Re: [PATCH v2 3/5] chain kill signals for cleanup functions

[Jeff King <peff@peff.net>]

On Thu, Jan 22, 2009 at 01:02:35AM -0500, Jeff King wrote:

> diff --git a/test-sigchain.c b/test-sigchain.c
> new file mode 100644
> index 0000000..8747dea
> --- /dev/null
> +++ b/test-sigchain.c
> [...]
> +int main(int argc, char **argv) {
> +	sigchain_push(SIGINT, one);
> +	sigchain_push(SIGINT, two);
> +	sigchain_push(SIGINT, three);
> +	raise(SIGINT);
> +	return 0;
> +}

The signal-handling test was failing on my Solaris auto-build. After
much painful debugging, it seems that when running without a controlling
terminal (such as under cron), the signal handler for terminal related
signals (including SIGINT) is initialized to SIG_IGN. Thus after popping
all of our signal handlers, we restore the SIG_IGN, the program is _not_
killed by the signal, and the test fails.

One fix would be to just "signal(SIGINT, SIG_DFL)" at the top. But I
think it makes the test cleaner to just switch to a more reliable
signal. The patch would look something like what is below. But I need to
know what exit code Windows generates for SIGTERM. Johannes?

---
diff --git a/t/t0005-signals.sh b/t/t0005-signals.sh
index 9707af7..09f855a 100755
--- a/t/t0005-signals.sh
+++ b/t/t0005-signals.sh
@@ -12,7 +12,7 @@ EOF
 test_expect_success 'sigchain works' '
 	test-sigchain >actual
 	case "$?" in
-	130) true ;; # POSIX w/ SIGINT=2
+	143) true ;; # POSIX w/ SIGTERM=15
 	  3) true ;; # Windows
 	  *) false ;;
 	esac &&
diff --git a/test-sigchain.c b/test-sigchain.c
index 8747dea..42db234 100644
--- a/test-sigchain.c
+++ b/test-sigchain.c
@@ -14,9 +14,9 @@ X(three)
 #undef X
 
 int main(int argc, char **argv) {
-	sigchain_push(SIGINT, one);
-	sigchain_push(SIGINT, two);
-	sigchain_push(SIGINT, three);
-	raise(SIGINT);
+	sigchain_push(SIGTERM, one);
+	sigchain_push(SIGTERM, two);
+	sigchain_push(SIGTERM, three);
+	raise(SIGTERM);
 	return 0;
 }