Re: EXIT trap not executed on error

[Bart Schaefer <schaefer@brasslantern.com>]

[-- Attachment #1: Type: text/plain, Size: 615 bytes --]

On Sun, Dec 11, 2022 at 8:52 PM Bart Schaefer <schaefer@brasslantern.com> wrote:
>
> I don't recall why this suppresses traps on errflag, if I ever knew.
> However, the comment suggests the calling code should be clearing
> errflag around the dotrap().
>
> The question is whether its OK here to just clear errflag like this,
> or if it should be saved and restored?

Seems reasonable to errflag on the side of safety (ahem) and do the restore.

The new first hunk covers another possible case; immediately outside
that diff context we call realexit(), so there is no possibility of
errflag being referenced again.

[-- Attachment #2: special_exit_trap.txt --]
[-- Type: text/plain, Size: 666 bytes --]

diff --git a/Src/exec.c b/Src/exec.c
index 7001fd615..2b7e0c7c5 100644
--- a/Src/exec.c
+++ b/Src/exec.c
@@ -1598,6 +1598,7 @@ sublist_done:
 			       (isset(ERRRETURN) && !errreturn)) &&
 		    !(noerrexit & NOERREXIT_EXIT);
 		if (errexit) {
+		    errflag = 0;
 		    if (sigtrapped[SIGEXIT])
 			dotrap(SIGEXIT);
 		    if (mypid != getpid())
@@ -1630,9 +1631,12 @@ sublist_done:
 	thisjob = cj;
 
     if (exiting && sigtrapped[SIGEXIT]) {
+	int eflag = errflag;
+	errflag = 0;	/* Clear the context for trap */
 	dotrap(SIGEXIT);
 	/* Make sure this doesn't get executed again. */
 	sigtrapped[SIGEXIT] = 0;
+	errflag = eflag;
     }
 
     unqueue_signals();