Bob Friesenhahn <bfriesen@xxxxxxxxxxxxxxxxxxx> writes: > Since modern OS targets should all be compliant to the current Open Group > specifications it seems that almost all of Autoconf related to OS-related > porting must be obsolete! You're exaggerating to make a point, but indeed, this is mostly true of the Autoconf macros that many of us used when we first started using Autoconf. There's a great deal of stuff that no longer makes any sense whatsoever to check for. I no longer bother checking whether memcmp works, or whether you can include <time.h> and <sys/time.h> at the same time, or if you have to include <strings.h> instead of <string.h>, or if the system is missing strdup or strerror, or if bcopy has to be used instead of memcpy and memmove, or if the arguments of setvbuf are reversed, or if I need to link with -lintl for strftime. Just to name a few. I consider this progress. :) It doesn't make Autoconf less useful. Autoconf has moved on and has new tests for new problems that still exist, and as new interfaces are introduced, there will continue to be portability problems with them. And of course one still has to probe for -lsocket and -lnsl. :) But indeed, much of the C portability mess that I learned through trial and error when porting code between SunOS, ULTRIX, AIX, and NeXTSTEP in the mid 1990s is now completely irrelevant and obsolete, and thank heavens for that. -- Russ Allbery (rra@xxxxxxxxxxxx) <http://www.eyrie.org/~eagle/> _______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
