StringByAddingPercentEscapesUsingEncoding encoding differently

Question

Hi,
I have an Objective-C function that URL encodes data to be sent to a Dav server. Sometimes it encodes properly, sometimes it doesn
NSString *URLEncode(NSString * url)
  NSString* sUrlCopy = [[NSString alloc] initWithFormat:@"%@",url];
  return [sUrlCopy stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding];
sometimes if I call it as URLEncode(@"/XæøåX.txt");
I'll get the proper encoding of ENCODED=/X%C3%A6%C3%B8%C3%A5X.txt
Sometimes though, I get back ENCODED=/X%C3%A6%C3%B8a%CC%8AX.txt
Decoding both of these gives me the same source though.
does the "ø" character have different unicode values? how can I get a consistant URL encoded value to send to web servers?
Thanks!

Answer 1

Hi,
Yes our program is breaking because of this. Our filesystem will retrieve a list of files from the DAV server, those filenames are UTF-8 encoded as they come over the wire from the server based on the unicode composition of the characters on the remote NTFS drive. We decode the filenames returned from the server and then return them to OSX. At some later point in time, OSX (Finder) will need to know the attributes of a file, so it calls in to us through attributesOfItemAtPath and will send in the decomposed version of the precomposed filename that we gave back to them moments earlier.
The problem happens becuse the DAV server has returned us a UTF8 encoding based on the precomposed filename as it lives on the server. We've decoded it and returned the precomposed filename to the OS. Somehow, from the time we've returned it to the OS and the time the OS sends it back into us for more processing, it has now become a decomposed filename.
Since we honor the filename composition, we UTF8 encode it, send it to the DAV server, and the server responds with File Not Found, because NTFS was doing a lookup on the decomposed version instead of the precomposed version.
Granted it could be a server problem and the server should figure it out, but OSX is doing something with the precomposed filenames it receives. It's decomposing them and sending them back to us.