Comment 33 for bug 362186

I have tried it on the HP LaserJet 3390 with the PostScript PPD (also a PostScript printer from HP). I could not observe the problem.

Looks like corruptions of the fonts. The scans of the test pages of both reporters show that a small piece of the lower border of each letter is added to the upper border.

In the table shown by the first reporter it seems that the right border of the "1" got added on the left hand side (is this printed in Landscape?).

The incoming data (in most cases PostScript) is converted to PDF at first by the Ghostscript-based pdftops CUPS filter and after applying page management by pdftopdf it is converted back to PostScript by the also Ghostscript-based cpdftocps CUPS filter. Somewhere on this filter chain (or already in the originally supplied file) all fonts get embedded in the file, so the printer-internal fonts are not used.

Can you please do the following tests:

Send a PDF file directly to CUPS (with the "lp" or "lpr" command). This leads to the pstopdf filter not being used. Does this lead to correct printouts?

Try the following manual file conversions in a terminal window:

gs -q -sstdout=%stderr -sDEVICE=pdfwrite -sOutputFile=- -dBATCH -dNOPAUSE
-dPARANOIDSAFER /usr/share/system-config-printer/testpage-a4.ps > test1.pdf
gs -q -sstdout=%stderr -sDEVICE=pswrite -sOutputFile=- -dBATCH -dNOPAUSE
-dPARANOIDSAFER test1.pdf > test2.ps
gs -q -sstdout=%stderr -sDEVICE=ps2write -sOutputFile=- -dBATCH -dNOPAUSE
-dPARANOIDSAFER test1.pdf > test3.ps

Display the files test1.pdf, test2.ps, and test3,ps on the screen, for example with evince. Which files are correct? Which files show the corrupted characters.