Decoding of path parameters does not work correctly for some non-english characters

Requested URL:
http://localhost:8888/search/éa

//Encoded into :
http://localhost:8888/search/%C3%A9a

http://www.w3schools.com/TAGS/ref_urlencode.asp
(see title : URL Encoding Functions )

Although the "URL Encoding Reference" shows each char is mapped to some value in the form of '%xx'
and é is shown to be %E9 whereas the encoding function input field converts é to %C3%A9

Same thing happens in java as well
(org.apache.commons.httpclient.util.URIUtil) using this one from commons-httpclient-3.1

Decoded in RestEasy into ;
/search/éa

Encode.decodePath(String path) use a

Pattern encodedChars = Pattern.compile("%([a-fA-F0-9][a-fA-F0-9])");

which is not always correct, some chars contain 2 %'s which causes the problem like in this case