On 08/10/14 11:57, Till Kamppeter wrote:

On 10/07/2014 11:45 PM, Tobias Hoffmann wrote:

I'm not sure what exactly is happening in your setup.
Maybe your PPD is missing something (- and I'm not an expert for PPD/PJL)?
Maybe you don't have a recent cups-filters (>= 1.0.34)?
Maybe you're hitting some edge-cases that are not handled correctly in
the current code?

AFAIK no-one else has reported a problem with PJL in a pure pdf
workflow, and I can't find an obvious flaw in the code which deals with
PJL in pdftopdf. The code is also basically copied over from the old
pdftopdf implementation (cups-filters <=1.0.20); if your PPD does hw
copy with the old implementation, but not with >=1.0.34, I will probably
be able to fix that.

Tobias, do you know for what the call

emitJCLOptions(stdout,ppd,param.deviceCopies);

in the emitPreamble() function in pdftopdf_jcl.cc is good for? For me it
looks like that it serves for inserting JCL options, but inserting JCL
options is already done by

ppdEmitJCL(ppd,stdout,param.jobId,param.user,param.title);

The code basically comes from the old pdftopdf implementation by Koji Otani. The same code is also present in imagetopdf.c .
There is also a section in cups-filter's README about that:

if (Copies option is specified in the PPD file) {
    mark Number of copies specified
} else if (pdftopdfJCLCopies is specified in the PPD file) {
    output JCL specified with JCLCopies
}

for (each marked options) {
    if (pdftopdfJCL<marked option's name> is specified in the PPD file) {
        output it's value as a JCL
    } else if (pdftopdfJCLBegin attributes is specified in the PPD file) {
        output "<option's name>=<marked choice>;" as a JCL
    }
}

AFAIUI Koji had to "invent" new pdftopdfJCL* keywords for the PJL + PDF case;
and for some reason ppdEmitJCL could not be used for them, so he added emitJCLOptions.

I can even comment out

emitJCLOptions(stdout,ppd,param.deviceCopies);

and the output stays the same.

Probably because the PPD neither contain *Copies nor *pdftopdfJCLCopies?

What I want to have is that if param.deviceCopies is not 1 (= multiple
hardware copies) that

@PJL SET COPIES=...

is added.

Have you tried adding something along this lines:

*OpenUI *Copies/Number of Copies: PickOne
*OrderDependency: 100 AnySetup *Copies
*DefaultCopies: 1
*Copies 1/1: "@PJL SET COPIES=1"
*Copies 2/2: "@PJL SET COPIES=2"
*Copies 3/3: "@PJL SET COPIES=3"
*Copies 4/4: "@PJL SET COPIES=4"
...
*CloseUI: *Copies

similar to how Foomatic-based work?

A real PPD also has to take other settings into account here (e.g. collate, staple, ...), if the printer-hardware does not support all combinations / has staple-limits / ...

Currently all works fine for PPDs with "*cupsManualCopies: True" as
pdftopdf generates software copies then. Without this entry pdftopdf
sends only one copy of the PDF content and adds only a comment for
subsequent filters. It does not add anything to generate the copies on a
PDF printer.

The PPD author (printer manufacturer, ...) chooses to EITHER set "*cupsManualCopes: True" OR add *Copies / *pdftopdfJCLCopies / ... (i.e. whatever is required to generate correct JCL output for their printer model).

Tobias