On Thu, 14 Nov 2002 21:03:05 +0100, Peter Kirchgessner wrote: Hi,
>I just compared your second error log with one from Linux. It looks like >the assertion failed in the child process. But the check that failed >there has been done before many times in the parent process. And there >it works. Could it be that for OS/2 the child process does not get a >copy of the parents memory ? Yes that's possible. fork() is problematic on OS/2 and should be avoided. Not all handles/objects are inherited to child-processes. So it's saver here if the reader-process would not be fork()ed. Perhaps the fork()ing of the reader-process could be made optional by something like #ifdef USE_FORK ... If this is too much work, the second solution could be to replace the the fork() approach by a os2-thread approach like allready done for mustek or microtek2. But first it has to be checked if fork() really is the cause of the problem, >If you like, I can send you a patch to hp.c to get some more detailed >debug print outs. But you need to recompile it on OS/2. Yes I'm interested and can recompile it. Regards, Franz
