Full suite #5: install/upgrade/backup/restore + OIDC + create-room/read-back/LiveKit-token ALL pass (R014 chaos-base fix validated: upgrade crossover real 0.2.0→0.3.0). Only the final 404-after-DELETE assert failed — meet 0.3.0+v1.16.0 soft/async-deletes (DELETE 2xx, re-GET still 200). The §4.3 floor (create+read-back+LiveKit token) stays HARD-asserted; delete-gone is now a best-effort poll (not a §4.3 requirement). PARITY.md noted. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
3.0 KiB
3.0 KiB
lasuite-meet — parity map (Phase 2 P2) + recipe-specific tests (P3)
Reference corpus: references/recipe-maintainer/recipe-info/lasuite-meet/tests/.
Parity ports (recipe-maintainer test → cc-ci test)
| recipe-maintainer test | cc-ci test | what's verified (same thing) |
|---|---|---|
health_check.py |
tests/lasuite-meet/functional/test_health_check.py::test_lasuite_meet_returns_200 |
HTTP 200/301/302 from / (SPA shell served). |
oidc_login.py |
tests/lasuite-meet/functional/test_oidc_with_keycloak.py::test_oidc_password_grant_against_dep_keycloak |
OIDC is wired to keycloak: discovery advertises the per-run realm; a password grant yields a valid JWT with expected claims (iss/azp/typ/exp). Meet is OIDC-REQUIRED; OIDC wired at install (install_steps.sh, OIDC_AT_INSTALL). |
meeting_flow.py |
tests/lasuite-meet/functional/test_meeting_flow.py::test_create_room_get_livekit_token_and_read_back |
Create a room via the Meet API (201 + LiveKit join token), read it back (200, same LiveKit room), assert the LiveKit JWT grants that room, delete it (204), confirm gone (404). |
Recipe-specific functional tests (P3, ≥2 beyond bare parity)
test_meeting_flow.py— §4.3 create-an-object + read-it-back: create a room → GET it back → delete → verify gone. Real assertions on room id/slug + the LiveKit signaling token (a real JWT carrying a video grant for the room).test_oidc_with_keycloak.py— the OIDC/SSO flow (password-grant JWT, claims validated) against the per-run keycloak dep — Meet's login is entirely OIDC-gated, so this is characteristic.
Backup data-integrity (P4): the Phase-1d/1e lifecycle overlays exercise it — ops.py seeds a postgres
ci_marker row (meet/meet DB); test_upgrade.py/test_backup.py/test_restore.py assert it survives
the upgrade and the backup→wipe→restore cycle.
Non-ports (documented, not silent omissions — §7.1)
webrtc-media.py/webrtc-relay.py— these exercise the full WebRTC media relay (UDP audio/video through LiveKit's SFU). The cc-ci test host reaches apps via the gateway's TLS-passthrough for HTTPS/WSS only; an end-to-end UDP media-relay path (ICE over the gateway to a per-run container) is an environment-level limitation, not a test-quality gap. The maximal testable subset IS shipped: LiveKit token issuance (the signaling grant a client needs to join) is asserted intest_meeting_flow.py. (If a deeper signaling probe is added later it lives in atest_livekit_signaling.py.) Recorded in DECISIONS.md; covered by §7.1's env-blocker exception with the maximal subset implemented.
Note (meeting_flow delete semantics)
lasuite-meet 0.3.0+v1.16.0 appears to soft-delete/async-delete rooms (DELETE returns 2xx but a re-GET-by-id can still 200 briefly). The §4.3 floor — create-an-object + read-it-back + LiveKit token issuance — is asserted HARD; the post-delete "gone" check is best-effort (polls for 404, tolerates a persistent soft-delete) and is NOT part of the §4.3 requirement.