In investigating the pamcmyk4 vs plank differences, I came across problems when running the following file at 300dpi, both banded and unbanded with both plank and pamcymk4. tests_private/pcl/pcl5cats/Subset/PWEP6SC1.BIN If you examine gx_dc_ht_colored_fill_rectangle (in gxcht.c) you'll see that it falls into 2 cases; the first is an 'optimisation' case for the second. As such disabling the first case should cause no differences in the output at all - just a slowdown. Unfortunately, testing shows this not to be true. This isn't a plank vs pamcmyk4 bug, so I'm trying to avoid burning more time on it now - but I am raising this bug so the problem isn't forgotten.
Created attachment 9230 [details] 0001-Latest-unicode-handling.patch This fixes all Sags comments, except for the 'a)' section of the environment variables, I think.
Comment on attachment 9230 [details] 0001-Latest-unicode-handling.patch Wrong bug, sorry!