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

feat(FN-4039): recover worktree db scratch bootstrap on hydration failure

Browse Source 
Codifies the worktree database open-path hydration failure with a recovery bootstrap mechanism, documented the behavior in storage docs, and added test coverage for the executor hydration degradation path — all gated behind a patch changeset for `@runfusion/fusion`.

Fusion-Task-Id: FN-4039
This commit is contained in:
Fusion
2026-05-11 18:09:35 -07:00
committed by gsxdsm
parent c501e00fdf
commit a8b904c993
5 changed files with 72 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/storage.md
View File

@@ -378,4 +378,6 @@ Expected executor log entry on success:
Hydrated worktree DB: 4 tasks, 12 task_documents
```
Failure policy is strict non-blocking: hydration warnings are logged, but worktree creation/execution continues. Canonical task data remains the root project TaskStore DB; if an agent needs non-hydrated rows immediately, `fn_task_show` remains the canonical fallback path.
A concrete recovered failure mode now covered by tests: when a worktree directory exists but its local `.fusion/` scratch state is missing, opening `DatabaseSync(<worktree>/.fusion/fusion.db)` can fail with `unable to open database file`. Hydration now performs destination bootstrap (`mkdir -p .fusion` + schema init) and retries the destination open once before degrading.
Failure policy remains strict non-blocking for genuinely unrecoverable cases: hydration warnings are logged, but worktree creation/execution continues. Examples that still intentionally degrade include source DB missing, destination write-permission failures, and irreconcilable schema/open errors after bootstrap retry. Canonical task data remains the root project TaskStore DB; if an agent needs non-hydrated rows immediately, `fn_task_show` remains the canonical fallback path.