upside down duplexed page -- CUPS issue (was Re: ps problem)

[Polytropon]

On Sun, 26 Jan 2014 23:27:55 -0700, Gary Aitken wrote:
> On 01/26/14 22:05, Polytropon wrote:
> > On Sun, 26 Jan 2014 21:51:54 -0700, Gary Aitken wrote:
> 
> >> 2. If I print it using
> >>      /usr/local/bin/lpr duplex.ps
> >>    it comes out with the duplexed page upside down.
> > 
> > This will use CUPS's settings which may add a "preample"
> > of PS to the printer to reflect those settings, for example
> > which "duplexing manner" (long or short edge "wrap") will
> > be used.
> 
> Except that in this case the %cupsJobsTicket: comment should
> instruct it to use long-edge duplexing, which should override
> any cups default.  The default for this printer is no-duplex,
> so if it overrides it shouldn't duplex at all.

Yes, "should", but the reality of complex software teaches that
in many cases, things don't work as they should. :-)

It would be required to capture the "postprocessed" PS that
CUPS sends to the printer to be sure about what _exactly_ will
be sent to the printer. Sometimes printers themselves can
change things (duplex on/off, short/long edge etc.).



> > To find out if those settings exist, open a web browser (bah)
> > and go to http://localhost:631 where you can find the CUPS
> > administration web pages. Check your printer's settings
> > carefully. (I know there's also a lpadmin tool, but haven't
> > bothered learning it because I tend to avoid CUPS whenever
> > possible, because when you have a PS-capable printer, it's
> > probably just useless overhead.)
> 
> Checked that already, and it is set by default to not duplex.
> So presumably its preamble would not change the duplexing.
> But the document itself has embedded duplexing along the long
> edge.  Why would that result in the short-edge behavior?

The wonder of modern software. :-)

No, honestly: If you already have a good printer that can
understand PS, and you don't require a _specific_ CUPS-only
feature, I'd suggest to get rid of CUPS and use the system's
printer spooler. You'll probably don't need any printer filter
here.



> >> If I create a simple 2 page document in openoffice,
> >> when printing from openoffice, I see the following:
> >>
> >> 3. If printed duplexed direct from OO, duplexed pages are upside down.
> > 
> > As expected.
> 
> Why?  Default is no duplex, but I turned long edge duplex on
> in the print dialog.

Strange case of conflicting pre- and postprocessing maybe?
PS content changes and precedence?



> >> 4. If printed to a pdf file and /usr/bin/lpr is used on the pdf, 
> >>    no duplexing occurs -- I get two separate sheets.
> > 
> > This could be explained by the way OO generates PDF and how
> > the printer filters turn that into PS. Or if your printer is
> > also capable of speaking PDF (directly), no mentioning that
> > it should use duplex is in the PDF input. 
> 
> ok, at least that part makes sense.  /usr/bin/lpr will ignore the 
> $cupsJobTicket: statement.

Correct. It will send the PS _as is_ to the printer and let it
deal with the PS content, which is defined by the PS interpreter's
implementation "inside the printer".



> >>From your PS source:
> > 
> >> %cupsJobTicket: media=Letter sides=two-sided-long-edge
> > 
> > There's also two-sided-short-edge (corresponding to the setting
> > accessible via localhost:631).
> 
> That's what I don't understand.
> It says two-sided-long-edge.  Why would it be performing as if
> it said two-sided-short-edge?

I'd imagine that there's a conflicting setting somewhere in the
CUPS configuration. Instead of manually searching through the
web interface, you could also search for options related to
"duplex", "edge", or "tumble" in /usr/local/etc/cups in the
configuration files and maybe the PPDs employed. For example,
the string "*DefaultDuplex: DuplexNoTumble" looks promising.
There are other entries related to that concept.


-- 
Polytropon
Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...