Zsh Mailing List Archive
Messages sorted by: Reverse Date, Date, Thread, Author

Re: Strange function/pipestatus behavior, maybe a scope bug?



On Thu, 24 Oct 2013 09:57:13 +0100
Peter Stephenson <p.stephenson@xxxxxxxxxxx> wrote:
> > Oh, one other bug which I don't believe is new:  the PIPEFAIL option
> > doesn't work right if the last command is a builtin.  Maybe that
> > test should be moved into storepipestats() as well, in which case
> > the redundancy I mentioned above might also be eliminated.
> 
> Are you saying that $pipestatus is correct but the overall status is
> wrong if you have something like "true | true | true" or "true | true |
> false"?  It certainly sounds like they ought to run at the same point.
> The last command ought to be the easy one --- that's where the status
> has always come from, so if those two pipelines are giving different
> answers with and without PIPEFAIL something is definitely odd.

I can't see any problem here, which might not be too surprising if it's
a race, but I don't see how the following fix can be wrong if pipestatus
is now working: the most complicated case appears to be if it gets
called multiple times and STAT_CURSH is set, but even there it looks
like it's idempotent with respect to lastval (assuming the information
hasn't changed, of course).

I don't know if this potentially makes one of the calls removable.

There is some bad interaction with completion, at least on my setup
(happens from a vanilla setup, too):

% true | true | false
% print $pipest<TAB>
0

This doesn't seem to happen with hooks, so completion is failing to save
something it should.

diff --git a/Src/jobs.c b/Src/jobs.c
index 82ffdf2..200e38e 100644
--- a/Src/jobs.c
+++ b/Src/jobs.c
@@ -377,7 +377,7 @@ check_cursh_sig(int sig)
 }
 
 /**/
-int
+static void
 storepipestats(Job jn, int inforeground)
 {
     int i, pipefail = 0, jpipestats[MAX_PIPESTATS];
@@ -397,7 +397,13 @@ storepipestats(Job jn, int inforeground)
 	numpipestats = i;
     }
 
-    return pipefail;
+    if (jn == jobtab+thisjob) {
+	if (jn->stat & STAT_CURSH) {
+	    if (!lastval && isset(PIPEFAIL))
+		lastval = pipefail;
+	} else if (isset(PIPEFAIL))
+	    lastval = pipefail;
+    }
 }
 
 /* Update status of job, possibly printing it */
@@ -531,17 +537,8 @@ update_job(Job jn)
 	return;
     jn->stat |= (somestopped) ? STAT_CHANGED | STAT_STOPPED :
 	STAT_CHANGED | STAT_DONE;
-    if (jn->stat & STAT_DONE) {
-	int newlastval = storepipestats(jn, inforeground);
-
-	if (job == thisjob) {
-	    if (jn->stat & STAT_CURSH) {
-		if (!lastval && isset(PIPEFAIL))
-		    lastval = newlastval;
-	    } else if (isset(PIPEFAIL))
-		lastval = newlastval;
-	}
-    }
+    if (jn->stat & STAT_DONE)
+	storepipestats(jn, inforeground);
     if (!inforeground &&
 	(jn->stat & (STAT_SUBJOB | STAT_DONE)) == (STAT_SUBJOB | STAT_DONE)) {
 	int su;
@@ -991,7 +988,7 @@ printjob(Job jn, int lng, int synch)
 
     if (skip_print) {
 	if (jn->stat & STAT_DONE) {
-	  (void) storepipestats(jn, job == thisjob);
+	    storepipestats(jn, job == thisjob);
 	    if (should_report_time(jn))
 		dumptime(jn);
 	    deletejob(jn, 0);
@@ -1122,7 +1119,7 @@ printjob(Job jn, int lng, int synch)
 /* delete job if done */
 
     if (jn->stat & STAT_DONE) {
-	(void) storepipestats(jn, job == thisjob);
+	storepipestats(jn, job == thisjob);
 	if (should_report_time(jn))
 	    dumptime(jn);
 	deletejob(jn, 0);

pws



Messages sorted by: Reverse Date, Date, Thread, Author