Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(tasks): bind cleanup and notices to exact owners

Browse Source 
Task records previously retained only ownerSession. If an old agent scope unwound after another agent reused the same agent and session ids, cleanup selected both records and could cancel replacement work. The completion surface also re-resolved the session at settlement, which could inject an old task notice into the replacement agent.

Retain the exact Agent instance for lifecycle work, select owner cleanup by object identity, and pass that exact owner to completion listeners. Keep read, list, kill, and wait authorization session-based as the runtime RFC intends. Add regressions for cleanup and notice routing under id reuse, then update the public docs and generated API catalogs.
This commit is contained in:
Tianyi Cui
2026-07-15 11:40:37 +08:00
parent c2d5a5be80
commit 1a69a3debe
11 changed files with 117 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
packages/tasks/tool-tasks/README.md
View File

@@ -12,7 +12,7 @@ ACP render intent: all three are `generic` cards (`read`/`read`/`execute`) — a
## Completion notices
On `onTaskDone`, injects `background task <id> (<kind>: <label>) finished [status: …]. Read its output with task_output.` into the owning agent's session (`agent.inject()` — durable context for the next request, not a wake-up). Suppressed when the snapshot is `reported` (the model already killed it, or a read/wait returned the end) — never a redundant "finished". The disposed-owner race is contained; a missing agent registry drops the notice.
On `onTaskDone`, injects `background task <id> (<kind>: <label>) finished [status: …]. Read its output with task_output.` through the exact owner `Agent` captured at task start (`agent.inject()` — durable context for the next request, not a wake-up). It never re-resolves a reusable agent/session id to a replacement. Suppressed when the snapshot is `reported` (the model already killed it, or a read/wait returned the end) — never a redundant "finished"; the disposed-owner race is contained.
## Config