Hi,
The man page of the curl_unescape function say :
----------
All input characters that are URL encoded (%XX where XX is a two-digit
hexadecimal number, or +)
       will be converted to their plain text versions (up to a ? letter,
no + letters to the right of a ? letter will be
       converted).
-----------

But the function code actually don't touch any "+" character, before or
after the "?" character.It only look up for "%XX" encodings.

Whatever is correct, the doc or the code should be made consistant.

Anyway, thanks for your cool and powerful library !

-- 
Renaud Duhaut <rd_at_duhaut.com>