MDQS and lpr

[Phil Budne]

In article <2590 at randvax.UUCP> urban at rand.org (Mike Urban) writes:
>Are there any MDQS shops out there that have successfully and
>transparently merged lpr and MDQS?

At one point I tried to use MDQS to drive a pair of PostScript
printers from one queue (before PLP was around).  I wrote to hacks;
bsdlpd which listens on the lpd port and queues jobs to MDQS and
bsdserver which contains code ripped from the guts of lpd's printjob.c
and uses printcap (and associated) filters to print jobs.

Whether I did so "successfully and transparently" is another question.
I have made the stuff available, but remember YOU GET WHAT YOU PAY
FOR.

The code was never used in production and if my (cloudy) memory
serves, has outstanding problems when it comes to UIDs and the
tradeoffs of being just like lpd (and the MDQS daemon not being able
to signal the bsdserver) and fitting into the MDQS paradigm (and not
being able to, for example send jobs to remove lpd's because the
bsdserver can't bind a low numbered port).  No doubt security problems
pervade.

AND since the code in bsdserver derrives from 4.3T lpd it will no
doubt object to the chicanery pulled by (for example) InterLeaf (which
insisists on issuing prints of symlinked files which it subquently
fools with!)

The files reside on BU.EDU in ~ftp/src/PRINTING/MDQS/Part0[1-3]

AGAIN! Caveat Emptor!  I did the work before I had come the the
inescapable conslusion that no reasonable Un*x spooling system exists.
Today I would shrug and say "you can't get there from here". Ahh for
youth's enthusiasm!!!

	-Phil Budne, Boston University