Using CUPS to print to a file instead of an actual printer is a great idea for troubleshooting and archiving purposes. This saves trees but, most importantly, saves you the walk to the printer. So, how does one go about setting this up? I am assuming you already have working CUPS server all setup and properly tested. Here’s what you need to do:

Shutdown cupsd via /etc/init.d on Linux and older Solaris versions or via “svcadm disable” on Solaris 10+.  Add the following new printer to /etc/cups/printers.conf as the very first printer, before any other <printer></printer> entry.

<Printer fileprint>
Info FilePrinter
DeviceURI fileprint.sh
State Idle
Accepting Yes
Jobsheets none none
QuotaPeriod 0
PageLimit 0
KLimit 0
</printer>

Create a directory that will store your printouts:

mkdir /var/cups_printouts ; chmod 777 /var/cups_printouts

Create this script: /usr/lib/cups/backend/fileprint.sh with the following content:

#!/bin/bash
cat $6 > /var/cups_printouts_`date +'%Y-%m-%d_%M%H%S'`.print

Make sure the script is executable by non-privileged users:

chmod 777 /usr/lib/cups/backend/fileprint.sh

In case you’re wondering what “$6” is, here’s a list of CUPS command-line arguments:

# $0 - name of the script
# $1 - job number
# $2 - user
# $3 - filename
# $4 - number of copies
# $5 - ?
# $6 - spooler file name

Start cupd and try printing something:

lp -d fileprint /etc/hosts

Check in /var/cups_printouts and, if everything went as expected, you should have a *.print file waiting for you there. Hopefully, this makes your life slightly easier.