root/geo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add MediaSupplyFavorite types and update task plan for generated-code drift diagnosis
Backend CI / Backend (push) Successful in 17m27s
Details

Browse Source
This commit is contained in:
liangxu
2026-07-15 00:04:13 +08:00
parent 3cb15ab5a3
commit 3037cfe059
4 changed files with 55 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
findings.md
+9
View File
@@ -1,6 +1,7 @@
# Findings & Decisions
## Requirements
- 2026-07-15: Fix CI `Main Verify generated code is committed` failure; reported diff begins around generated `MediaSupplyOrder`.
- 2026-07-14 follow-up: with exactly one usable account for a platform, the first human-verification challenge must prevent every later local task for that platform from reaching the adapter.
- 2026-07-14: When any account hits a human-verification challenge, retry the current collection with another usable account for the same AI model/platform.
- The client/platform error must always produce a desktop system notification, regardless of whether account failover succeeds.
@@ -11,6 +12,13 @@
- For desktop AI monitoring, put scheduling authority on the client, keep scraping silent/background-only, allow stale cross-day tasks to be dropped, and avoid fanning one question to all models at once.
## Research Findings
- 2026-07-15 generated-code CI: local `sqlc` is v1.30.0 and `server/Makefile` invokes the unpinned system binary (`cd internal/tenant/repository && sqlc generate`).
- The current workspace has no root `.github/workflows/backend-ci.yml`; the active `Main Verify` workflow must be located elsewhere (likely Gitea workflow config) before choosing a generator version.
- Active backend workflow is `.gitea/workflows/backend-ci.yml`; the earlier missing `.github` path was historical documentation drift.
- `generated/models.go` currently has no local diff before regeneration; `MediaSupplyOrder` fields are aligned by the committed generator output.
- Gitea Backend CI pins `sqlc v1.30.0`, matching the local binary exactly, then runs `make sqlc-generate` and `git diff --exit-code -- server/internal/tenant/repository/generated`.
- Version-mismatch hypothesis is falsified. The next deterministic probe is the exact CI generation command.
- The reported `MediaSupplyOrder` is generated in `server/internal/tenant/repository/generated/models.go` around line 699.
- Screenshot evidence: on client `e83077a0...`, `monitoring_collect:7995` for Doubao failed with `doubao_challenge_required`, while later Doubao task `monitoring_collect:7989` also appeared as failed with the same message.
- Screenshot refinement: the first challenged desktop task `#7224` shows attempt 1, while later `monitoring_collect:7989` shows attempt 0 and the same copied challenge message. This indicates the later task was bulk-terminalized without adapter execution, but recorded as `failed`, making cancellation look like another failed attempt.
- The product-correct state model should be: the triggering task is `failed`; later same-model tasks are `cancelled` in `desktop_tasks` and `skipped` (with a human-verification skip reason) in `monitoring_collect_tasks`, and cancellation events remove them from the local scheduler.
@@ -185,6 +193,7 @@
## Issues Encountered
| Issue | Resolution |
|-------|------------|
| Initial CI workflow lookup assumed root `.github`, but that directory does not exist in this checkout | Search all workflow/config paths before reproducing the generator command |
| A zsh glob for `apps/desktop-client/vitest.config.*` had no matches during read-only inspection | Switched to `rg --files`/explicit paths; no repository state changed |
| First planning-file patch used a stale placeholder row and failed verification | Re-read the exact table context and applied a narrower patch |
| Repository is fully untracked in git status output | Avoid relying on git history; verify via filesystem and test commands instead |