hush: explain why wait5.tests is failing

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

diff --git a/shell/hush.c b/shell/hush.c
index df96e6fde016882122ae2c80f54c63f4a9ada794..b131f6095b75dcc2469a712b12b043d64839c2e7 100644 (file)
--- a/shell/hush.c
+++ b/shell/hush.c
@@ -7401,6 +7401,10 @@ static int process_wait_result(struct pipe *fg_pipe, pid_t childpid, int status)
                                printf(JOB_STATUS_FORMAT, pi->jobid,
                                                "Done", pi->cmdtext);
                        delete_finished_bg_job(pi);
+//bash deletes finished jobs from job table only in interactive mode, after "jobs" cmd,
+//or if pid of a new process matches one of the old ones
+//(see cleanup_dead_jobs(), delete_old_job(), J_NOTIFIED in bash source).
+//Testcase script: "(exit 3) & sleep 1; wait %1; echo $?" prints 3 in bash.
                }
        } else {
                /* child stopped */
@@ -9899,6 +9903,15 @@ static int FAST_FUNC builtin_wait(char **argv)
                                        ret = job_exited_or_stopped(wait_pipe);
                                        if (ret < 0)
                                                ret = wait_for_child_or_signal(wait_pipe, 0);
+//bash immediately deletes finished jobs from job table only in interactive mode,
+//we _always_ delete them at once. If we'd start doing that, this (and more)
+//would be necessary to avoid accumulating dead jobs:
+# if 0
+                                       else {
+                                               if (!wait_pipe->alive_cmds)
+                                                       delete_finished_bg_job(wait_pipe);
+                                       }
+# endif
                                }
                                /* else: parse_jobspec() already emitted error msg */
                                continue;