smbwall

[David Lee]

On Fri, 25 Jan 2002, Garry J. Garrett wrote:

> Torsten Curdt wrote:
> > 
> > [snip]
> > 
> > > It is good to see that there is someone else who wishes to have this sort
> > > of functionality associated with Samba.
> 
> 
> Apparently it's already in Samba...  Keith Farrar pointed me to
> the "examples" directory where there is a "misc" directory that
> has a "wall.perl".  I haven't tested it yet, but it appears to
> be the same type of thing that my shell script does, only written
> in Perl.  Keith said that it's been included in Samba since 1.9.
> I had never looked in "examples" because primarily it seems to
> contain examples of configuration files (smb.conf, etc.).

But my proposal to avoid such "specials".  Rather that the UN*X "wall",
"write", etc. should simply work, automatically, to SMB clients (should a
site so choose). 

> My gut reaction to this is, if smbd were to allocate a pseudo terminal
> when it starts up, then it would get the "wall" message when it is
> sent out and could pass along that wall message as a windows pop-up
> message.  Of course, this makes a lot of sense if you run samba as
> a stand-alone daemon.  ...

That's the idea.  Hold onto it!  (But there is debate, understandable,
concerning the implementation detail "if smbd were to allocate...".  See
another message on this.)

>                   ...  If you run Samba out of inetd, then wouldn't
> each instance of smbd that is fired off by inetd end up soaking up
> it's own pseudo terminal? (seems wasteful).

Soaking?  It would just be another file descriptor in smbd's main process
loop "select()".


> If this is your desire, the perhaps it makes sense to allocate a pseudo
> terminal per drive mapping?  Then you could send a wall or a issue a
> write to the terminal associated with the drive mapping and Samba
> would turn around and send it out as a windows pop-up.

Again, that's the general idea, in principle.

There is at least one particular implementation detail that worries some
of the Samba Team.  It is important that we pay attention to their
concerns and to their advice. 

In another message, sent a few minutes ago, I tried to address (a) the
functionality I am trying to achieve, and (b) some implementation
possibilities.  I am trying to keep discussion of those reasonably
separate (obviously they overlap!), because there is a danger that a
controversial implementation detail might jeopardise the general
principles. 

Hope that helps.

-- 

:  David Lee                                I.T. Service          :
:  Systems Programmer                       Computer Centre       :
:                                           University of Durham  :
:  http://www.dur.ac.uk/t.d.lee/            South Road            :
:                                           Durham                :
:  Phone: +44 191 374 2882                  U.K.                  :