Single conductor scenarios

.github/workflows/test.yaml

@@ -94,6 +94,56 @@ jobs:
 
           pkill --echo hc && pkill --echo holochain && pkill --echo lair-keystore
 
+      - name: Smoke test - app_install
+        run: |
+          # Start a sandbox conductor and run it in the background
+          nix develop --command bash -c "hc s clean && echo "1234" | hc s --piped create && echo "1234" | hc s --piped -f 8888 run &"
+
+          # Run the scenario for 5 seconds
+          RUST_LOG=info nix run .#app_install -- --connection-string ws://localhost:8888 --agents 2 --behaviour minimal:1 --behaviour large:1 --duration 5 --no-progress
+
+          pkill --echo hc && pkill --echo holochain && pkill --echo lair-keystore
+
+      - name: Smoke test - first_call
+        run: |
+          # Start a sandbox conductor and run it in the background
+          nix develop --command bash -c "hc s clean && echo "1234" | hc s --piped create && echo "1234" | hc s --piped -f 8888 run &"
+
+          # Run the scenario for 5 seconds
+          RUST_LOG=info nix run .#first_call -- --connection-string ws://localhost:8888 --agents 1 --behaviour local:1 --duration 5 --no-progress
+
+          pkill --echo hc && pkill --echo holochain && pkill --echo lair-keystore
+
+      - name: Smoke test - write_read
+        run: |
+          # Start a sandbox conductor and run it in the background
+          nix develop --command bash -c "hc s clean && echo "1234" | hc s --piped create && echo "1234" | hc s --piped -f 8888 run &"
+
+          # Run the scenario for 5 seconds
+          RUST_LOG=info nix run .#write_read -- --connection-string ws://localhost:8888 --duration 5 --no-progress
+
+          pkill --echo hc && pkill --echo holochain && pkill --echo lair-keystore
+
+      - name: Smoke test - write_query
+        run: |
+          # Start a sandbox conductor and run it in the background
+          nix develop --command bash -c "hc s clean && echo "1234" | hc s --piped create && echo "1234" | hc s --piped -f 8888 run &"
+
+          # Run the scenario for 5 seconds
+          RUST_LOG=info nix run .#write_query -- --connection-string ws://localhost:8888 --duration 5 --no-progress
+
+          pkill --echo hc && pkill --echo holochain && pkill --echo lair-keystore
+
+      - name: Smoke test - local_signals
+        run: |
+          # Start a sandbox conductor and run it in the background
+          nix develop --command bash -c "hc s clean && echo "1234" | hc s --piped create && echo "1234" | hc s --piped -f 8888 run &"
+
+          # Run the scenario for 5 seconds
+          RUST_LOG=info nix run .#local_signals -- --connection-string ws://localhost:8888 --duration 5 --no-progress
+
+          pkill --echo hc && pkill --echo holochain && pkill --echo lair-keystore
+
       - name: Build scenario bundles
         run: |
           set -euxo pipefail

CHANGELOG.md

@@ -8,7 +8,24 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [Unreleased]
 
 ### Added
+- Exposed `on_signal` from the app websocket in the instrumented websocket.
+- New handler function `handle_api_err` which can be used with `map_err` to deal with `ConductorApiError`s and convert
+  them into `anyhow` errors or panic when the error is fatal.
+- New common helper `uninstall_app`, see its rustdoc for details.
+- Each run will now generate a unique run ID which is used to keep report data separate between runs. At some point it
+  will be possible to specify a run ID to use but for now it is generated automatically.
+- Check in the `happ_builder` whether `hc` and `cargo` are available. This is used by the scenario build script to skip
+  building happs if build tools are not available. This allows the project to be loaded in an environment where the
+  tools aren't available.
+
 ### Changed
+- Updated to new Holochain client version 0.5.0-alpha.4 which allowed `&mut self` to be replaced with `&self` in admin
+  and app instrumented websockets.
+- `ShutdownHandle` now hides its implementation. It works the same way that it did but you can no longer access the 
+  broadcast channel that it uses internally. Shutdown failures used to panic but it a `ShutdownHandle` happens to not
+  have any subscribers then that should not be considered a fatal error. It will now log a warning instead.
+- Metrics now automatically include `run_id` and `scenario_name` tags.
+
 ### Deprecated
 ### Removed
 ### Fixed