We're using the Samba client 3.0.0 API to get file server access. We have some Japanese Win2k servers that we test with. This server has some share names that are in Japanese. When I call cli_RNetShareEnum against this server, the names passed to my enum function are in the Shift_JIS character set. This is even though the CAP_UNICODE capabilities flag is set after cli_negprot is called. This causes the pull_ascii funtion to mash the names when it converts from CH_DOS to CH_UNIX inside of convert_string. For files with Japanese names, the Samba API properly passes the names in Unicode. To summarize, there's actually two problems: 1. The Unicode flag is set even though the share names are in Shift_JIS. 2. The names get mashed under the convert_string function hierarchy using the internal character conversion routines (utf8_pull, etc.). We do not have ICONV (or GICONV) support compiled in. I searched bugzilla for anything relating to this but found nothing.
I was able to get around this by using the RPS NetShareEnum call. RPC uses Unicode. I don't have authority to change the bug status. Feel free to close it however you see fit.
closing. Lot of mb string work done since original 3.0.0 release as well.
sorry for the same, cleaning up the database to prevent unecessary reopens of bugs.