Architcecture for winbindd client character conversion.

[Gerald (Jerry) Carter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

simo wrote:

>> Just to mention a point, Jeremy's proposal assumes a
>> similar boundary for string conversion as we use for
>> incoming CIFS packets.  Instead of the wire being the
>> boundary, the incoming winbindd pipe request is the
>> boundary.

You didn't answer this question.

>> And while it may be a really bad idea for Unix to allow
>> per user locale settings, the fact is that it does.
>> All this discussion really just seems hinged on the
>> point that it is a bad idea to do this.  But we have fixed
>> so many issues where it was a bad idea in Unix/Windows,
>> what makes this one different?
> 
> The difference is that in this case we can't fix nothing 
> really by allowing multiple charsets per session.
> 
> The filesystem is one, and if you allow people to 
> login with different locales it means you are allowing people
> to make a mess in the filesystem where some people write
> files in utf8, some other in Japanase charsets and some
> others in iso8859-15.

Got to say I'm really surprised at the push back on
this one.






cheers, jerry
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFG1INlIR7qMdg1EfYRAqDCAJ9feBULaHzkt//cCsZXP7h9tnUgCACgm7yt
Yhti6bXg5+HZ9/b0Vco8loA=
=XJ7X
-----END PGP SIGNATURE-----