tor, 03.02.2005 kl. 17.42 skrev Remco Poelstra:
Hi,
Let me first introduce myself, sine you people probably never heard
about me :).
I'm a student at the Eindhoven University of Technology. With a team of
4 students we are responsible for Linux support at the faculty of
Electrical Engineering.
Tim Waugh asked me to help thinking about a new printing solution, as a
reply on a mail of me on the config list.
I read the posts here about that subject and I would like to share my
opinion on this matter:
First I would like to state that there won't be any solution if CUPS
doesn't improve.
The problem that eggcups tries to adress (authentication) isn't solved
if there is a chain of CUPS servers involved. If CUPS can't communicate
back, then the user still can't authenticate. So CUPS should implement a
2-way IPP protocol.
Then eggcups can be reduced to listening to HAL messages in order to
display the print queue and pop up authentication dialogs. By not
implementing a full CUPS server, it has less system resource usage,
which is a good thing I think.
Second if CUPS supports signed browse packets, than we have an 'easy'
solution of identifying fake printers on a large network.
For the configuration part: gnome-cups-manager does quite a nice job on
this.
Than we only need some decend tool to configure the server itself.
Just my 2 cents,
Remco Poelstra
Another part which would make it much easier to configure for smaller
networks (ie. those without a local DNS server), and avoids disasters if
hosts named "localhost" shares printers - do a test on a recived
hostname immediatly after it is recived, check that it does look up the
correct IP - or an IP at all.
If it does not resolve - use IP adress directly instead.
Kyrre