Mercurial > emacs
view src/s/isc3-0.h @ 8475:77d5e055dce1
(PTY_TTY_NAME_SPRINTF): New definition.
(sigsetmask): Add #undef.
(SOLARIS2_4): Defined.
Discard old contents of file. Instead, undef C_SWITCH_SYSTEM and
const, so that we use GCC without -traditional.
(HAVE_VFORK): Defined.
author | Richard M. Stallman <rms@gnu.org> |
---|---|
date | Sun, 07 Aug 1994 07:58:18 +0000 |
parents | f5aa65299489 |
children | 4ae54c3e1dc3 |
line wrap: on
line source
/* s- file for Interactive (ISC) Unix version 3.0 on the 386. */ #include "isc2-2.h" /* These have been moved into isc2-2.h. */ /* #define HAVE_SOCKETS #define HAVE_SELECT */ /* This appears on 3.0, presumably as part of what SunSoft call X2. */ #undef NO_X_DESTROY_DATABASE /* mt00@etherm.co.uk says this is needed for process.c. */ #define HAVE_TIMEVAL /* People say that using -traditional causes lossage with `const', so we might as well try getting rid of -traditional. */ #undef C_SWITCH_SYSTEM /* We indirectly #include s/usg5-3.h, which says to use libX11_s and libc_s. Martin Tomes <mt00@controls.eurotherm.co.uk> says that ISC has no libX11_s, and that linking with libc_s causes sbrk not to work. */ #undef LIB_X11_LIB #undef LIBX11_SYSTEM #define LIBX11_SYSTEM -lpt -lnls -lnsl_s -lc /* marko@tekelec.com (Marko Rauhamaa) says that his linker couldn't find memmove, but that sounds crazy - I thought all SYSV descendants had that. Let us know if this turns out to be wrong. */ /* It is safe to have no parens around the args in the safe_bcopy call, and parens would screw up the prototype decl for memmove. */ #define memmove(d, s, n) safe_bcopy (s, d, n) /* This works around a bug in ISC 4.0 and 3.0; it fails to clear the "POSIX process" flag on an exec. */ #define EXTRA_INITIALIZE __setostype (0)