Skip to content

fix(task): restore delegation link across interrupt/resume + type NewTaskTool provider call - #27

Open
easonLiangWorldedtech wants to merge 9 commits into
mainfrom
up-8
Open

fix(task): restore delegation link across interrupt/resume + type NewTaskTool provider call#27
easonLiangWorldedtech wants to merge 9 commits into
mainfrom
up-8

Conversation

@easonLiangWorldedtech

Copy link
Copy Markdown
Owner

Part 8/8 of the task-tree series (builds on parts 1-7).

Description

Fixes two correctness gaps in the delegation lifecycle:

  • Interrupting and resuming a parent no longer loses the child's delegation link, so resumed parents keep their children attached.
  • The NewTaskTool provider call is typed instead of cast, removing an untyped boundary.

Test Procedure

  • pnpm --dir src exec vitest run (delegation interrupt/resume cases in the task suites)

Pre-Submission Checklist

  • Scope: delegation-link fix + typing only
  • Self-review performed
  • Tests added for the regression
  • Visual snapshot: not applicable (no UI change)
  • Documentation impact: none required

easonliang28 and others added 9 commits August 17, 2026 10:32
Part 1/8 of the task-tree series (upstream-ready recomposition).

Adds `depth` to HistoryItem and a cycle-safe `backfillTaskDepth()` that
propagates parent depth through the delegation tree, so every task knows its
nesting level. Depth is surfaced for later use by settings validation,
environment details, and history-tree display.

Includes the single-open-invariant spec mock for backfillTaskDepth (folded in
from the series' CI fix) so this PR passes unit tests standalone.

Depth contract hardening:
- Children inherit their live parent's depth authority instead of assuming it,
  so a legacy non-authoritative parent never stamps its placeholder depth onto
  the child as a persisted fact (regression test added).
- computeTaskDepth now processes an ancestor reached after exactly
  MAX_DEPTH_WALK hops; only strictly longer chains are rejected (boundary
  regression test added).
…ound-trip

Part 2/8 of the task-tree series (upstream-ready recomposition).

Adds two global settings that control subtask delegation:

- `maxNestingDepth` (default 2, range 0-5): how many levels a subtask may
  nest. A value of 0 disables delegation entirely — every new_task runs inline.
- `autoFlattenOnLimit` (default true): when the limit is reached, flatten the
  subtask into the current conversation instead of opening a new tab; when
  disabled such requests are rejected so you can continue directly.

Full round trip: global-settings schema + defaults, ExtensionState,
SettingsView control bound to cachedState with save payload, webviewMessageHandler
persistence through ContextProxy, ClineProvider getState/getStateToPostToWebview,
runtime consumers using the shared defaults. Includes focused tests for UI
binding/save, persistence, and the saved value returned to the webview.

Also completes the taskTree translation keys in all 17 non-English locales so
check-translations passes standalone (folded in from the series' CI fix).
…ached (#12)

Co-authored-by: Eason Liang <easonliang28@gmail.com>
Adds a three-level new_task chain e2e (root -> child -> depth-2 task) where the
depth-2 task's own new_task call exceeds maxNestingDepth and is flattened inline.
Asserts exactly 3 tasks exist, the flattened work completes in the depth-2 task's
own conversation, parent/child resume correctly, and the task stack never holds a
fourth task. Hand-written aimock fixtures (no API key needed) use unique
FLATTEN_E2E_ markers with predicate matching to avoid cross-suite collisions.
…n history tree (#13)

Co-authored-by: Eason Liang <easonliang28@gmail.com>
…hods (#14)

Co-authored-by: Eason Liang <easonliang28@gmail.com>
Co-authored-by: Eason Liang <easonliang28@gmail.com>
…nners (#17)

* feat(webview): surface inline-subtask transitions as distinct chat banners

* fix(webview): localize inline-subtask banner details and task-tree settings keys

---------

Co-authored-by: Eason Liang <easonliang28@gmail.com>
…ovider call (#18)

## Problem A/C — delegation link lost across interrupt/resume

When a delegated child is interrupted (cancelTask / evictCurrentTask), the parent's
`awaitingChildId` link is preserved only while the parent is still `delegated`. After a
crash or resume cycle the parent can be left `active` with no `awaitingChildId`, so when
the user resumes the child, AttemptCompletionTool refuses to route its completion back
(it requires `parent.awaitingChildId === this child`). The result: a resumed subtask's
result is silently stranded and never reported to the parent.

Fix: re-establish the link in `createTaskWithHistoryItem` (the common funnel for every
resume path) when resuming an *interrupted* child. New private helper
`reestablishDelegationLinkOnResume` transitions a demoted `active` parent back to
`delegated`, gated so it:
  - skips children whose delegation was intentionally severed (`cancelledDelegationChildIds`),
  - never clobbers a live delegation to a different child,
  - only performs the legal `active -> delegated` transition.
Non-fatal: any failure is logged and the resume proceeds without the link. Because
AttemptCompletionTool already accepts an `active` parent whose `awaitingChildId` matches,
routing (Problem C) works automatically once the link is restored.

## Problem D — remove `as any` in NewTaskTool

`task.providerRef` is a `WeakRef<ClineProvider>`, so after the null-check `provider` is
already typed `ClineProvider`. The `(provider as any).delegateParentAndOpenChild(...)`
cast was unnecessary; it is now a typed method call. This removes the last `as any` in
NewTaskTool.ts, so its stale `no-explicit-any` suppression entry is dropped from
eslint-suppressions.json (count never increases).

## Tests

Added 4 focused regression tests for `reestablishDelegationLinkOnResume` covering: link
restored on resume of an interrupted child with a demoted parent; no-op when already
delegated to the same child; never clobbers a live delegation to another child; and no
reattach for intentionally-severed children.

Co-authored-by: Eason Liang <easonliang28@gmail.com>
@easonLiangWorldedtech
easonLiangWorldedtech marked this pull request as ready for review August 17, 2026 02:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants