bug#26397: [PATCH] Use vfork if possible on Darwin (bug#26397)

[YAMAMOTO Mitsuharu <mituharu@math.s.chiba-u.ac.jp>]

>>>>> On Sun, 9 Apr 2017 17:06:21 -0700, Aaron Jensen <aaronjensen@gmail.com> said:

> For me, this slows down call-process in TTY emacs by 3ms (from
> 2.5ms->5.5ms). I typically launch a GUI emacs and connect to the
> server in TTY via emacsclient, and in that scenario this patch does
> not slow down call-process in the TTY, but it will affect those who
> use TTY only, it seems.

Then use "obsolete" ioctl TIOCNOTTY instead of setsid?  I'm not
familiar with this matter, so could someone check if this is OK?  I
confirmed that only the controlling terminal of the child process was
detached.

				     YAMAMOTO Mitsuharu
				mituharu@math.s.chiba-u.ac.jp

diff --git a/src/callproc.c b/src/callproc.c
index 05048576ce..3f6df422f3 100644
--- a/src/callproc.c
+++ b/src/callproc.c
@@ -52,6 +52,8 @@ along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.  */
 #include "syswait.h"
 #include "blockinput.h"
 #include "frame.h"
+#include "systty.h"
+#include "keyboard.h"
 
 #ifdef MSDOS
 #include "msdos.h"
@@ -626,7 +628,18 @@ call_process (ptrdiff_t nargs, Lisp_Object *args, int filefd,
     {
       unblock_child_signal (&oldset);
 
+#ifdef DARWIN_OS
+      /* Darwin doesn't let us run setsid after a vfork, so use
+	 TIOCNOTTY when necessary. */
+      int j = emacs_open (DEV_TTY, O_RDWR, 0);
+      if (j >= 0)
+	{
+	  ioctl (j, TIOCNOTTY, 0);
+	  emacs_close (j);
+	}
+#else
       setsid ();
+#endif
 
       /* Emacs ignores SIGPIPE, but the child should not.  */
       signal (SIGPIPE, SIG_DFL);