Duplex problem

[Helge Blischke]

stelmo at savba.sk wrote:
> 
> Helge Blischke wrote:
> > stelmo at savba.sk wrote:
> > >
> > > Helge Blischke wrote:
> > > > stelmo at savba.sk wrote:
> > > > >
> > > > > Anonymous wrote:
> > > > > > Helge Blischke wrote:
> > > > > > > stelmo at savba.sk wrote:
> > > > > > > >
> > > > > > > > This is a summary of what I have found out so far
> > > > > > > >
> > > > > > > > 1) The duplex is ok for text and other format files
> > > > > > > > so it seems that the problem is with the ps files only
> > > > > > > >
> > > > > > > > 2) with a file produced with dvips (from a dvi file) I get no duplex
> > > > > > > > with the same file after I had run ps2ps I get one blank page
> > > > > > > >
> > > > > > > > 3) if I replace pstops filter of cups 1.1.20 with the filter from
> > > > > > > > the 1.1.14 release I get duplex right.
> > > > > > > >
> > > > > > > > I run debian sarge, cups version 1.1.20, printer is xerox phaser 6250dp and I use win's ppd.
> > > > > > >
> > > > > > >
> > > > > > > Post (an URL to) a sample file. Whithout knowing the details, it is
> > > > > > > virtually impossible to give any advice.
> > > > >
> > > > > the test files can be found at
> > > > >
> > > > > www.quniverse.sk/stelmo/cups
> > > > >
> > > > > this first file was generated usig latex and then dvips while the second
> > > > > ps file was dowloaded from the internet.
> > > > >
> > > > > Peter
> > > >
> > > > Well, both files print duplex if requested from the command line (-o
> > > > sides=two-sided-long-edge)
> > > > at least with my installateion (CUPS 1.1.19 and the alternate pstops
> > > > filter). But it may be
> > > > that the 1.1.20 pstops filter is more regirous with respect to DSC
> > > > comnpliance.
> > > > Both files lack the "%%BeginProlog" comment line which may screw up
> > > > things belonging into the
> > > > prolog or setup section, respectively.
> > > >
> > > > Perhaps you could contact the radical eyes people to make dvips really
> > > > DSC compliant (it should
> > > > be not too difficult, believe me).
> > > >
> > > > Helge
> > > >
> > > > --
> > > > H.Blischke at srz-berlin.de
> > > > H.Blischke at srz-berlin.com
> > > > H.Blischke at acm.org
> > >
> > > In addition I have tried the following:
> > > I downloaded fresh cupsys-1.1.20 source package from debian mirror and I have replaced the source file pstops.c with an older pstops.c
> > > from the 1.1.14 version of cupsys and compile the whole thing (cupsys).
> > > In this case (the binary files pstops differ, cupsys 1.1.20 vs. 1.1.14 but the same source file pstops.c)  I get duplex right.
> > > So it seems to me that the problem lies in the filter pstops indeed (I just wanted to be sure and see clearly what you see right away)
> > >
> > Please, post or send me the PPD for the printer you have trouble with. I
> > just looked at
> > the latest three versions of pstops.c (the ones distributed with 1.1.19,
> > 1.1.20, and
> > the CVS one) and didn't find anything that could disturb the duplex
> > printing.
> > As I don't know which source version of this filter was in 1.1.14, I
> > have no idea -
> > perhaps you could send me a diff.
> >
> > As for the PPD, if that contains a JobPatchFile statement, in case of a
> > missing
> > %%BeginProlog statement the job patch file stuff may be emitted _after_
> > the setup stuff
> > (the defaults you configured for the printer).
> >
> > Helge
> >
> >
> > --
> > H.Blischke at srz-berlin.de
> > H.Blischke at srz-berlin.com
> > H.Blischke at acm.org
> 
> I have put everything on the web under the link
> www.quniverse.sk/stelmo/cups
> 
> there are two pstops.c files one from the cupsys 1.1.20 and
> the other from the cupsys 1.1.14 debian source packages
> my cupsys configuration files including ppd and Readme containing description of the files and comments.
> Just a small comment. I have tried a few versions of ppds for
> this printer but the other versions needed to be modified before
> I could use them. So this seems to me the best.
> 

Thanks for the sample files. The PPD you posted backs my suspicion
regarding the job patch file:
It resets the dictionary xerox$pagedevice to the initial values; this
any previously set 
features like duplex etc. get lost.

Verify this by inserting the line
---snip---
%%BeginProlog
---snip---
after the %%EndComments line in one of your offending PS files - you'll
se then it prints duplex.

If fixing dvips isn't an issue (ant it seems so), pleas post an RFE to
the CUPS home page
requesting that, within the pstops filter, the do_setup function
initially calls do_prolog() if
the prolog stuff has not yet been sent. Perhaps you may succeed in
getting this into the final
1.1.21 release.

Helge


-- 
H.Blischke at srz-berlin.de
H.Blischke at srz-berlin.com
H.Blischke at acm.org