Matthew,
> Thanks for the update. As it happens, the IJS module for the Braille > embosser I'm working on can accept colour data (but converts it to > greyscale). This is because the embosser is capable of printing at a > number of different ``intensity'' levels. So it's not usually a 1-bit > device (as the test input I'm using is in colour).
What do you mean by "can accept colour"? ghostscript's ijs device-independent core (if *nobody* messes it around with extra patches!) has always been able to take 24-bit color, 8-bit color, 32-bit cmyk or whatever is valid postscript, and down-sample to whatever the ijs server (the device-sepecific part) needs. This is part of the ijs protocol. Your ijs driver can negotiate with ghostscript's ijs core to get it to send 24-bit color-data, but it is very stupid to do so (because it is 120MB of data being passed around, instead of 5MB, for a 24-bit device versus a 1-bit device). If your device is capable of a number of gray levels, your driver should negotiate with the ijs core to send data out in DeviceGray with that number of gray-depth. It is all part of the IJS protocol - your ijs device driver can tell ghostscript's ijs core to send data out in 4-bit gray, 2-bit gray, etc, if that's the maximum gray-depth of your device.
If this is not the same bug, I'll file one separately as the KRGB patch is in Debian's gs-gpl (I checked debian's diff), and had been experienced by our users so I would appreciate if you would test whether the unmodified gs 8.00/8.14 debian packages our project provides works without segfault for you. if our 8.00 and 8.14 works, and debian-shipped gs 8.01 segfaults, then the answer is clear.
Regards, Hin-Tak
-- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
