> It works, thanks. I used the (unmodified) genextern.sh and str.c from 1.0.3 > and it compiles (and works!) fine now (BTW, I'm on a intel platform, so the > GNU 2.7.2 cpp is my one and only cpp, I checked it). I'm convinced CPP should be taken out and shot. As a macro/syntatic language it is very restrictive. I've re-written genextern.sh using only sed. > I only had problems with "make clean" ;-). The "xargs rm" should really be > "xargs -r rm" and "make clean" should also remove the Makefile to make it > clear that a fresh ./configure is needed. Use make spotless to get rid of the makefiles too. > After that only the "usual" compiler complaints showed up: > - missing "#include <wait.h>" in mmap_tests.c (since 1.0) > - sscanf in list.c with %u instead of %lu (since 1.0.2) These two were deliberate. the argument passed to wait is always a pointer to an int, but on some platforms the int is encoded in a struct wstatus. This will breaks wait() unless it is implicity defined. %lu on 64 bit machines is a 64 bit number this breaks time_t. -- Prof. Julian Assange |If you want to build a ship, don't drum up people |together to collect wood and don't assign them tasks proff@iq.org |and work, but rather teach them to long for the endless proff@gnu.ai.mit.edu |immensity of the sea. -- Antoine de Saint Exupery
