fix job control on FreeBSD and macOS: do not use WCONTINUED

Go's syscall.WaitStatus decodes the "continued" status wrongly on BSD:
on FreeBSD it is read as killed by SIGCONT, so after bg the job was
reported as "Exit 147" and removed from the job table while it was
still running (fg: "no such job"); on macOS it is read as stopped.

The shell marks jobs as running itself when it continues them, so wait4
is called without WCONTINUED and only exited/signaled statuses finish a
process.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

README.md

@@ -256,6 +256,7 @@ $ bg %2                       # oder: angehaltenen Job im Hintergrund weiterlauf
 - `exit` (oder Ctrl+D) mit angehaltenen Jobs warnt einmal (`there are stopped jobs`), ein zweites `exit` direkt danach beendet gosh.
 - `cd`, `set`, `export`, `exit`, `source`, `fg`, `bg` und `wait` lassen sich nicht mit `&` im Hintergrund starten.
 - In Skripten und mit `gosh -c` gibt es keine Job-Control (kein Ctrl+Z, kein `fg`/`bg`), `&` und `wait` funktionieren aber. Hintergrund-Jobs lesen dort von `/dev/null`.
+- Wird ein angehaltener Job von außen fortgesetzt (`kill -CONT`), zeigt `jobs` ihn weiter als `Stopped`, bis er endet oder erneut angehalten wird; `fg` und `bg` funktionieren trotzdem.
 - Einen Job beenden lässt sich mit `kill` über die Prozess-ID (z. B. `kill 12345`); `kill %1` versteht gosh noch nicht.
 
 ### Farben

jobs.go

@@ -162,9 +162,7 @@ func (j *job) update(pid int, ws syscall.WaitStatus) {
 		switch {
 		case ws.Stopped():
 			p.stopped = true
-		case ws.Continued():
+		case ws.Exited() || ws.Signaled():
-			p.stopped = false
-		default:
 			p.finished, p.stopped = true, false
 			p.err = waitStatusError(ws)
 			p.cmd.Process.Release()
@@ -185,17 +183,30 @@ func waitStatusError(ws syscall.WaitStatus) error {
 	return nil
 }
 
-// wait waits until no process of the job is running any more (all finished
+// waitFlags returns the options for wait4: WUNTRACED to see stopped
-// or, with job control, stopped). With nohang it only collects the status
+// processes (with job control) and WNOHANG to not block.
-// changes that are already available and does not block.
+//
-func (j *job) wait(nohang bool) {
+// WCONTINUED is not used: Go's syscall.WaitStatus decodes the "continued"
+// status wrongly on FreeBSD (as killed by SIGCONT) and macOS (as stopped).
+// The shell marks jobs as running itself when it continues them (fg, bg); a
+// job continued from outside (kill -CONT) stays "Stopped" in the job table
+// until it stops again or exits.
+func waitFlags(j *job, nohang bool) int {
 	flags := 0
 	if jobControl && j.pgid != 0 {
-		flags |= syscall.WUNTRACED | syscall.WCONTINUED
+		flags |= syscall.WUNTRACED
 	}
 	if nohang {
 		flags |= syscall.WNOHANG
 	}
+	return flags
+}
+
+// wait waits until no process of the job is running any more (all finished
+// or, with job control, stopped). With nohang it only collects the status
+// changes that are already available and does not block.
+func (j *job) wait(nohang bool) {
+	flags := waitFlags(j, nohang)
 
 	for {
 		// With job control all processes of the job are in its process

jobs_linux_test.go

@@ -105,6 +105,21 @@ func (s *ptyShell) expect(want string) {
 	s.t.Fatalf("timeout waiting for %q, output:\n%s", want, s.out.String())
 }
 
+func TestWaitFlags_NoWCONTINUED(t *testing.T) {
+	defer func(saved bool) { jobControl = saved }(jobControl)
+	jobControl = true
+	j := &job{pgid: 1}
+	for _, nohang := range []bool{false, true} {
+		flags := waitFlags(j, nohang)
+		if flags&syscall.WCONTINUED != 0 {
+			t.Errorf("waitFlags(nohang=%v) contains WCONTINUED, which Go decodes wrongly on BSD", nohang)
+		}
+		if flags&syscall.WUNTRACED == 0 {
+			t.Errorf("waitFlags(nohang=%v) misses WUNTRACED", nohang)
+		}
+	}
+}
+
 func TestJobControl(t *testing.T) {
 	s := startPtyShell(t)
 
@@ -120,6 +135,13 @@ func TestJobControl(t *testing.T) {
 	s.send("bg\n")
 	s.expect("[1]+  Running                 sleep 30 &")
 
+	// After bg the job keeps running and is not reported as finished
+	// (regression: on FreeBSD the "continued" status was read as killed by
+	// SIGCONT, "Exit 147")
+	time.Sleep(300 * time.Millisecond)
+	s.send("jobs\n")
+	s.expect("[1]+  Running                 sleep 30 &")
+
 	// fg brings it back, Ctrl+C reaches only the job
 	s.send("fg %1\n")
 	s.expect("sleep 30")