aquavm/crates/air-lib/trace-handler
Mike Voronov a8b227caf5
refactor(aquavm): getting rid of CallOutputValue in call merger (#353)
This PR refactors call merger of `TraceHandler`. Previously it requires `CallOutputValue` to determine a type of call output value (stream or scalar). And internally it checked correspondence between data result and call output type and return a error if they are not equal. Although execution engine component also had a match over these values and does nothing if they are not matched since `TraceHandler` did this job. This PR eliminate such behaviour and improve isolation of AquaVM modules.
2022-10-06 19:59:47 +03:00
..
src refactor(aquavm): getting rid of CallOutputValue in call merger (#353) 2022-10-06 19:59:47 +03:00
.rustfmt.toml Decouple log targets to a separate crate (#152) 2021-10-05 16:55:04 +03:00
Cargo.toml Update all non-major Rust dependencies (#330) 2022-09-19 14:07:34 +03:00
README.md Rename subtree to subgraph (#265) 2022-05-17 15:53:33 +03:00

AIR trace handler

This crate contains implementation of the CRDT-based merging data algorithm. It exposes the TraceHandler struct that based on the visitor pattern and has public methods that should be called in certain places of AIR instructions execution. Internally TraceHandler contains several FSM and each of such public methods does state transitioning of one or more these FSMs. Below are state transition sequences for all instructions that caller must follow.

Ap instruction

Expected sequence of TraceHandler calls for the ap instruction:

meet_ap_start
    -> meet_ap_end

Call instruction

Expected sequence of TraceHandler calls for the call instruction:

meet_call_start
    -> meet_call_end

Par instruction

Expected sequence of TraceHandler calls for the par instruction:

meet_par_start
    -> meet_par_subgraph_end(..., SubgraphType::Left)
    -> meet_par_subgraph_end(..., SubgraphType::Right)

Fold instruction

Expected sequence of TraceHandler calls for the fold instruction:

meet_fold_start.1 ->
    meet_generation_start.N ->
        meet_next.M ->
        meet_prev.M ->
    meet_generation_end.N ->
meet_fold_end.1

where .T means that this function should be called exactly T times.