I've been going through the thousands of differences caused by r11549 and so far have come across one file which appears to be a regression. See the attached 11548.png and 11549.png files. The command lines used: bin/gs -sDEVICE=pdfwrite -o test.pdf -r300 ./Bug687044.ps bin/gs -sDEVICE=ppmraw -o test.ppm -r300 ./test.pdf
Created attachment 6610 [details] 11549.png
(In reply to comment #0) > I've been going through the thousands of differences caused by r11549 and so > far have come across one file which appears to be a regression. See the > attached 11548.png and 11549.png files. > > The command lines used: > > bin/gs -sDEVICE=pdfwrite -o test.pdf -r300 ./Bug687044.ps > bin/gs -sDEVICE=ppmraw -o test.ppm -r300 ./test.pdf Actually this file has always been wrong. By some weird co-incidence Ghostscript was previously able to produce the expected result, but Acrobat (our reference) always missed the offending text. This is believed to be another manifestation of type 3 font problems with pdfwrite, and we already have an open report for this (which already includes this file). *** This bug has been marked as a duplicate of bug 691383 ***