This page lists where each test layer lives and the command that runs it. The principles and the layers are in Testing.
| Layer | Where it lives | Local command |
|---|---|---|
| Build scripts | build.rs in each crate |
mise build |
| Unit tests | #[cfg(test)] modules next to the code |
mise test:fast |
| Compiler integration | crates/integration-tests/tests/compiler/ (fixtures), crates/integration-tests/tests/compiler_legacy/ (Rust tests) |
mise test:local |
| Code-graph suites | crates/integration-tests-codegraph/ |
mise test:integration:codegraph |
| CLI integration | crates/integration-tests/tests/cli.rs |
mise test:cli |
| Container tests | crates/integration-tests/tests/ |
mise test:integration |
| E2E | e2e/ |
e2e/scripts/setup.sh, then e2e/scripts/test.sh |
| Load test | scripts/ci/orbit-perf/, crates/integration-tests/tests/server/performance/scenarios/ |
bash scripts/ci/orbit-perf.sh |
| Performance tools | code-indexing-benchmark.yaml, crates/query-engine/profiler/ |
mise query:profile |
| Fuzzing | crates/fuzz/ |
mise test:fuzz, mise fuzz:<target> |
Other commands:
mise test:compilerruns the compiler fixtures.mise bench:runmeasures code indexing on real repositories, andmise bench:smokeruns one indexing pass for each repository.mise query:diffcompares two query profiles.bash scripts/ci/orbit-perf.shruns the load test on your machine with the small synthetic profile. It deploys the Orbit image tag from the Caproni values. SetGKG_IMAGEandGKG_IMAGE_TAGto test another published image, andSYNTH_CONFIGorROUNDSto change the size or the length of the run.mise perf:gtimeandmise perf:rssrecord the time and memory oforbit workspace indexfor before and after comparisons.mise test:fastskips the integration-test, code-graph, fuzz, and profiler crates, which have their own jobs.
Container tests
Run container tests through mise. The mise runner starts one ClickHouse container and passes its address to the tests. On macOS, start the Colima gkg profile first:
colima start gkg --memory 12
mise test:integration| Task | What it runs |
|---|---|
mise test:integration |
All container tests |
mise test:integration:data |
Query scenarios and data correctness |
mise test:integration:server |
Data correctness, redaction, and graph formatting |
mise test:integration:indexer |
All indexer tests |
mise test:integration:indexer:sdlc:scenario <filter> |
SDLC indexer scenarios whose path contains the filter |
mise test:integration:corpus |
The query corpus and the executable docs |
mise test:integration:overlay <name> |
Data correctness with an ontology overlay |
Harness references
| Harness | Reference |
|---|---|
| Test binaries, CI jobs, lanes, and Rust-only data tests | integration-tests/README.md |
| Compiler fixtures | compiler/README.md |
| Code-graph suites | integration-tests-codegraph/README.md |
| Container harness and scenario format | integration-testkit/README.md |
| Indexer scenario schema | indexer_scenario.schema.json |
| Query corpus | fixtures/queries/corpus |
| End-to-end harness | E2E testing harness |
| Code-indexing benchmarks | code-indexing-benchmark.yaml |
| Reference architecture bench | bench/README.md |
| Fuzzing, corpus promotion, and crash triage | fuzz/README.md |
| SOX billing boundary | SOX billing boundary |