Graphs + Events for Translations

I am wary of putting extra events into ExecutionInfo, how are you stopping them from reaching the SMT solver? A clean way to go might be to do something like:

#[derive(Debug)]
pub struct ExecutionInfo<'ev, B> {
    /// A vector containing all the events in a candidate execution that the axiomatic concurrency model can use
    pub events: Vec<AxEvent<'ev, B>>,
    /// A vector containing extra events that can appear in the graph, but won't be given to Z3
    pub extra_events: Vec<ExtraEvent<'ev, B>>,
    /// A vector of po-ordered instruction opcodes for each thread
    pub thread_opcodes: Vec<Vec<B>>,
    /// The final write for each register in each thread (if written at all)
    pub final_writes: HashMap<(Name, ThreadId), &'ev Val<B>>,
}

Could have ExtraEvent be a separate type as above or just be AxEvent if that's easiest

rems-project / isla

Graphs + Events for Translations #41