On my amd64 linux box the following two commands generate different output files: pcl6 -r72 -sDEVICE=bitrgb -sOutputFile=test1.bit -dNOPAUSE ./c327.bin pcl6 -r72 -sDEVICE=bitrgb -sOutputFile="| cat >test2.bit" -dNOPAUSE ./c327.bin The differences appear on pages 28 and 29 (see attached screenshots for differences on page 29). Note that this is not a non-deterministic problems, I've run each command 100 times and each md5sum is identical. I get the same results on my MacBook Pro under Mac OS X, but on the same machine running Ubuntu 8.10 via VMware Fusion neither command line is deterministic. I believe this problem is related to bug #690299 since the areas that differ are the same (at least on page 29, I've seen differences on page 28 with -r600 and multiple runs but they appear to be infrequent).
Created attachment 4830 [details] screenshot_72.png Screen shot of 72 dpi output for -sOutputFile=test1.bit and -sOutputFile="| cat >test2.bit". Differences are highlighted in red.
Created attachment 4831 [details] screenshot_500.png Screenshot with differences highlighted in red of multiple runs with -sOutputFile=test.bit with -r500 (see bug #690299).
The two test files that exhibit this behavior not report errors: PCL XL error Subsystem: KERNEL Error: ColorSpaceMismatch Operator: SetBrushSource So I'm going to assume the difference is related to how the files are closed when the error occurs and close this bug.