upload to avoid
System.IO.PathTooLongException when using the URL uploader with an
encoded string

[Jaex]

Your code can stuck in loop forever.

[rrkpp]

Sorry, under what use case could an infinite loop occur? Just so I don't make the same mistake again. Thanks

[Jaex]

Also doing loop like that can be problem like if file name is %50.jpg then it will decode it too, isn't it?

[rrkpp]

As far as I can tell it's impossible to support multi-encoded URLs while simultaneously protecting against pseudo-encoded file names. Either you loop to support multiple encoding and break files like %50.jpg or you give up the ability to decode URLs that have been encoded multiple times. By definition %50.jpg is an encoded URL which makes it virtually impossible to detect as being a decoded string. I think Unicode filenames like the G+ example and multiple encoded strings which are widespread on many hosts are a more popular use case than %50.jpg style names.

[Jaex]

I was wrong about infinite loop sorry.

But I don't think url decoding download path is good idea. File name was from url so I will just url decode url first then get file name from it to make download path so won't need second url decoding.

Edit: Actually url decoding for url is not required at all? Just file name decoding for download path is needed for issue #249

Edit2: Fixed issue now.