[Afpfs-ng-devel] accented characters in UTF-8 filenames
Status: Alpha
Brought to you by:
alexthepuffin
From: Michael U. <mu...@re...> - 2007-05-15 19:40:28
|
Hi Alex, Hi list, from what I've learned today MAC OSX volumes (HFS+) use decomposed UTF-8 characters in file and directory names (base character and accent are not encoded as a single multibyte character [precomposed Unicode], but instead the base character stands for itself and is followed by a multibyte "accent only" character [decomposed Unicode]). http://developer.apple.com/qa/qa2001/qa1173.html gives some comments on this topic. When mounting a OSX volume from a linux box via afp-client, I get a correct listing of volume names in a terminal set to UTF-8 charset, even if accented characters are contained in the names. Saving this listing output to a text file and trying to convert this file from UTF-8 to a different charset (e.g. ISO-8859-2) with recode or iconv leads to 'invalid input' error messages. Maybe I'm completely wrong here and somehow don't see the obvious ... OTOH I feel that there should be some decomposed -> precomposed conversion in afpfs-ng (or somewhere else on the OSX side ...?) to deliver filenames in 'standard' (precomposed) UTF-8, which are correctly coded for further processing on the linux side. Please let me know, if I shall provide further information or give any other assistance. Best regards ... Michael |