fix supervisors restarting temporary children

In the current implementation of supervisors, temporary children
should never be restarted. However, when a temporary child is
restarted as part of a one_for_all or rest_for_one strategy where
the failing process is not the temporary child, the supervisor
still tries to restart it.

Because the supervisor doesn't keep some of the MFA information
of temporary children, this causes the supervisor to hit its
restart limit and crash.

This patch fixes the behaviour by inserting a clause in
terminate_children/2-3 (private function) that will omit temporary
children when building a list of killed processes, to avoid having
the supervisor trying to restart them again.

Only supervisors in need of restarting children used the list, so
the change should be of no impact for the functions that called
terminate_children/2-3 only to kill all children.

The documentation has been modified to make this behaviour
more explicit.

1 files changed, 7 insertions, 0 deletions

diff --git a/lib/stdlib/src/supervisor.erl b/lib/stdlib/src/supervisor.erl
index e60706ed05..dc31647eb5 100644
--- a/lib/stdlib/src/supervisor.erl
+++ b/lib/stdlib/src/supervisor.erl
@@ -735,6 +735,13 @@ restart(one_for_all, Child, State) ->
 terminate_children(Children, SupName) ->
     terminate_children(Children, SupName, []).
 
+%% Temporary children should not be restarted and thus should
+%% be skipped when building the list of terminated children, although
+%% we do want them to be shut down as many functions from this module
+%% use this function to just clear everything.
+terminate_children([Child = #child{restart_type=temporary} | Children], SupName, Res) ->
+    do_terminate(Child, SupName),
+    terminate_children(Children, SupName, Res);
 terminate_children([Child | Children], SupName, Res) ->
     NChild = do_terminate(Child, SupName),
     terminate_children(Children, SupName, [NChild | Res]);