On Sun 22 Mar 2026 at 13:51:13 (-0400), Haines Brown wrote: > On Sun, Mar 22, 2026 at 02:59:30PM -0000, CGS wrote: > > On 2026-03-22, CGS <[email protected]> wrote: > > > On 2026-03-22, Haines Brown <[email protected]> wrote: > > >> On Sat, Mar 21, 2026 at 02:50:17PM -0000, CGS wrote: > > >> > > >>> Is this driverless printing or is there a ppd file? > > >> > > >> Driverless > > >> > > >>> You said '-o > > >>> orientation-requested=3' is ignored? The thing is to verify whether or > > >>> not that orientation request is being passed on to CUPS or not in order > > >>> to narrow things down. > > >> > > >> How can that be verified? It is logged somewhere? > > >> > > > > > > lpr -o orientation-requested=3 file.txt > > > lpstat -W all > > > # grab job ID, then: > > > sudo strings /var/spool/cups/cN | grep -i orient > > > > > > I think that should work. > > When I ran lpstat -W all I hot this: > > HP_LaserJet_Pro_M428f_M429f_8264A8-37 haines 1024 Sun 22 Mar > 2026 06:22:14 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-38 haines 46080 Sun 22 Mar > 2026 08:01:04 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-39 haines 395264 Sun 22 Mar > 2026 10:26:01 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-40 haines 1024 Sun 22 Mar > 2026 10:27:59 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-41 haines 1024 Sun 22 Mar > 2026 10:29:35 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-42 haines 1024 Sun 22 Mar > 2026 10:30:07 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-43 haines 1024 Sun 22 Mar > 2026 10:31:26 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-44 haines 0 Sun 22 Mar > 2026 10:33:49 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-45 haines 1024 Sun 22 Mar > 2026 10:34:18 AM EDT > HP_LaserJet_Pro_M428f_M429f_8264A8-46 haines 1024 Sun 22 Mar > 2026 01:40:15 PM EDT > > I don't see any IDs here.
They're the numbers 37–46. > # ls /var/spool/cups > c00037 c00040 c00043 c00046 d00039-001 d00042-001 d00046-001 > c00038 c00041 c00044 d00037-001 d00040-001 d00043-001 tmp > c00039 c00042 c00045 d00038-001 d00041-001 d00045-001 > > Looks like c00046 was tbe last job sent to the printer. The c files are the (binary) control files, and the d files are the corresponding data files, basically meaning that a file, once queued to print, can be modified/removed/whatever. In the case of piping stuff in, the d file would be the only copy. I think the 1024 sizes above merely reflect the pipe's buffer size: the corresponding d file will be the true size. CUPS is based on a PDF workflow, so the first step with text is to convert it to a PDF. I think that's the filter which is producing landscape pages, and I think that the second filter, pdftopdf, is autorotating such pages before it gets sent to the printer by the final filter. I'm assuming that when you print an ordinary portrait PDF, you get a normal-looking portrait document. My work pattern would be to convert text or text documents to a PDF first, and printing that. > > Anyway, we now know that orientation-requested=3 works. So make it the > > permanent default. > > > > The flag is documented in the RFC. > > Where is the RFC? $ wget https://www.ietf.org/rfc/rfc8011.txt at page 122. Before you make a portrait option permanent, I would find out what happens with documents that have mixed orientation from page to page. (There's a "mixed.pdf" document available on the wiki page referenced earlier.) You could find that landscape pages that would normally print rotated at 100% scale are instead shrunk to fit across a portrait page instead, which would be undesirable. You can write functions with parameters to print in different ways, rather than setting the options with lpoptions. As an example, here are three that I use. Note that they're NOT forcing an orientation, but only indicate which way round (and if) to print doublesided. print-1s () { lp -d brother -o media=letter -o print-quality=5 "$@" } print-2s-portrait () { lp -d brother -o media=letter -o print-quality=5 -o sides=two-sided-long-edge "$@" } print-2s-landscape () { lp -d brother -o media=letter -o print-quality=5 -o sides=two-sided-short-edge "$@" } You could write one to force portrait when you know that's what you want for the entire document. I'm not sure whether there's a way on unsetting an orientation that you previously set in the lpoptions configuration file (there might be, for all I know). I'm not in the habit of printing directly from applications (or pipes), as I would always want a corresponding PDF file archived of anything I physically print. So I don't know whether trixie has introduced a change here. Cheers, David.
