Skip to content

test(embed): isolate runtime integration scenarios - #620

Merged
senamakel merged 5 commits into
epic/8092-inference-outfrom
epic-8092-final-union
Oct 22, 2026
Merged

senamakel merged 5 commits into
epic/8092-inference-outfrom
epic-8092-final-union

Conversation

@senamakel

Copy link
Copy Markdown
Owner

Summary

Split the multi-scenario openhuman-embed integration tests into one test binary per runtime scenario. Each process now owns its process-global runtime slot, so independent tests no longer collide with one another; shared scenario implementations live under tests/support/.

Validation

  • cargo test -p openhuman-embed --tests --no-fail-fast — all targets passed serially.
  • cargo clippy -p openhuman-embed --all-targets -- -D warnings — passed.
  • cargo fmt --all -- --check and the repository pre-push format/lint/Clippy hooks — passed.
  • pnpm rust:layout, pnpm saas:ambient, and pnpm agent:runtime-boundary — passed.
  • Full product coverage produced 91% changed-line coverage (4,580 measured lines; 402 missed), above the 80% threshold. The broad coverage runner itself reported unrelated fixture/runtime failures and a missing output-directory issue; the lcov was generated afterward and coverage-presence passed.

Part of tinyhumansai#8092.

senamakel and others added 5 commits October 22, 2026 06:25
… test

Replace the bare unwrap on runtime construction with unwrap_or_else that panics with the underlying error, so failures in the attached tools test report the actual cause instead of an opaque panic.

Auto-committed-on: cuttlefish
Co-authored-by: Medulla <medulla@tinyhumans.ai>
The attached tools test now prints the runtime build error to stderr before panicking, so the underlying failure is visible in test output instead of being buried in the panic message.

Auto-committed-on: cuttlefish
Co-authored-by: Medulla <medulla@tinyhumans.ai>
…se files

The four large integration test modules for attached tools, cron agents,
host-only tools, and structured turns were broken up into individual
per-scenario test files with shared support modules, so each case can be
read and run in isolation.

Auto-committed-on: cuttlefish
Co-authored-by: Medulla <medulla@tinyhumans.ai>
Wire the shared common module into the attached tools, host-only tools, and
structured turns integration tests so they can reuse the common test helpers.
Also adjust the cron agents support module to match the new shared setup.

Auto-committed-on: cuttlefish
Co-authored-by: Medulla <medulla@tinyhumans.ai>
Each test target includes the shared scenario modules but only runs one
case, so the unused helpers in the other modules triggered dead code
warnings. Adding a crate-level allow silences them without changing what
the tests exercise.

Auto-committed-on: cuttlefish
Co-authored-by: Medulla <medulla@tinyhumans.ai>
@senamakel
senamakel merged commit f1e71e9 into epic/8092-inference-out Oct 22, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant