Is there a fix for Windows non-ASCII filenames via smbfs?

[Jason Haar]

I've just started delving into rsync again - we use it via Linux servers to
synchronize Windows data over a WAN via smbfs.

i.e.

Windows --> Linux == Rsync == Linux --> Windows

Basically there are a bunch of "8 bit" filenames on the Windows server that
rsync (or "find", or "ls") cannot see: rsync reports:

rsync: readdir(Agiting/11718_1?2 Ag ad 11.99_Folder): No such file 
 or directory (2)

The question-mark can be anything (I say "8-bit char" - you say "unicode", I
say potato,...). To be honest I think they're primarily fat-fingered
mistakes by English-speaking Windows developers - but it doesn't change the
fact that these are valid Windows filenames, but cannot be accessed by
standard Unix filesystem calls. We are talking "English installs" of Window
2000 servers here BTW.

I have tried mounting these Windows share a variety of ways, 
"-o iocharset=iso8859-1,codepage=cp850" doesn't help, and "-o unicode" just
doesn't work - I either get empty shares, or shares consisting of single
letters where filenames should be(!!)

I must be doing something wrong - this can't be so unusual. So can anyone
suggest what magic options I should be using to get these Windows filenames
to show up under Linux?

This has been tested from RH-7.1 (2.2.8a) to Fedora Core 2 (3.0.7) with the
same results. 

I first reported this problem in 1999 and am amazed I'm still seeing it.

Thanks!

-- 
Cheers

Jason Haar
Information Security Manager, Trimble Navigation Ltd.
Phone: +64 3 9635 377 Fax: +64 3 9635 417
PGP Fingerprint: 7A2E 0407 C9A6 CAF6 2B9F 8422 C063 5EBB FE1D 66D1