How to print landscape files from another server via LPD?
Helge Blischke
h.blischke at acm.org
Fri Jul 30 02:08:51 PDT 2010
John MacLerran wrote:
>> John MacLerran wrote:
>>
>> >> John MacLerran wrote:
>> >>
>> >> >> John MacLerran wrote:
>> >> >>
>> >> >> > Dear CUPS Forum members
>> >> >> >
>> >> >> > I'm trying to set up a queue in CUPS to print landscape all the
>> >> >> > time. I have searched the documentation and forums, and found
>> >> >> > several similar questions, but the solutions presented there did
>> >> >> > not help.
>> >> >> >
>> >> >> > The reason I need this is that our ERP system sends its reports
>> >> >> > in 132
>> >> >> > column landscape format by default, and I'm trying to set up a
>> >> >> > CUPS queue that will accept that output and print it.
>> >> >> > Alternatively, I'm trying to get a defnintive answer of "no, this
>> >> >> > can't be done", so that I can present other, better options (like
>> >> >> > pdf printing and electronic delivery of reports), and still have
>> >> >> > done due diligence.
>> >> >> >
>> >> >> > There are a few wrinkles, however. Our ERP servers are running
>> >> >> > Solaris 10, using LPD to send the output to a linux box that acts
>> >> >> > as our print server.
>> >> >> >
>> >> >> > On Solaris, the print queue is configured like this (the lines
>> >> >> > don't wrap):
>> >> >> >
>> >> >> > ccenterapp_land:\
>> >> >> > :printer-uri-
>> >> >>
supported=lpd\://anaerobic.isu.edu/printers/ccenterapp_land#Solaris:\
>> >> >> > :bsdaddr=anaerobic.isu.edu,ccenterapp_land,Solaris:
>> >> >> >
>> >> >> > (Anaerobic is our Linux box. It is Red Hat Enterprise 5.5 and
>> >> >> > cups 1.3.7, patch level 18, the Red Hat default.)
>> >> >> >
>> >> >> > On Linux, our lpoptions file for this queue looks like this
>> >> >> > (again, no wrap in the real file):
>> >> >> >
>> >> >> > Default ccenterapp_land landscape=true cpi=14 lpi=8
>> >> >> > page-left=36
>> >> >> > page-right=36 page-top=36 page-bottom=36
>> >> >> >
>> >> >> >
>> >> >> > When I send a 132-column text file from Linux, it prints
>> >> >> > correctly -- output is flipped landscape, and extends out to the
>> >> >> > 132-column mark, using this lp command:
>> >> >> >
>> >> >> > lp -d ccenterapp_land testprint.txt
>> >> >> >
>> >> >> > However, when I send that exact same test file from the Solaris
>> >> >> > boxes,
>> >> >> > using the same lp command, I get output that is partially
>> >> >> > correct. It is flipped landscape, but the lines are wrapped into
>> >> >> > portrait margins -- 101 columns, to be exact.
>> >> >> >
>> >> >> > I specified various options in the cups-lpd config file (xinitd,
>> >> >> > I think, I'm not the linux box's sysadmin), but, frankly, I'm not
>> >> >> > sure what options to try, and neither a search here, nor a google
>> >> >> > search, turned up a list of what options are valid.
>> >> >> >
>> >> >> > The current cups-lpd config entry looks like this:
>> >> >> >
>> >> >> > service printer
>> >> >> > {
>> >> >> > disable = no
>> >> >> > socket_type = stream
>> >> >> > protocol = tcp
>> >> >> > wait = no
>> >> >> > user = lp
>> >> >> > server = /usr/lib/cups/daemon/cups-lpd
>> >> >> > server_args = -o job-sheets=none,none
>> >> >> > }
>> >> >> >
>> >> >> > In the server_args parameter, I've tried the following options:
>> >> >> >
>> >> >> > -o document-format=application/octet-stream -- as indicated by
>> >> >> > the man page
>> >> >> > -o document-format=application/vnd.cups-raw -- from a google
>> >> >> > search
>> >> >> > -o landscape -- a s.w.a.g. on my
>> >> >> > part
>> >> >> >
>> >> >> > Neither of which helped. The vnd.cups-raw one really clobbered
>> >> >> > the output -- reducing it to one line, portrait, that ran off the
>> >> >> > right side of the page and was dropped.
>> >> >> >
>> >> >> > What am I missing? How can I make the linux queue print
>> >> >> > landscape, all 132 columns, without wrapping, when the file is
>> >> >> > sent from Solaris?
>> >> >> >
>> >> >> > Thank you
>> >> >> > -John MacLerran
>> >> >> > macljohn at isu.edu
>> >> >>
>> >> >> I think you are using the default system V printing system on your
>> >> >> solaris box. In that system, every print queue is associated with
>> >> >> one or more of the following descriptions:
>> >> >> - printer type
>> >> >> - content type (ASCII, PostScript, etc)
>> >> >> - form name (especially for ASCII content type defining cpi, width
>> >> >> (chars per line, etc.)
>> >> >>
>> >> >> Please post how the respective printer on your Solaris box is
>> >> >> defined with respect to these attributes. Perhaps you only need to
>> >> >> modify one of these.
>> >> >>
>> >> >> Helge
>> >> >>
>> >> >>
>> >> > ------
>> >> > Thanks Helge.
>> >> > I'm not sure those attributes are set up for this print queue.
>> >> >
>> >> > I looked in this directory:
>> >> > /var/spool/lp/admins/lp/printers/
>> >> > and saw directories for some of our printers -- the ones that are
>> >> > defined locally on the Solaris box, but the queue in question is not
>> >> > related to a printer that is local on the Solaris box (which is
>> >> > named Poseidon), but is just a 'remote' queue that resides on
>> >> > Anerobic, a completely different server.
>> >> >
>> >> > I've asked our unix sysadmin if he's aware of another directory
>> >> > where
>> >> > those attributes reside, and we'll check for them. Do you know if
>> >> > the /var/spool/lp/admins/lp/printers/ directory is the one I should
>> >> > be looking in? -john -----------
>> >>
>> >> John,
>> >>
>> >> I currently do not have access to a real Solaris 10 box, which means
>> >> I'm not able to do real tests on that issue.
>> >> But you could try the following:
>> >> - stop the respective printer on your print server
>> >> - print a (small) sample file from your Solaris box
>> >> - post (an URL to) both the job's control and data file from the
>> >> Linux box (the cxxxxx and the dxxxxx-001 file from the
>> >> /var/spool/cups directory).
>> >>
>> >> Looking into these files, I should be able to determine if the
>> >> unwanted wrapping occurs on the Solaris or the Linix box and, perhaps,
>> >> give hints how to settle down this issue.
>> >>
>> >> Helge
>> >>
>> > ---
>> > Thanks, Helge.
>> > You can view the files at:
>> > http://www.isu.edu/~macljohn/c11535
>> > http://www.isu.edu/~macljohn/d11535-001
>> >
>> > To me, the data in d11535-001 does not look wrapped. I'm not sure what
>> > I'm looking at in the control file, but it doesn't look obviously
>> > wrong.
>> >
>> > -john
>> > ---------------
>>
>> John,
>>
>> the two files look OK, so the job should print OK. For further
>> investigation, could you , on the Linux box, execute the command
>> cupsctl --debug-logging
>> and print this (or a similar) job (with the printer enaqbled, of course),
>> then, if the print is finished, execute
>> cupsctl --no-debug-logging
>> and post (an URL to) the current /var/log/cups/error_log ?
>> From that log file, I hope to be able to determine what your CUPS is
>> really doing.
>>
>> Helge
>>
> -------
> Hi Helge,
>
> I've posted the following files in the location specified:
>
> http://www.isu.edu/~macljohn/error_log
> http://www.isu.edu/~macljohn/cupsd.conf
> http://www.isu.edu/~macljohn/ccenterapp_land.ppd
>
> The error log contains material since July 26. Today's information begins
> around line 500 or so. In the material related to today, I see that it
> uses the cupsd.conf file, the ppd file for the printer, and a few other
> filters - textopaps and pstops, I think. I think those are binary program
> files that are standard with cups, so I didn't include those.
>
> Let me know if you need anything else.
> Thanks for being so helpful with this. I really appreciate it.
> -john
> --------------------
Hi John,
from your error_log I learned that your CUPS install uses the texttopaps
filter to convert plain text to PostScript. I did some research in the
internet regarding this filter and found a lot of complaints especially with
respect to landscape printing.
As a first approach to settle your issue I'd suggest the following:
- verify that /usr/lib/cups/filter/texttops exists and is executable.
- in /etc/cups/mime.convs replace all occurences of textopaps
by texttops
- restart cups if necessary
Let me know if it works now.
Helge
P.S.: what strikes me is that you told in a previous post that the problem
does not show up if you print a plain text file from your Linux box locally
using the lp command. Perhaps you could send me a log file (with debug
enabled) for just such a test job.
More information about the cups
mailing list