[jcifs] URL encoding

[Michael B Allen]

On Sat, 09 Mar 2002 12:32:56 +0900
"Talbot David" <chukhonets at hotmail.com> wrote:

> 
> 
> >What "standard windows client"? The smbclient program will not negotiate
> >Unicode (actually the latest and greatest might but I don't think vanilla
> >2.2 does).
> 
> I traced the packets generated when the same directory is accessed via 
> "Network Computer" on Win 98.

Win98 does not negotiate Unicode (for most things). Does this machine
properly exhibit the behavior you seek?

> > > and the non-ascii characters seem to have their native
> > > encoding values. I suppose treating the path as a byte array should 
> >work.
> > > I'll give it go.
> 
> 
> >It sounds like the server is negotiating ASCII (a.k.a the platform
> >dependent 8 bit codepage). If this is indeed the case you are in uncharted
> >territory.
> 
> Probably not the right territory for me just yet ;)
> 
> By the way which class actually has the job of converting the url string 
> into bytes that make up the SMB request packet? I assume that's where I'd 
> need to start.

The codepage thing has nothing to do with jCIFS. On the other hand I
can't help but think that's NOT the problem because it is only used to
support foreign character sets. For example cp866 is Cryllic Russian:

  http://czyborra.com/charsets/codepages.html#CP866

so if you were trying to support Hebrew or Nordic Win 98 clients or
somthing that's probably something you would have mentioned already
right? Otherwise, again, you're just not encoding the paths properly
... try the SmbCrawler.

Mike

-- 
May The Source be with you.

#################################################################
#################################################################
#################################################################
#####
#####
#####
#################################################################
#################################################################
#################################################################