Comment 22 for bug 369503

My alternative pdftops filter is a shell script which does principally the same as the original pdftops filter which is a compiled C program. It calls Ghostscript to convert PDF to PostScript. I have written the shell script for quick testing. One can edit it quickly and attach it to bug reports to test the changes. The original filter uses the Ghostscript "pswrite" output device and the alternative filter uses the "ps2write" output device. The PostScript output is different: The former delivers DSC-conforminmg PostScript but converts all text characters into bitmaps, the latter conserves text as text and conserves the fonts, but its PostScript output is not DSC conforming.

The "pswrite" output device has several compatibility problems which got discovered recently: This bug, bug 362186, bug 44989, and bug 361772. The "ps2write" output device solves these problems but its PostScript is not DSC-conforming, so it cannot be used in the PostScript printing workflow of CUPS used by most distributions (PDF input --pdftops--> Non-DSC-PS --pstops--> ...) as the pstops filter which does the page management (N-up, selected pages, reverse order, ...) is fed with non-DSC-conforming PostScript and PostScript must be DSC-conforming to reliably tell pages apart. In the PDF-centric printing workflow in Ubuntu (PDF input --pdftopdf--> page-managed PDF --pdftops--> ...) This is no problem.