Comment 2 for bug 1299951

Revision history for this message
Peter Selinger (selinger) wrote :

I investigated this issue further, and collected additional information that may help in reproducing and finding the cause of this bug.

Specifically, the bug only occurs with files printed from, or saved as PostScript from, Acrobat Reader. As a testfile, I used the attached gridsynth.ps and gridsynth.pdf

* Test 1: print gridsynth.ps from the command line (lpr gridsynth.ps). It prints just fine, at full resolution, and all pages will print. Even duplexing works.

* Test 2: print gridsynth.pdf from within Acrobat Reader. It will print at a much lower resolution (maybe 150dpi), and only the first page will print - the back of the first page will be blank, and there will be no further pages.

* Test 3: print gridsynth.pdf from within Acrobat Reader by "printing to file", with output gridsynth-acro.ps. Then print gridsynth-acro.ps from the command line (using lpr). It will print at a much lower resolution (maybe 150dpi), and only the first page will print. The printer output is the same as for Test 2.

* Test 4: Convert gridsynth.pdf to PostScript using some other method, such as pdf2ps. Printing the resulting PostScript file from the command line, it prints fine, at full resolution, and the printer output is the same as Test 1.

* Test 5: take the file gridsynth-acro.ps from Test 3, convert it to PDF using ps2pdf, then convert the result back to PS using pdf2ps. Print the resulting file from the command line. It prints fine, at full resolution, with the same output as in Test 1.

I repeated similar tests with other input files, and the outcome is the same. All PostScript that was generated by Acrobat Reader will print at low resolution and stop after the first page. All other PostScript will print fine. This shows that there is nothing wrong with the input files per se.

Tests 1 and 4 show that there is nothing wrong with the original input files gridsynth.ps and gridsynth.pdf. Test 5 shows that the PostScript file generated by Test 3 does not lose any actual information; one can re-distill correct PostScript from this file which will print correctly. So it is not the case, for example, that Acrobat Reader itself lowers the resolution of the file.

Nevertheless, any PostScript that is directly generated by Acrobat Reader will print incorrectly.

I don't think that these tests show that the bug is 100% in Acrobat Reader. Rather, they point to some kind of incompatibility between the PostScript generated by Acrobat Reader and the CUPS/HPLIP filters that are supposed to process it. Especially Test 5 shows that the Acrobat-generated PostScript is salvagable.

For reference, I am attaching all the files:

* gridsynth.ps: used in Test 1

* gridsynth.pdf: used in Test 2

* gridsynth-acro.ps: generated by Acrobat Reader from gridsynth.pdf in Test 3

* gridsynth-pdf2ps.ps: generated by ps2pdf from gridsynth.ps in Test 4

* gridsynth-acro-ps2pdf-pdf2ps.ps: generated from gridsynth-acro.ps in Test 5