Skip to content

Commit 8abc76f

Browse files
ryansolidClaude via Cursorcursoragent
committed
test(signals): pin #4569 — a first load answered while pending by propagation is committed and pending (R2 ruled)
Ruled 2026-10-31: a node's question changes only when the node itself re-runs (#3373). a2:11, landing while a2 is pending by propagation from a0's re-ask and has no committed value, initializes a2: committed and pending on a0. Content that reads only its verdict reveals with pending (S-40); content that reads a2 holds until the chain settles (F-2, no screen to keep). The asymmetry is the rule. Frames: loading ×3, pending, pending, settled. The rejected reading (loading ×5, settled) is documented, not pinned: it needs a supersession by propagation no rule has and a new resting state, treats an uninitialized node more strictly than a committed one, and contradicts #4459/#4495's p:true first reveals. Docs: D-23 clause + Collapses ref; F-9 'What it does not say' line (on creation is the node's, not a boundary's first frame); S-40 verdict form with the frames and the plain-read control. No src change. Closes #4569 Co-authored-by: Claude via Cursor <[email protected]> Co-authored-by: Cursor <[email protected]>
1 parent 265a01c commit 8abc76f

5 files changed

Lines changed: 68 additions & 92 deletions

File tree

‎packages/signals/docs/RULES-DERIVED.md‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -188,9 +188,9 @@ An armed boundary catches a lane the same way it catches a transition (D-12, D-1
188188

189189
**From:** F-9, F-1. **Scenarios:** S-26, S-27.
190190

191-
`isPending(x)` is `true` while any of three causes holds: (i) a write held by a live hold; (ii) the node's own async in flight for a new question; (iii) a fresh value landed but held uncommitted by a hold it is entangled with — and `false` the moment none does. Exception 1: a node with no committed value — a first load, or a memo born held over a held value (D-2) — is loading, not pending. A node answered and then re-asked in the same flush — an async first answer that lands in the tick a write re-asks it — is not a first load (amended 2026-10-30): its answer is a produced value, so `latest` of it shows the answer and `isPending` of it is `true`, the re-ask being a new question; so is a derivation of it with no value of its own, at any depth (F-9, D-22). It has no verdict: inside an owner `isPending` of it does what its read does — a first load's `NotReadyError` goes to a boundary or holds the change (F-2); a reader of a memo born held is held with it, or caught by a boundary that has not shown content — and the reader runs at the commit, where the answer is final; unowned it is `false` (D-24). A node with a committed value is judged wherever its readers are: on screen, behind a fallback, or in a branch not yet attached. A verdict on content that may have no committed value is read inside the boundary that owns it — `<Loading><p>{m()}</p><span>{isPending(m) ? "updating" : ""}</span></Loading>` — where its reader is caught with the content; read outside, it holds the change that mounted it. Either way no answer is shown before the commit: `<input disabled={isPending(data)}>` is never seen enabled over an unjudged value (`packages/web/test/ispending-guard-born-held.spec.tsx`). No verdict is the honest answer: `true` would say "updating" for something never shown, and `false` would read as settled to a guard like `if (!isPending(x)) save(x())`. Exception 2: a re-ask of the same question is quiet; an `affects()` mark is a declared new question and pends exactly the marked data and what derives from it, released when the declaring hold settles (or at flush end, ambient). `[isPending(x), x()]` read in one scope never pairs `true` with the fresh value: a verdict reader sees the screen — in either read order — unless it is a node a hold's fetch is waiting on (D-25). `isPending(() => latest(x))` follows `x`'s own non-quiet async only — on a signal or sync memo the latest form is never pending — so `[isPending(() => latest(x)), latest(x)]` never pairs `true` with the fresh value either. A spinner driven by `isPending` fires while the owner's async is still in flight (it does not wait for the hold).
191+
`isPending(x)` is `true` while any of three causes holds: (i) a write held by a live hold; (ii) the node's own async in flight for a new question; (iii) a fresh value landed but held uncommitted by a hold it is entangled with — and `false` the moment none does. Exception 1: a node with no committed value — a first load, or a memo born held over a held value (D-2) — is loading, not pending. A node answered and then re-asked in the same flush — an async first answer that lands in the tick a write re-asks it — is not a first load (amended 2026-10-30): its answer is a produced value, so `latest` of it shows the answer and `isPending` of it is `true`, the re-ask being a new question; so is a derivation of it with no value of its own, at any depth (F-9, D-22). It has no verdict: inside an owner `isPending` of it does what its read does — a first load's `NotReadyError` goes to a boundary or holds the change (F-2); a reader of a memo born held is held with it, or caught by a boundary that has not shown content — and the reader runs at the commit, where the answer is final; unowned it is `false` (D-24). A node with a committed value is judged wherever its readers are: on screen, behind a fallback, or in a branch not yet attached. **A node initialized while pending by propagation is committed and pending (ruled 2026-10-31, #4569):** an answer that lands on a node with no committed value while an upstream re-ask has marked it pending by propagation — its own flight still current, nothing superseded it (#3373) — initializes the node: a node's question changes only when the node itself re-runs, so the answer is current, and the node is committed and pending on the upstream flight. Content that reads only its verdict reveals with `true` (S-40: a fresh boundary reading only a display-ahead channel shows what it has); content that reads the node holds until the chain settles — a first reveal has no screen to keep (F-2). That asymmetry is the rule, not a bug; a reading that kept the node loading would need a supersession by propagation no rule has, a new resting state, and would treat an uninitialized node more strictly than a committed one. A verdict on content that may have no committed value is read inside the boundary that owns it — `<Loading><p>{m()}</p><span>{isPending(m) ? "updating" : ""}</span></Loading>` — where its reader is caught with the content; read outside, it holds the change that mounted it. Either way no answer is shown before the commit: `<input disabled={isPending(data)}>` is never seen enabled over an unjudged value (`packages/web/test/ispending-guard-born-held.spec.tsx`). No verdict is the honest answer: `true` would say "updating" for something never shown, and `false` would read as settled to a guard like `if (!isPending(x)) save(x())`. Exception 2: a re-ask of the same question is quiet; an `affects()` mark is a declared new question and pends exactly the marked data and what derives from it, released when the declaring hold settles (or at flush end, ambient). `[isPending(x), x()]` read in one scope never pairs `true` with the fresh value: a verdict reader sees the screen — in either read order — unless it is a node a hold's fetch is waiting on (D-25). `isPending(() => latest(x))` follows `x`'s own non-quiet async only — on a signal or sync memo the latest form is never pending — so `[isPending(() => latest(x)), latest(x)]` never pairs `true` with the fresh value either. A spinner driven by `isPending` fires while the owner's async is still in flight (it does not wait for the hold).
192192

193-
**Collapses:** [A19](./SPEC-ASYNC-SEMANTICS.md#a19-ispendingx--the-observable-value-is-not-final-three-causes), [A24](./SPEC-ASYNC-SEMANTICS.md#a24-question-scoped-pending-pending-iff-a-value-change-is-in-flight-or-an-affects-mark-is-live), [A10](./SPEC-ASYNC-SEMANTICS.md#a10-ispendingx-x-is-atomic-within-one-scope), [A8](./SPEC-ASYNC-SEMANTICS.md#a8-ispending--latestx-follows-xs-own-async-only--verdicts-are-per-channel), [A14](./SPEC-ASYNC-SEMANTICS.md#a14-companion-nodes-get-child-lanes-that-do-not-merge-with-the-owner), [A20](./SPEC-ASYNC-SEMANTICS.md#a20-superseded-optimistic-writes-announce-a-store-wide-pending)/[A21](./SPEC-ASYNC-SEMANTICS.md#a21-superseded-the-store-wide-mask) (superseded — the mask is gone), recovery 6.3. **Status:** standing. A probe alone holds nothing; a plain read beside it holds (`optimistic-settle-verdicts.test.ts` #3409, `question-scoped-pending.test.ts` 3.4; #3977). `isPending` of a memo born held, read inside and outside the boundary: `tests/fresh-boundary-born-held-reader.test.ts` (#3977; the engine joins the reader to the hold as the read does, which is the throw's observable effect). **Known deviation (#4469, 2026-10-30):** a verdict reader is re-run at every landing of a hold it observed, even when the landing changes nothing it read; the gate that would skip it needs a deps walk. Measured 2026-10-30: two cheaper gates broke F-9/D-23 (verdict-memo-flight-no-reask) and #4189 (errored-store-probe-settled-3962) pins and failed the frozen floors by +31/+36 B. A plain reader of the same recovered hold is not in this: it shows the committed value for the whole hold and runs once, at the commit, since a recovery to a value equal to the last good one is still a change of answer (D-41, D-9). Pinned in `tests/hold-recovers-landing-reruns-4469.test.ts` (the verdict rows `it.fails`). A verdict reader of a signal whose writes net to its committed value in the tick an action opens a hold is not in this either (#4485, fixed 2026-10-30): the writes made no proposal (D-8), so the reader is not the hold's stale reader and `isPending(s)` stays `false` (`tests/netted-verdict-reader-4485.test.ts`).
193+
**Collapses:** [A19](./SPEC-ASYNC-SEMANTICS.md#a19-ispendingx--the-observable-value-is-not-final-three-causes), [A24](./SPEC-ASYNC-SEMANTICS.md#a24-question-scoped-pending-pending-iff-a-value-change-is-in-flight-or-an-affects-mark-is-live), [A10](./SPEC-ASYNC-SEMANTICS.md#a10-ispendingx-x-is-atomic-within-one-scope), [A8](./SPEC-ASYNC-SEMANTICS.md#a8-ispending--latestx-follows-xs-own-async-only--verdicts-are-per-channel), [A14](./SPEC-ASYNC-SEMANTICS.md#a14-companion-nodes-get-child-lanes-that-do-not-merge-with-the-owner), [A20](./SPEC-ASYNC-SEMANTICS.md#a20-superseded-optimistic-writes-announce-a-store-wide-pending)/[A21](./SPEC-ASYNC-SEMANTICS.md#a21-superseded-the-store-wide-mask) (superseded — the mask is gone), recovery 6.3. **Status:** standing. A probe alone holds nothing; a plain read beside it holds (`optimistic-settle-verdicts.test.ts` #3409, `question-scoped-pending.test.ts` 3.4; #3977). `isPending` of a memo born held, read inside and outside the boundary: `tests/fresh-boundary-born-held-reader.test.ts` (#3977; the engine joins the reader to the hold as the read does, which is the throw's observable effect). A first load answered while pending by propagation from an upstream re-ask, under a boundary (the issue's verdict-only content, a 3-deep chain, the plain-read controls that hold) and in a root: `tests/first-load-superseded-answer-4569.test.ts`, `packages/web/test/first-load-superseded-answer-4569.spec.tsx` (#4569, ruled 2026-10-31). **Known deviation (#4469, 2026-10-30):** a verdict reader is re-run at every landing of a hold it observed, even when the landing changes nothing it read; the gate that would skip it needs a deps walk. Measured 2026-10-30: two cheaper gates broke F-9/D-23 (verdict-memo-flight-no-reask) and #4189 (errored-store-probe-settled-3962) pins and failed the frozen floors by +31/+36 B. A plain reader of the same recovered hold is not in this: it shows the committed value for the whole hold and runs once, at the commit, since a recovery to a value equal to the last good one is still a change of answer (D-41, D-9). Pinned in `tests/hold-recovers-landing-reruns-4469.test.ts` (the verdict rows `it.fails`). A verdict reader of a signal whose writes net to its committed value in the tick an action opens a hold is not in this either (#4485, fixed 2026-10-30): the writes made no proposal (D-8), so the reader is not the hold's stale reader and `isPending(s)` stays `false` (`tests/netted-verdict-reader-4485.test.ts`).
194194

195195
### D-24. `isPending` never throws unowned; the probe is reads-only
196196

‎packages/signals/docs/RULES-FUNDAMENTALS.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -98,7 +98,7 @@ These are the only words the fundamentals use. Internal names (transaction, stag
9898

9999
**Immutable.** `isPending(x)` is `true` iff `x`'s value is not the final answer to a _new question_: an input changed and the new answer has not been committed, a landed answer waits in a hold, or in-flight work declared it will change the data (`affects()`). It is `false` the moment the value is final. Re-asking the same question — `refresh()`, polling, an action's confirming refetch — is quiet: the shown value still answers, `isPending` stays `false`, and the fresh value reveals silently; to make a reload read as pending, declare it (`affects(x); refresh(x)`). A node with no committed value has nothing to judge, so it has no verdict: a first load, or a memo born held over a held value (D-2). It is _loading_ — inside an owner `isPending` of it does what the read itself does: a first load's `NotReadyError` goes to a boundary or holds, and a reader of a memo born held is held with it (caught by a boundary that has not shown content, F-4, or held with the change, F-2) and runs at the commit, where the answer is final. No answer is shown before then. Unowned, `isPending` of it is `false` (D-24). `latest` of it is its pending value if it has one (a memo born held) and throws if it has none (a first load) (D-22). A declared first paint (`loadingValue`) is loading too. Optimistic values are verdict-inert: an optimistic write never makes its own slot pending and never silences anything. Pending is per node — per store key — never store-wide. For a node with a committed value, painting does not decide it: where `x`'s readers are — on screen, behind a fallback, in a branch not yet attached — does not change `isPending(x)`, and `latest(x)` is the newest value it has produced (D-22). A verdict on content that may have no committed value belongs inside the boundary that owns that content (D-23).
100100

101-
**What it does not say.** It does not say `isPending` can be `true` on creation (it cannot — hydration parity; a memo born held has no verdict until its commit). It does not say an action in progress is pending data (process affordances are co-written state, not verdicts).
101+
**What it does not say.** It does not say `isPending` can be `true` on creation (it cannot — hydration parity; a memo born held has no verdict until its commit). It does not say an action in progress is pending data (process affordances are co-written state, not verdicts). It does not say a boundary's first frame cannot show `true`: "on creation" is the node's, not the screen's — a node whose first answer lands beside the write that re-asks it (#4459), or while an upstream re-ask has it pending by propagation (#4569), is committed and pending, and content that reads only its verdict reveals with it (D-23, S-40).
102102

103103
**Collapses:** A19 (three causes, two exceptions), A24 (question-scoped; `affects`), A27 (the loading window), A7, A8 (per channel), A9, A12, A13, A14, A16, A22, A23, OL-R12, OS-R17/R40, PJ-R28–R31, recovery 6.3/6.4. **Derived:** D-22–D-26, D-30, D-36. **Scenarios:** S-14, S-25–S-29, S-36, S-37.
104104

‎packages/signals/docs/SCENARIOS.md‎

Lines changed: 17 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -698,6 +698,23 @@ _Ruled 2026-10-09, from the read-order brief's sibling case (416). Not D-28: thi
698698
| ✗ | | sibling `0`, slot `(empty)` | neither the fallback nor `1` — the boundary waited on `n0`, which it did not read |
699699
| ✗ | | sibling `0 \| loading` | the same wait, shown as a fallback |
700700

701+
**The verdict form (ruled 2026-10-31, #4569; D-23, #3373):** a fresh `<Loading>` whose content reads only `isPending` of a first load reveals when the verdict exists. `a0 = async(s)`, `a2 = async(await a0())`; `a0:0` lands, `a2` asks `a2:11`; `setS(1)` re-asks `a0` and marks `a2` pending by propagation while `a2:11` is up and `a2` has no value; `a2:11` lands. `a2`'s question changes only when `a2` re-runs, so the answer is current: `a2` is committed and pending on `a0`. Content that reads `a2` holds (F-2: no screen to keep) — the asymmetry is the rule.
702+
703+
```tsx
704+
<Loading fallback="loading">{isPending(() => a2()) ? "pending" : "settled"}</Loading>
705+
```
706+
707+
| # | Action | verdict-only content | control: `{a2()}` — alone or beside the probe | Notes |
708+
| --- | ---------------------------------------- | -------------------- | --------------------------------------------- | --------------------------------------------------------------------- |
709+
| 1 | mount; `a0:0` lands; `a2` asks `a2:11` | `loading` | `loading` | first load |
710+
| 2 | `setS(1)` — re-asks `a0:1` | `loading` | `loading` | `a2` pending by propagation, no value, own flight still current |
711+
| 3 | `a2:11` lands | `pending` | `loading` | `a2` initializes, committed and pending on `a0`; the plain read waits |
712+
| 4 | `a0:1` lands; `a2` re-runs, asks `a2:20` | `pending` | `loading` | |
713+
| 5 | `a2:20` lands | `settled` | `21` / `21 settled` | |
714+
| ✗ | | after 3: `loading` | | rejected: a first load kept loading by a supersession no rule has |
715+
716+
Pinned: `tests/first-load-superseded-answer-4569.test.ts` (also a 3-deep chain and the no-boundary root, which attaches at 3 with `pending`), `packages/web/test/first-load-superseded-answer-4569.spec.tsx`.
717+
701718
### S-45. A first answer landing in the same tick as an action that writes its input: `latest` shows it, the plain read waits — D-22, F-9 (async generator E9)
702719

703720
_Ruled 2026-10-28. The leaf form and the through-a-memo form show the same screen; today the leaf keeps the fallback up until the action ends — that is the ✗ row._

0 commit comments

Comments
 (0)