redirect print job to file and process - how?

[Gary]

I want to share a printer queue to receive a PCL job and save the job as
a file. The name of the file containing the job is then passed to a
process that manipulates the PCL, modifying certain pieces. The “new”
print job saved as a file and is optionally later sent to another queue
for printing.
I can do this in Windows fairly easily using a few tools such as RedMon.
It seems that it would be even simpler to do this under Linux, but I am
very new to Linux and I don’t know how, nor where to look for
information as to how.
I can locate the jobs in the queue, so I could presumably write a script
to pick them out and do what I want to them. I am not sure if I can tell
the spooler the job is gone, or whether I need to. I know that if I do
it this way to Windows it will eventually cause problems.

Thanks for advice.

[Lew Pitcher]

Post by Gary
I want to share a printer queue to receive a PCL job and save the job as
a file. The name of the file containing the job is then passed to a
process that manipulates the PCL, modifying certain pieces. The “new”
print job saved as a file and is optionally later sent to another queue
for printing.
I can do this in Windows fairly easily using a few tools such as RedMon.
It seems that it would be even simpler to do this under Linux, but I am
very new to Linux and I don’t know how, nor where to look for
information as to how.
I can locate the jobs in the queue, so I could presumably write a script
to pick them out and do what I want to them. I am not sure if I can tell
the spooler the job is gone, or whether I need to. I know that if I do
it this way to Windows it will eventually cause problems.

There are likely many different ways to do this in Linux.

FWIW, I'd tackle it this way:

1) write a script to process the PCL and create a file

2) create a printer queue and printer within the spooler that, instead
of printing the data, runs the script on it. This is fairly easy to
do in lpr or lprng (I've done it before), but I don't know if CUPS is
as easy do set up.

3) set up a process (you define how it is invoked) that selectivly
spools transformed files to a real printer. Likely, just a script
that 'lpr's the file, but to printer instead of the pseudo-printer
from step 2.

How it works
a) user queues output to pseudo-printer queue
b) pseudo-printer reads output, process PCL, saves new PCL in file.
pseudo-printer /does not/ print anything, but just empties the queue
c) user runs 'selective print' script to choose one of the saved files
d) 'selective print' script queues file as output to real printer queue
e) real printer reads output, processes and prints it



- --

Lew Pitcher, IT Specialist, Enterprise Data Systems
Enterprise Technology Solutions, TD Bank Financial Group

(Opinions expressed here are my own, not my employer's)

[Gary]

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

Post by Gary
I want to share a printer queue to receive a PCL job and save the job as
a file. The name of the file containing the job is then passed to a
process that manipulates the PCL, modifying certain pieces. The “new”
print job saved as a file and is optionally later sent to another queue
for printing.
I can do this in Windows fairly easily using a few tools such as RedMon.
It seems that it would be even simpler to do this under Linux, but I am
very new to Linux and I don’t know how, nor where to look for
information as to how.
I can locate the jobs in the queue, so I could presumably write a script
to pick them out and do what I want to them. I am not sure if I can tell
the spooler the job is gone, or whether I need to. I know that if I do
it this way to Windows it will eventually cause problems.

There are likely many different ways to do this in Linux.
1) write a script to process the PCL and create a file
2) create a printer queue and printer within the spooler that, instead
of printing the data, runs the script on it. This is fairly easy to
do in lpr or lprng (I've done it before), but I don't know if CUPS is
as easy do set up.
3) set up a process (you define how it is invoked) that selectivly
spools transformed files to a real printer. Likely, just a script
that 'lpr's the file, but to printer instead of the pseudo-printer
from step 2.
How it works
a) user queues output to pseudo-printer queue
b) pseudo-printer reads output, process PCL, saves new PCL in file.
pseudo-printer /does not/ print anything, but just empties the queue
c) user runs 'selective print' script to choose one of the saved files
d) 'selective print' script queues file as output to real printer queue
e) real printer reads output, processes and prints it

That's exactly what I intended to do. What I am hoping for is some hint
as to how to do it, or where to find documentation as to how.

Post by Gary
Specifically, I don't see how to set up a printer queue to output a

print file to disk and then trigger a script, like RedMon does for Windows.
I do see that under Redhat, CUPs normally runs GhostScript as a filter
on a spooled file... This is 90% there, and it seems logical that I
could substitute my own program (or script) for GhostScript. But how?
And will this gracefully remove the job from the queue?

An example that illustrates what I am afraid of: In Windows, if I just
put a queue on hold and grab the print job out of the queue folder,
deleting it when I am done, this works perfectly for a long while. But
the paused Windows queue does not forget the "pending" job even when the
job is deleted by another process, and it has a hissy fit when the
spooled job count exceeds four digits. So, after I have processed 9,999
files, undefined bad stuff happens. It's possible to do it that way only
if I also implement some mechanism to clear the queue counter as well as
the actual queue.
I am afraid that in Linux if I just do something that seems to work, I
will set a time bomb for myself.