Timer event order pollution

This PR fixes a bug (found by https://groups.google.com/g/mpf-users/c/JtZxG7oOceQ/m/P3htggdmAwAJ) where the order of a Timer's control_events could cause a crash.

The Problem

There has been a longstanding (7+ years) flaw in the timer control event code where the event handler kwargs were a singleton dictionary, so each event handler would alter/amend the same dictionary and attach it as kwargs. This oversight slipped by largely unnoticed until earlier this year, when the timer code was revamped to pass more event kwargs through (to support dynamic placeholder values for tick value, tick interval, et cetera., see #1761).

Even then, this oversight slipped through because the timer kwarg pollution is dependent on the order in which the timer's control_events are listed. Only if a control event with a value comes before a reset/restart will the former event's timer_value kwarg be polluted into the reset/restart handler, causing a duplicate argument error.

The Fix

This PR simply moves the kwargs = {} declaration to be inside the control event iteration block, so each control event starts with a fresh, empty, unpolluted kwarg dictionary. With this change, the order of control events doesn't matter and pollution is avoided.

tick tock tick tock

missionpinball / mpf

Timer event order pollution #1851

The Problem

The Fix

Quality Gate passed