On 5/6/07, Per Inge Mathisen <[EMAIL PROTECTED]> wrote: > On 5/6/07, Dennis Schridde <[EMAIL PROTECTED]> wrote: > > > <sys/cdefs.h> does this work. Are we including that? If so, maybe > > > _XOPEN_SOURCE is not being defined consistently in our code. > > On GNU/Linux it is done by features.h, which is included by unistd.h. > > The problem might be that somewhere in frame.h sys/cdefs.h is included and > > when we include unistd.h in exceptionhandler.c, it's inclusion is protected > > by header-guards... > > I had the same problem earlier with _GNU_SOURCE, and I solved it by > making sure the relevant headers were included in frame.h and that > frame.h was always loaded first. > > Given that we have now, I think, six build systems in active use(*), > we should not just dump tests and defines into configure.ac as a way > of fixing problems, even though this would be the right autoconf way > of doing things. > > - Per > > *) autoconf, raw makefiles, waf, cmake, xcode and msvc. (Yes, this is > not good state of affairs.)
Okay, by setting _XOPEN_SOURCE=600 in the Xcode project, I got a bit further. Now, I get an error about NSIG not being defined in exceptionhandler.c. As it turns out, <sys/signal.h> only defines NSIG if *neither* _ANSI_SOURCE nor _POSIX_C_SOURCE is defined. I am not sure what to use in its place. See http://www.opengroup.org/onlinepubs/009695399/basedefs/signal.h.html _______________________________________________ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
