Hpps and page logging

Hi Konstantin,

Can you please more elaborate the query?

Regards,
Santhosh

Hi Santhosh,

When a printer job uses pstops and hpps, nothing is logged out in /var/log/cups/page_log.

Debugging cups give the following chain:

   D [07/Jul/2018:18:22:59 +0200] [Job 284] 3 filters for job:
   D [07/Jul/2018:18:22:59 +0200] [Job 284] pstops (application/postscript to application/vnd.cups-postscript, cost 66)
   D [07/Jul/2018:18:22:59 +0200] [Job 284] hpps (application/vnd.cups-postscript to printer/test2/application/vnd.cups-postscript, cost 0)
   D [07/Jul/2018:18:22:59 +0200] [Job 284] - (printer/test2/application/vnd.cups-postscript to printer/test2, cost 0)

On the other hand when a printer job uses pdftopdf, there is a specific code in this binary that logs the printed page instead of hpps filter:

   D [07/Jul/2018:18:15:32 +0200] [Job 283] bannertopdf (application/vnd.cups-pdf-banner to application/pdf, cost 32)
   D [07/Jul/2018:18:15:32 +0200] [Job 283] pdftopdf (application/pdf to application/vnd.cups-pdf, cost 66)
   D [07/Jul/2018:18:15:32 +0200] [Job 283] pdftops (application/vnd.cups-pdf to application/vnd.cups-postscript, cost 100)
   D [07/Jul/2018:18:15:32 +0200] [Job 283] hpps (application/vnd.cups-postscript to printer/test2/application/vnd.cups-postscript, cost 0)
   D [07/Jul/2018:18:15:32 +0200] [Job 283] - (printer/test2/application/vnd.cups-postscript to printer/test2, cost 0)

In cups-filters-1.20.3/filter/pdftopdf/pdftopdf.cc, checkFeature function says:

// CUPS standard is that the last filter (not the backend, usually the
  // printer driver) does page logging in the /var/log/cups/page_log file
  // by outputting "PAGE: <# of current page> <# of copies>" to stderr.

And there is a piece of code for the hpps case:

} else if (!strcasecmp(lastfilter, "hpps")) {
            // hpps is last filter, hpps is part of HPLIP and it is a bug that
            // it does not do the page logging.
            param.page_logging = 1;

I think the best solution should be that hpps logs correctly and would like to know if there are plans to do it.

Thanks.

I was thinking about something like this in hppsfilter.c (sorry line numbers are shifted due to other patches):

@@ -889,9 +895,29 @@ int main (int argc, char **argv)
          hpwrite("@PJL ENTER LANGUAGE=POSTSCRIPT\x0a", strlen("@PJL ENTER LANGUAGE=POSTSCRIPT\x0a"));
     }
     unsigned int numBytes = 0;
+ unsigned int pageCount = 0;
+ unsigned int numCopies = 0;
+ char NbCopies[MAX_BUFFER] = {0};
+
+
     while ( (numBytes = cupsFileGetLine(fp_input, line, sizeof(line))) > 0)
+ {
+ // Sometimes it is an argument, sometimes it is in the page content
+ if (!strncmp(line, "%%BeginNonPPDFeature: NumCopies", strlen("%%BeginNonPPDFeature: NumCopies"))) {
+ // Line is not defined by \n at the end, so need to take one line substring
+ strncpy(NbCopies,
+ line+strlen("%%BeginNonPPDFeature: NumCopies")+1,
+ strchr(line,'\n')-line-strlen("%%BeginNonPPDFeature: NumCopies")-2);
+ numCopies=atoi(NbCopies);
+ }
+ // Counting pages
+ if (!strncmp(line, "%%Page:", 7))
+ {
+ pageCount++;
+ fprintf(stderr, "PAGE: %d %d\n", pageCount,numCopies?numCopies:atoi(argv[4]) );
+ }
         hpwrite (line, numBytes);
-
+ }

This question was expired because it remained in the 'Open' state without activity for the last 15 days.