Matthew Kennedy
Matthew Kennedy
- We can't sleep less than 1ms (calling `chThdSleepMicroseconds(999)` is the same as calling `yield()`) - Tick mode forces a high priority interrupt every 1ms, adding some unnecessary CPU load...
Cranking and running modes are both sequential, and this mess happens during the transition from cranking to running: 
$title nothing there is used, do we still need it?
```c++ class MyClass { public: MyClass(lua_State* l) { luaL_error(l, "first constructor called"); } MyClass(lua_State* l, int x) { if (x > 100) { luaL_error(l, "oops, X too large!"); } }...
simulator, 12-1, "use only rise" set to false:  "use only rise" set to true: 
Do we need an option for two wire wasted? Why not just *always* do two wire wasted, and allow users to ignore some pins?
Like #4208, but for the dwell start. Next step is spark firing too (that's still trigger tooth index based). progress to #4195 probably related to #4477 probably related to #4421
We're currently using this enum's values for both the "subsystem" and "index" in executeTSCommand, which is asking to be causing a bug.
Slow RPM is fine, but has some limitations, and instant RPM looks great. Why compute both?