chore(wip): état runtime .ideai (flux conversation live, layouts)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
2026-06-20 18:52:27 +02:00
parent c60060494d
commit 5cb99fd353
9 changed files with 48 additions and 20 deletions

View File

@ -62,3 +62,5 @@
{"id":"aab45df1-f0ac-4eca-850b-c89b688c3211","conversation":"08336578-5b47-09d2-2f41-5cc483f101f4","atMs":1781945801522,"source":{"kind":"agent","agentId":"aefdbd61-e3d4-4bc1-9f42-c259446a97b5"},"role":"response","text":"Verdict QA final : vert accepté avec réserve environnementale.\n\nJe ne maintiens pas le KO global, car le rouge restant est maintenant expliqué et borné : dans cet environnement sandbox, un bind de socket Unix réel échoue `PermissionDenied` même pour un smoke minimal `UnixListener::bind` dans `/tmp` et `/run/user/1000`. Les tests qui exigent un vrai socket Unix loopback ne peuvent donc pas être probants ici. Le reste du périmètre applicatif, hotfix submit/logs, frontend, wiring hors vrai socket et app-tauri lib hors tests loopback réels est vert.\n\nCommandes vertes déjà validées QA sur ce périmètre :\n\n```text\ncargo test -p application compose_convention_file --no-fail-fast\n```\nRésultat : `15 passed; 0 failed; 28 filtered out`.\n\n```text\ncargo test -p application\n```\nRésultat : `43 passed; 0 failed` pour les unit tests application, tous les tests dintégration application et doctests passés. Warning existant : `method writes is never used` dans `crates/application/tests/orchestrator_service.rs:1642`.\n\n```text\ncargo test -p infrastructure input\n```\nRésultat : `35 passed; 0 failed; 187 filtered out`.\n\n```text\ncargo test -p application --test orchestrator_service\n```\nRésultat : `45 passed; 0 failed` avec le même warning existant `writes`.\n\n```text\ncd frontend && npx vitest run src/features/terminals/useWritePortal.test.tsx\n```\nRésultat : `1 passed`, `12 tests passed`. Warnings Vite existants sur options `esbuild` dépréciées / `oxc` prioritaire.\n\n```text\ncd frontend && npx vitest run src/features/terminals/TerminalView.portal.test.tsx src/features/terminals/TerminalView.test.tsx src/features/terminals/useWritePortal.test.tsx\n```\nRésultat : `3 passed`, `24 tests passed`. Même warnings Vite.\n\n```text\ncd frontend && npx tsc --noEmit\n```\nRésultat : succès, aucune sortie.\n\n```text\ncargo check -p app-tauri\n```\nRésultat : succès, `Finished dev profile`.\n\nCommandes de confirmation après diagnostic sandbox, relancées par QA :\n\n```text\ncargo test -p app-tauri --test orchestrator_wiring -- --skip open_binds_the_project_loopback_endpoint --skip double_open_keeps_a_single_endpoint_no_address_in_use --skip close_cleans_up_the_endpoint_socket_file --skip file_watcher_and_loopback_endpoint_live_together\n```\nRésultat réel :\n```text\nrunning 9 tests\n...\ntest result: ok. 9 passed; 0 failed; 0 ignored; 0 measured; 4 filtered out; finished in 0.01s\n```\n\n```text\ncargo test -p app-tauri --lib -- --skip state::mcp_e2e_loopback_tests --skip state::bind_endpoint_d1_tests --skip mcp_bridge::tests::end_to_end_over_real_loopback\n```\nRésultat réel :\n```text\nrunning 39 tests\n...\ntest result: ok. 39 passed; 0 failed; 0 ignored; 0 measured; 8 filtered out; finished in 0.01s\n```\n\nRéserve exacte : non-validé dans ce sandbox uniquement pour les tests nécessitant la création/binding dun vrai socket Unix loopback :\n- `crates/app-tauri/tests/orchestrator_wiring.rs` : `open_binds_the_project_loopback_endpoint`, `double_open_keeps_a_single_endpoint_no_address_in_use`, `close_cleans_up_the_endpoint_socket_file`, `file_watcher_and_loopback_endpoint_live_together`.\n- `app-tauri --lib` : `state::mcp_e2e_loopback_tests`, `state::bind_endpoint_d1_tests`, `mcp_bridge::tests::end_to_end_over_real_loopback`.\n\nCes tests restent à exécuter dans un environnement non sandboxé permettant `UnixListener::bind` sur socket filesystem. Je nai modifié aucun fichier."}
{"id":"ea85a3be-0cf0-4ad7-b8a7-7344426844b7","conversation":"08336578-5b47-09d2-2f41-5cc483f101f4","atMs":1781971470719,"source":{"kind":"agent","agentId":"a6ced819-b893-4213-b003-9e9dc79b9641"},"role":"prompt","text":"QA globale pour le chantier `feature/conversation-ux-live-state` Lot A. Périmètre implémenté: backend/Tauri read-model `get_project_work_state(projectId)` + frontend onglet/panneau Work read-only. Backend a ajouté `crates/application/src/workstate/mod.rs`, snapshots live, commande/DTO Tauri; frontend a ajouté types/domain/port/adapter/mock/hook/panel et onglet Work.\n\nMerci de relire le périmètre: read-only, pas nouvelle persistance, pas lecture log/handoff, pas FIFO complète, ordre manifeste, live agents hors manifeste ignorés. Exécute les tests pertinents réels. Minimum:\n- `cargo test -p application workstate`\n- `cargo test -p application`\n- `cargo check -p app-tauri`\n- tests app-tauri DTO/live pertinents (`dto_agents`, `list_live_agents_r0b`) et si nécessaire app-tauri lib avec skips socket connus\n- `cd frontend && npx vitest run src/features/workstate/workstate.test.tsx src/features/projects/projects.test.tsx`\n- `cd frontend && npx tsc --noEmit`\nRetourne verdict, commandes exactes/résultats, risques, et si les tests socket Unix restent filtrés pour raison sandbox déjà diagnostiquée. Ne modifie pas le code sauf correction de test strictement nécessaire."}
{"id":"39f05602-a660-4632-accc-42d4887ce1ed","conversation":"08336578-5b47-09d2-2f41-5cc483f101f4","atMs":1781971551021,"source":{"kind":"agent","agentId":"aefdbd61-e3d4-4bc1-9f42-c259446a97b5"},"role":"response","text":"Verdict QA globale Lot A `feature/conversation-ux-live-state` : vert, avec réserve environnementale non bloquante sur les tests socket Unix réels.\n\nPérimètre relu :\n- Backend application : `GetProjectWorkState` agrège en read-only depuis `AgentContextStore::load_manifest`, `LiveSessions::live_agent_snapshots()` et `InputMediator::busy_state`.\n- Pas de nouvelle persistance observée dans le read-model ; pas de lecture log/handoff ; pas de modélisation FIFO complète.\n- Ordre manifeste conservé par itération sur `manifest.entries`.\n- Live agents hors manifeste ignorés par construction, car seuls les agents du manifeste sont émis.\n- Snapshot live typé ajouté sans changer `live_agents()` existant.\n- Tauri : commande `get_project_work_state(projectId)` câblée dans `commands.rs`, `state.rs`, `lib.rs`, DTO camelCase avec `live.kind` et `busy.state/sinceMs`.\n- Frontend : onglet `Work`, port/gateway/adapters/mock/hook/panel ajoutés ; panneau read-only, appelle uniquement `getProjectWorkState`, affiche live/offline + busy/idle, refresh manuel et refresh sur événements existants.\n\nCommandes exécutées et résultats :\n\n```text\ncargo fmt -p application -p app-tauri -- --check\n```\nRésultat : OK, aucune sortie.\n\n```text\ncargo test -p application workstate\n```\nRésultat :\n```text\nrunning 5 tests\ntest workstate_attaches_live_pty_session_to_manifest_agent ... ok\ntest workstate_includes_busy_state_from_input_mediator ... ok\ntest workstate_attaches_live_structured_session_to_manifest_agent ... ok\ntest workstate_ignores_live_agents_absent_from_manifest ... ok\ntest workstate_lists_manifest_agents_idle_without_live_sessions ... ok\n\ntest result: ok. 5 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out\n```\nWarning existant observé :\n```text\nwarning: method `writes` is never used\n --> crates/application/tests/orchestrator_service.rs:1642:8\n```\n\n```text\ncargo test -p application\n```\nRésultat :\n```text\nrunning 43 tests\n...\ntest result: ok. 43 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out\n...\nRunning tests/workstate.rs\nrunning 5 tests\n...\ntest result: ok. 5 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out\n\nDoc-tests application\nrunning 0 tests\n\ntest result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out\n```\nMême warning existant `CapturingFs::writes` unused.\n\n```text\ncargo check -p app-tauri\n```\nRésultat :\n```text\nFinished `dev` profile [unoptimized + debuginfo] target(s) in 1.70s\n```\n\n```text\ncargo test -p app-tauri --test dto_agents\n```\nRésultat :\n```text\nrunning 19 tests\n...\ntest project_work_state_dto_serialises_live_and_busy_camelcase ... ok\n...\ntest result: ok. 19 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out\n```\n\n```text\ncargo test -p app-tauri --test list_live_agents_r0b\n```\nRésultat :\n```text\nrunning 5 tests\ntest no_sessions_yields_empty_list ... ok\ntest both_kinds_live_listed_without_duplicates ... ok\ntest pty_live_agent_is_listed ... ok\ntest same_agent_in_both_registries_is_deduplicated ... ok\ntest structured_live_agent_is_listed ... ok\n\ntest result: ok. 5 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out\n```\n\n```text\ncargo test -p app-tauri --lib -- --skip state::mcp_e2e_loopback_tests --skip state::bind_endpoint_d1_tests --skip mcp_bridge::tests::end_to_end_over_real_loopback\n```\nRésultat :\n```text\nrunning 39 tests\n...\ntest result: ok. 39 passed; 0 failed; 0 ignored; 0 measured; 8 filtered out; finished in 0.01s\n```\nLes 8 tests filtrés restent les tests nécessitant un vrai socket Unix loopback ; le sandbox courant refuse `UnixListener::bind` avec `PermissionDenied`, diagnostic déjà établi. À rejouer hors sandbox permissif socket Unix, non bloquant pour ce Lot A.\n\n```text\ncd frontend && npx vitest run src/features/workstate/workstate.test.tsx src/features/projects/projects.test.tsx\n```\nRésultat :\n```text\nTest Files 2 passed (2)\nTests 13 passed (13)\n```\nWarnings Vite existants : options `esbuild` dépréciées / `oxc` prioritaire.\n\n```text\ncd frontend && npx tsc --noEmit\n```\nRésultat : succès, aucune sortie.\n\nRisques restants :\n- Couverture volontairement limitée au read-model minimal live/busy ; pas de FIFO complète, pas dhistorique conversationnel, pas de lecture log/handoff.\n- Tests socket Unix réels filtrés pour contrainte sandbox connue ; ils restent à exécuter dans un environnement qui autorise le bind de sockets filesystem.\n- Warning existant `CapturingFs::writes` unused dans `orchestrator_service.rs`, sans lien avec ce lot.\n\nAucune correction requise. Je nai modifié aucun fichier."}
{"id":"81d3494b-a647-40ea-909d-7b4649094e81","conversation":"08336578-5b47-09d2-2f41-5cc483f101f4","atMs":1781974100258,"source":{"kind":"agent","agentId":"a6ced819-b893-4213-b003-9e9dc79b9641"},"role":"prompt","text":"Valide le Lot B complet `workstate delegation/queue snapshot` sur la branche active `feature/workstate-delegation-queue`.\n\nPérimètre à vérifier:\n- Backend/Tauri: port read-only `AgentQueueSnapshot`, snapshots FIFO d'`InMemoryMailbox`, `GetProjectWorkState.agents[].tickets`, DTO camelCase, wiring `AppState`.\n- Frontend: types `tickets`, mock normalisé, panneau Work affiche tickets FIFO inProgress/queued, refresh sur `delegationReady`.\n- Les tickets human et agent sont inclus; l'UI ne doit pas les présenter tous comme délégations agent.\n\nCommandes minimales demandées (adapte si le repo impose un cwd):\n- `cargo fmt --all -- --check`\n- `cargo test -p infrastructure mailbox --lib`\n- `cargo test -p application --test workstate`\n- `cargo test -p app-tauri --test dto_agents`\n- `cargo check -p app-tauri`\n- `cd frontend && npx vitest run src/features/workstate/workstate.test.tsx src/features/projects/projects.test.tsx`\n- `cd frontend && npx tsc --noEmit`\n\nTu peux ajouter des commandes pertinentes, mais ne committe pas et ne modifie pas les fichiers applicatifs sauf tests QA indispensables. Ignore le dirty runtime `.ideai/*` s'il est seulement live-session.\n\nRetour attendu: verdict clair, commandes exécutées avec résultats réels, anomalies éventuelles avec fichiers/lignes, et si rouge, donne la sortie d'erreur exploitable."}
{"id":"b916af4a-6f94-4777-92b1-c5f89f898cde","conversation":"08336578-5b47-09d2-2f41-5cc483f101f4","atMs":1781974271224,"source":{"kind":"agent","agentId":"aefdbd61-e3d4-4bc1-9f42-c259446a97b5"},"role":"response","text":"Verdict: VERT pour le Lot B `workstate delegation/queue snapshot` sur `feature/workstate-delegation-queue`.\n\nJe nai modifié aucun fichier. Dirty ignoré côté runtime `.ideai/*`; les fichiers applicatifs/test du lot étaient déjà modifiés dans le worktree.\n\nCommandes exécutées et résultats réels:\n\n1. `cargo fmt --all -- --check`\n - Exit 0, aucune sortie.\n\n2. `cargo test -p infrastructure mailbox --lib`\n - Exit 0.\n - `running 13 tests`\n - `test result: ok. 13 passed; 0 failed; 0 ignored; 0 measured; 215 filtered out; finished in 0.00s`\n - Couvre notamment `snapshot_preserves_fifo_order_and_positions`, `snapshot_carries_ticket_metadata`, `snapshot_is_read_only`, updates après cancel/resolve.\n\n3. `cargo test -p application --test workstate`\n - Exit 0.\n - `running 12 tests`\n - `test result: ok. 12 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s`\n - Couvre `workstate_lists_two_tickets_in_fifo_order`, `workstate_marks_busy_head_in_progress_and_rest_queued`, `workstate_marks_all_queued_when_agent_idle`, `workstate_maps_human_and_agent_ticket_sources`, queue absente/agent hors manifest.\n\n4. `cargo test -p app-tauri --test dto_agents`\n - Exit 0.\n - `running 20 tests`\n - `test result: ok. 20 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s`\n - Couvre `project_work_state_dto_serialises_tickets_camelcase`: `tickets`, `inProgress`, source `agent.agentId`, source `human`, absence de snake_case.\n\n5. `cargo check -p app-tauri`\n - Exit 0.\n - `Finished dev profile [unoptimized + debuginfo] target(s) in 0.59s`\n\n6. `cd frontend && npx vitest run src/features/workstate/workstate.test.tsx src/features/projects/projects.test.tsx`\n - Exit 0.\n - Warnings Vite non bloquants: options `esbuild` dépréciées/ignorées au profit de `oxc`.\n - `Test Files 2 passed (2)`\n - `Tests 17 passed (17)`\n - `Duration 2.02s`\n\n7. `cd frontend && npx tsc --noEmit`\n - Exit 0, aucune sortie.\n\n8. Commande QA ajoutée, car `crates/domain/src/mailbox.rs` est touché par le lot: `cargo test -p domain mailbox --lib`\n - Exit 0.\n - `running 6 tests`\n - `test result: ok. 6 passed; 0 failed; 0 ignored; 0 measured; 170 filtered out; finished in 0.00s`\n\nVérifications de périmètre:\n\n- Backend/Tauri:\n - Port read-only présent: `domain::mailbox::AgentQueueSnapshot` dans `crates/domain/src/mailbox.rs:69`.\n - `InMemoryMailbox` implémente la snapshot read-only dans `crates/infrastructure/src/mailbox/mod.rs:166`; tests FIFO/métadonnées/read-only verts.\n - `GetProjectWorkState` consomme `Arc<dyn AgentQueueSnapshot>` dans `crates/application/src/workstate/mod.rs:130` et mappe `agents[].tickets`; tests application verts.\n - DTO `tickets` camelCase dans `crates/app-tauri/src/dto.rs:1602` et test camelCase dédié vert.\n - Wiring `AppState`: même `InMemoryMailbox` exposé comme `AgentMailbox` et `AgentQueueSnapshot` dans `crates/app-tauri/src/state.rs:994-999`, puis injecté dans `GetProjectWorkState::new` vers `state.rs:1034-1038`.\n\n- Frontend:\n - Types `tickets` présents dans `frontend/src/domain/index.ts:153-179`.\n - Mock normalisé avec fallback `tickets: []` dans `frontend/src/adapters/mock/index.ts:1670-1684`.\n - Panneau Work trie les tickets par `position` avant affichage dans `frontend/src/features/workstate/ProjectWorkStatePanel.tsx:65`.\n - UI distingue source humaine et agent: `Human` pour `source.kind === \"human\"`, sinon `label (agent-id-prefix)` dans `ProjectWorkStatePanel.tsx:21-27`; test `renders an in-progress human ticket` et test FIFO agent verts.\n - Refresh sur `delegationReady` couvert par `frontend/src/features/workstate/workstate.test.tsx:223` et vert.\n\nAnomalies: aucune anomalie bloquante détectée. Les seuls messages observés sont les warnings Vite sur la config `esbuild` dépréciée, sans échec de test ni impact Lot B constaté."}