Hello Julian; thank you for your report.
On Sun 22 Oct 2017 at 09:20:28 +0100, Julian Gilbey wrote: > Package: cups > Version: 2.2.5-2 > Severity: important > > (I'm labelling this as "important" because it's wasting large amounts > of paper.) > > This is a weird bug, and I have little idea how to track down the > cause. > > About a week or so ago, on my Debian testing system, printing commands > started producing three copies of every page on my local printer (so a > 2-page document would come out at 3 copies of page 1 followed by 3 > copies of page 2). This is the case in at least all of the following > circumstances: Can this changed be pinned down to any particular version of cups or cups-filters? Or something you altered in cupsd.conf (for example) at that time? > * Connect the printer with a USB cable, print via usb://... > * Connect the printer with via ethernet, print via dnssd://... > * Print from evince > * Print using lpr (eg: echo hello | lpr) > * Going to http://localhost:603/ and doing Maintenance -> Print Test Page > > However, going to http://localhost:603/ and doing Maintenance -> Print > SELF Test Page only produces a single page of output! A self-test page is actually produced by the printer itself; it is often done from a button on the printer. The filtering system is not involved. Did you really use port 603 to bring up the CUPS web interface? You also mention Windows and Mac machines. I am not familiar with printing from either of these, but my understanding is that Windows does all the processing and sends the completed job to a CUPS server, which does no processing. That is, the filtering system is bypassed. I wonder whether your Mac is doing this too. > I have updated cups to the latest version today (2.2.5-2), and that > has made no difference, and nor has deleting the printer from CUPS and > reinstalling it. The PPD files have not changed in about a year > (except for the uninstalling and reinstalling, and even that installed > an identical copy to the existing PPD file). > > So I'm at a bit of a loss as to the cause of this. I also don't see > anything obvious in the log files, but I may be missing something. > > Any suggestions? We'll need the printer model, the PPD used and an error_log. Let's standardise on doing lp -d <print_queue> /etc/nsswitch with a USB connection to the printer. https://wiki.debian.org/DissectingandDebuggingtheCUPSPrintingSystem#The_CUPS_Error_Log gives you a way of getting the error_log. It compresses very nicely with gzip to make it acceptable to be sent to debian-printing as well as the bug. Would you also look in /var/spool/cups before and after printing and report on any change in the "d" files there, please. Regards, Brian.