Re: unsuccessfully install of printer brother hl-l2360dw

groenveld@acm.org (John D Groenveld) writes:

In article <87wmei6gyl.fsf@example.com>,
Salvador Mirzo  <smirzo@example.com> wrote:

I could never make this work, but I believe you're right.  Looking at my
printer settings (which I printed out using the printer's menu) I
see... On page 3/3, in network configuration, I see:
>
 <Active services>          BRN3C2AF43F4D7C
                            BINARY_P1
                            TEXT_P1
                            POSTSCRIPT_P1
                            PCL_P1
                            BRN3C2AF43F4D7C_AT
>
I've seen in the CUPS administration that a URL mentions BINARY_P1, so
it seems that CUPS was using this BINARY_P1 service.  Perhaps if I were
to change it to POSTSCRIPT_P1, it would then accept PostScript from the
network.

>
Spew a little PostScript to it.
ISTR ghostscript includes some samples or you can use enscript, a2ps
and friends to generate some.

CUPS and the printer were installed from scratch using the IPP
``driver'' as you instructed me in this thread.  I used a2ps to
transform a text file into a PS one.  I sent the document to the printer
directly on port 9100 and I got out the PostScript code in plain text.

Then I changed the printer's CUPS configuration to use the LPD protocol
and I manually set the queue to POSTSCRIPT_P1, having chosen (as the
driver) the Brother HL-L2360D series from the list installed by the
/brsaler/ OpenBSD package.  Then I used CUPS's lpr to send the document
and I got out a nice looking PostScript document made by a2ps.  This
proves this printer indeed talks PostScript as you said.

Then I went to try two-sided-long-edge.  I began by trying to print with
programs such as evince (a GTK PDF viewer), qpdfview (a QT PDF viewer)
and Firefox (another GTK program) and I explicitly asked for
two-sided-long-edge always getting two-sided-short-edge.  I tried to ask
for one-sided and I got a one-sided print out.  In other words, things
behave almost as one would expect; the only problem being that
two-sided-long-edge always behaves like two-sided-short-edge.

I then tried CUPS's lpr on the command line---below.

But my experience with CUPS was not very good because I could never
print two-sided-long-edge.  It always printed two-sided-short-edge (or
one-sided).  Since I was able to make it always print
two-sided-long-edge by not using CUPS at all, I gave up on CUPS.

>
OpenBSD's ipptool(1) claims your printer supports short and long edge.
Try with cups-lpr(1)
<URL:https://openprinting.github.io/cups/doc/man-lpr.html>
| -o sides=two-sided-long-edge
|     Prints on both sides of the paper for portrait output.
| -o sides=two-sided-short-edge
|     Prints on both sides of the paper for landscape output.

I used CUPS's lpr to print a two-page PostScript document in both
options.  They both came out two-sided-short-edge (that is, landscape).
I said:

$ /usr/local/bin/lpr -P BRWB052162167A6 -o sides=two-sided-long-edge manual.ps
$ /usr/local/bin/lpr -P BRWB052162167A6 -o sides=two-sided-short-edge manual.ps

I got two A4 papers out, both printed the same way,
two-sided-short-edge.  Finally, I said

$ /usr/local/bin/lpr -P BRWB052162167A6 -o sides=one-sided manual.ps

and two A4 papers out, one sided as expected.

I wonder if there's a mistake or something that can be done in the PPD
file of the printer.  On OpenBSD, the PPD file is installed at

  /etc/cups/ppd/BRWB052162167A6.ppd

Here are lines 119--125:

*OpenUI *Duplex/2-Sided Printing: PickOne
*OrderDependency: 10 AnySetup *Duplex
*DefaultDuplex: None
*Duplex None/Off (1-Sided): "<</Duplex false>>setpagedevice"
*Duplex DuplexNoTumble/Long-Edge (Portrait): "<</Duplex true/Tumble false>>setpagedevice"
*Duplex DuplexTumble/Short-Edge (Landscape): "<</Duplex true/Tumble true>>setpagedevice"
*CloseUI: *Duplex

It looks right, but I wish I knew more about this file.  For instance,
if I could somehow set it to always do two-sided-long-edge, I'd be
nearly 100% happy.

My imagination says that these values in quotes are PostScript code.  I
then changed the one that says /Tumble true/ to /Tumble false/.  Like
this:

*Duplex DuplexNoTumble/Long-Edge (Portrait): "<</Duplex true/Tumble false>>setpagedevice"
*Duplex DuplexTumble/Short-Edge (Landscape): "<</Duplex true/Tumble false>>setpagedevice"

I restarted cupsd and said:

$ /usr/local/bin/lpr -P BRWB052162167A6 -o sides=one-sided manual.ps
$ /usr/local/bin/lpr -P BRWB052162167A6 -o sides=two-sided-long-edge manual.ps
$ /usr/local/bin/lpr -P BRWB052162167A6 -o sides=two-sided-short-edge manual.ps

The first came out one-sided, as expected; the next two came out
two-sided-short-edge.  Lol---my hack didn't work.

By the way, I also tried landscape with two-sided-short-edge (using
Firefox).  It comes out right: landscape two-sided-short-edge.

I'm out of ideas for now.  But I still think there's a way to solve this
because the printer is able to do PostScript; it can do two-sided
including both long and short edge.  My intuition says that what's going
on is simply a mistake somewhere.  I was very hopeful with my hack
above, but it made no difference.

Perhaps what I'm going to try next is to study a little bit of
PostScript to be able to write a document myself that's very simple and
that it instructs the printer directly to see if I can get
two-sided-long-edge.  My feeling is that these options in the PPD file
are used to know what PostScript code to include in the document.  Since
I don't see the code that the printer actually gets in the end, that's a
margin for confusion.

You already know I switched to OpenBSD in the process of debugging this.
FWIW, I was able to set up the printer to use IPP.  I was able to print
with it, but with the same two-sided-short-edge behavior as using the
LPD:// URL in CUPS.
>
All in all, CUPS worked, but it turns out my life has been better with
just using /etc/printcap and lpr.
>
I wanted to thank you for the help! 
>
If I find anything new, I'll post again.

>
If you get OpenBSD's cups to work, that will point to a bug in FreeBSD's.

At this point I don't think FreeBSD is at fault.

BTW my WAG is OpenBSD's vmm supports FreeBSD guests if you want test
FreeBSD without dual booting:
<URL:https://www.openbsd.org/faq/faq16.html#StartVm>

Not for testing CUPS at the moment, since OpenBSD is behaving exactly as
FreeBSD did.  (For everything else, I have a headless FreeBSD server,
but no printer attached to it.)

Thanks very much for the help!