+ Printing from W95 almost works. It queries the queue and thinks it is
busy. Have to figure out why.

+ NT: not tested.

+ Other SMB clients too?  Will people out there please report results,
positive or negative, on other clients please?

Here are known problems with smbpd 0.94 which will hopefully be fixed in
a later release:

+ Need to behave more like a B node, act on more NS messages. At least
defend own name.

+ Generates lots of debugging messages to the console. At the moment
this is a good thing. Later provide more control over debugging level?

+ Need to do buffer limit sanity checks in pack_struct.

+ Lots of general cleanup.

+ Maybe make all the signatures of the SMB action routines uniform,
then use a dispatch table instead of a case statement?

+ This server blocks other SMBs from the client while printing a block.
(Other clients are not affected.) Does this cause problems for the
client? What if the client times out?

+ What do we do if the printer has a hardware problem during printing,
e.g. paper out?  Block and wait for operator to fix or signal an error
on the command and close the print file?

+ Low priority: The way the time is computed in the negotiate_protocol
reply means it may be up to 7 minutes slow. Also it's the time the server
was started. As far as I know, the client doesn't do anything important
with the time, so I'm not in a hurry to fix this.

Things that probably won't be fixed:

+ Job sizes are cumulative printed bytes, not job size. So it will keep
increasing until end of job. Can't display job size because we don't
know the job size when the printing starts (remember smbpd doesn't have
local spooling). I don't think this can be fixed.

If you encounter a problem not listed above:

Please give as much detail as you can, including at least the client
software (e.g. WfWg), version, and any messages on the console. In
particular note any Unimplemented command messages from smbpd.

Mail the info to ken@syd.dit.csiro.au. Thanks.
