Re: Patches Applied for 21.1.8

Andy> At 06:38 PM 10/23/99 +0200, Adrian Aichner wrote: Andy> Bzzzt. The mingw32 build doesn't use ntheap.c! Obviously it Andy> needs to be included somewhere else. Hello Andy, hello Craig! Andy: Ah, Uh, that explains some. I looked around but still don't have a clue. What should we do? WINDOWSNT is defined in a mingw32 build, right? Will WINDOWS95 be defined in a MINGW32 build on Windows9[58]? I found src/s/windows95.h, which sets WINDOWS95 and includes src/s/windowsnt.h Craig: Are you using the configure script? I am ignorant of mingw32. What is the configure line you use? At this point, the whole PID handling isn't too elegant. Even on UNIX, an int PID is moved into Lisp_Int and back, although a Lisp_Int might not have room for an int. One way to solve this would be to return the PID from create_child() unmodified (negative for 9[58], positive for NT, which is an illusion, they both return a DWORD which is an unsigned int). All cases where UNIX flavor builds use negative pids would have to be distinguished from large Windows9[58] PIDs which are actually unsigned. Any thoughts? Regards, Adrian Andy> andy Andy> -------------------------------------------------------------- Andy> Dr Andy Piper Andy> Senior Consultant Architect, BEA Systems Ltd