How to stop :timer

I have this code, i want to stop_clock right now with erlang :timer. How can I do this ? Is it possible ?

 def handle_event("start_clock", %{"id" => id}, socket) do
    :timer.send_interval(1000, self(), :tick)

    card = Kanban.get_card_from_modal!(id)

    estimated_duration = card.task.estimated_duration

    time = estimated_duration / 60
    hour = trunc(time)
    e = time - hour
    minute = round(e * 60)
    second = socket.assigns.second

    {:noreply, socket |> assign(data_card_id: card.id, time: estimated_duration ,hour: hour, minute: minute, second: second)}
  end

  def handle_event("stop_clock", %{"id" => id}, socket) do
    :timer.cancel(:tick)
   {:noreply, socket}
  end

Timers are not identified my the message they send, but the reference returned when starting the timer.

ref = :timer.send_interval(1000, self(), :tick)
:timer.cancel(ref)

In your case you’d need to retain the reference in your state, e.g. in socket.private.

3 Likes

I think that its works but I have tried :timer.kill_after(0) and it works too

Thanks you

iex(1)> :timer.kill_after(0)
** (EXIT from #PID<0.109.0>) shell process exited with reason: killed

I think that’s because it kills the current process, not because it’s actually killing your timer properly.

10 Likes

Creating timers using the functions :erlang.send_after/3 and :erlang.start_timer/3 is much more efficient than using the timers provided by the :timer module.

10 Likes

Just perfect. Thank you!!!

Wow, thanks for that information. That’s the first time I’ve heard about this.
All the examples and docs on elixir part always showed timer.

There is an enormous number of cool functions in the :erlang module :smiley:

1 Like

Can you elaborate why this is? Feels like a place where the less efficient version should be made an alias to the more efficient one unless there’s some good reason to keep the less efficient one.

I think it because the :timer module is old and maybe predates the timer functions in the :erlang module. Originally it was a server running which managed all the timer requests and kept track of when to send messages and to whom. Now it has been heavily optimised and itself uses the timer functions n the :erlang module when it can so it is generally less efficient than calling the function directly. That would be one reason why it is still used in many examples. Also some of the functions like tc need some wrapping to get the same results.

EDIT: What I meant to say is that using :timer is generally almost as efficient as calling the timer functions directly. There is only a very small difference.

3 Likes