proposal for changing name mangling.

[Andrew Bartlett]

Michel Stoop wrote:
> 
> I have again ran into troubles with duplicate mangled names under W2k.
> We have some model libraries wich contain more than 4000 files
> with simmilair names. (like gatkr160x400.mod).
> Now when I type in a cmd prompt: del gat*.mod I get some errors.
> After some research I saw that the delete command builds it file list
> on the mangled name. I assume to save memory...
> What I also saw that when I copy these files to a NT server I get
> mangled names consisting of the 2 first characters in uppercase, then
> a 4 character hex value, then the mangling char and a number.
> This gave me the idea to change the way samba mangles names to match
> what W2k produces. So I did some work on smbd/mangle.c
> With the currrent name mangling of samba I got up to 6 equal mangled
> names for different files, but now I get only unique names.
> 
> So in the current version the file gatkr160x400.mod would be mangled
> to GATKR~9 at .MOD, but now it becomes GA494F~6.MOD
> Now I have included a diff file for samba-2.2.3a and want to ask if someone
> wants to look at it. And I want to know what the samba team thinks about
> changing the way names are mangled and how it should be implemented.

HEAD has a compleatly new name mangling scheme - if this is an area that
interests you we could really do with some help tracking down some bugs
in it.

Andrew Bartlett

-- 
Andrew Bartlett                                 abartlet at pcug.org.au
Manager, Authentication Subsystems, Samba Team  abartlet at samba.org
Student Network Administrator, Hawker College   abartlet at hawkerc.net
http://samba.org     http://build.samba.org     http://hawkerc.net