The recent thread about infinite loops when exiting has not addressed
workers/44007.

The workers/45656 segfault is in BUGS twice, but was fixed by workers/45660.

45396 for lseek() has been applied.

users/24904 appears to have been fixed by workers/44480, I can't reproduce.

45568 was fixed by 45616.

(Patch to remove the above four, attached.)

users/26150 (MULTIOS does not work with "exec") can be fixed by
tweaking "nowait" and "pline_level" in the "If nullexec is 1 we ..."
code at line 3894 in exec.c ... but they should only change if a
multio was in fact spawned.  How is that detect-able?

The foregoing is actually closely related to 42609, in both cases the
shell ends up waiting for exit of a job that is in turn waiting for
the shell.  Both probably are side-effects of a long-ago change to
create job table entries for process substitutions.  So maybe there's
a better fix that involves job table flags.