Re: execv bug???
Re: execv bug???
- Subject: Re: execv bug???
- From: Steve Checkoway <email@hidden>
- Date: Sun, 27 Jan 2008 04:38:23 -0800
On Jan 27, 2008, at 12:12 AM, Peter Seebach wrote:
In message <email@hidden>, Steve writes:
if (chdir(auto_qmail) == -1) _exit(61);
execv(*binqqargs,binqqargs);
_exit(120);
Calling _exit from user code is nearly always a mistake.
This is not true at all. If you call fork(2) and a subsequent exec*
fails, then you must call _exit(2). exit(3) calls functions registered
with atexit, flushes open output streams, closes open streams, and
unlinks files created with tmpfile(3).Calling exit(3) will cause the
open output streams to flush twice.
So, the chdir on all other platforms goes to /var/qmail, and, the
execv
executes from there bin/qmail-queue.
So the program you want to execute is /var/qmail/bin/qmail-queue?
Seems like a strange place for it.
No, because as a simple test program shows, that works:
static char *ev[] = { 0, 0 };
int
main(void) {
ev[0] = "bin/ls";
chdir("/");
execv(*ev, ev);
return 23;
}
You're assigning a const char * to a char *. I'm not sure why gcc
doesn't warn about this, but I suspect that it should. In fact, if you
assign "bin/ls" to a const char * first, then it does warn, "warning:
assignment discards qualifiers from pointer target type." That said,
execv(3) works for me without an absolute path as well.
--
Steve Checkoway
"Anyone who says that the solution is to educate the users
hasn't ever met an actual user." -- Bruce Schneier
Attachment:
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden