You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 290c551
Browse filesBrowse the repository at this point in the historyBrowse files
test(signals): shadow the per-hold release — a posture is toward the hold it was earned against, released at its landing
Per the 2026-11-01 rulings (D-25): a probe's posture is remembered per
(node, hold) and released when that hold lands (ruling 1); a pass that
probes T while holding a mere T staging is T's verdict reader (ruling 2,
today's order). The shadow table encodes both and the remaining classes
it surfaced were shadow gaps, fixed here:
- a probe made before the flush opens its hold is toward that hold
(`FLUSH_HOLD`, resolved at the open or at the flush's end);
- #3884: toward the hold whose fetch the node waits on at its head the
memory is spent, a body probe seats anew; toward other holds it stands;
- a boundary's display-ahead swap (#3540, #3528) seats the output's
passes as the hold's verdict reader while the hold lives;
- a frame (render effect, boundary structure) holding a mere T staging
re-run outside T's flush is a leaf toward T (D-3, #3404, #4283);
- a pass in T's verdict lane that did not probe T — entered by a read of
a lane node, or a render effect seated there by a verdict memo's
re-staging (D-26 #4284) — reads T's nodes as the screen (D-37; #4365)
but is no verdict reader of the flush's own proposals (`SEAT_VLANE`);
- `holdWork` (verdict.ts) drops the memory with CONFIG_VERDICT.
Gate: the 49 rows ruling 1 moves (4393 ×28, 4491 ×15, 4504 ×4,
flight-no-reask ×2) are the only rows in the 13 gate files; the full
suite's other non-lane rows are the #4544 clause (a probe of an unheld
node earns no posture) and D-37's amendment (lane rows).
Co-authored-by: Claude via Cursor <[email protected]>
Shadow seat (`__TEST__` only): a verdict posture is remembered per (node, hold) and released when that hold lands (D-25 as ruled 2026-11-01); a probe made before the flush opens its hold resolves to that hold; a lane pass that did not probe reads the hold's nodes as the screen but is no verdict reader of the flush's own proposals (`SEAT_VLANE`); a frame holding a mere staging outside its flush is a leaf toward it.
Copy file name to clipboardExpand all lines: packages/signals/docs/RULES-DERIVED.md
+2-2Lines changed: 2 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -204,9 +204,9 @@ An armed boundary catches a lane the same way it catches a transition (D-12, D-1
204
204
205
205
**From:** F-6, F-1. **Scenarios:** S-29.
206
206
207
-
What is rendered from a display-ahead read reveals together with it: an async memo over `latest(x)` holds the `latest` view's reveal until it lands; a combined `isPending(() => [fast(), copy()])` over two async memos, one wrapped in a sync memo, stays `true` until both land, because a memo's value is one shared slot — it computes under its own hold posture, never under the posture of whichever reader pulled it. A node whose fetch a hold is waiting on is that hold's work, not a verdict reader: a pass that re-runs it while its fetch is in flight reads plainly — following the change it is fetching for — even if an earlier pass of it called `isPending` or `latest` (#3884). A memo's posture is what it reads: one that reads a verdict reader's value is a derivation of a display-ahead read, and a verdict reader itself — `createMemo(() => [held(), V()])` with `V = createMemo(() => isPending(x) && latest(e))` shows `V`'s verdict now, beside the screen's `held()`, as the inline probe and a render effect reading both do (#4050: a Provider's children memo over a router shell). **The probe decides the pass's posture in either read order (clarified 2026-10-28):** a pass that probes is a verdict reader whether the probe came before or after a plain read of the same held node — `[x(), isPending(x)]` and `[isPending(x), x()]` both show `[committed, true]` now and re-derive at the landing (A10) — unless the pass is **that** hold's own work: holding a staging of its or a fetch it waits on (#3884), born in its world (D-27, #4341), or dirtied by its flush having probed nothing of it (D-2 in-flush). A pass dirtied by T's flush that _also_ probes T is T's verdict reader — the probe decides (F-6), and only an existing staging or birth overrides it. Work beats verdict, per hold: the seat of a pass toward a hold T is decided once, when the pass starts, as exactly one of _T's work_, _T's verdict reader_ (it probed a T-held node or inherited that from a verdict memo of T, #4050, and is not T's work), or _mainline toward T_. A pass may hold different seats toward different holds (U's verdict reader and T's work: #4544 timing edge). A probe of a node no hold holds earns no posture toward any hold (#4544). A pass that read two holds' display-ahead values is one reveal unit of both (A15 lanes corollary, #3335): it waits for either lane's flight, never showing one hold's `latest()` beside a half-computed derivation of the other's.
207
+
What is rendered from a display-ahead read reveals together with it: an async memo over `latest(x)` holds the `latest` view's reveal until it lands; a combined `isPending(() => [fast(), copy()])` over two async memos, one wrapped in a sync memo, stays `true` until both land, because a memo's value is one shared slot — it computes under its own hold posture, never under the posture of whichever reader pulled it. A node whose fetch a hold is waiting on is that hold's work, not a verdict reader: a pass that re-runs it while its fetch is in flight reads plainly — following the change it is fetching for — even if an earlier pass of it called `isPending` or `latest` (#3884). A memo's posture is what it reads: one that reads a verdict reader's value is a derivation of a display-ahead read, and a verdict reader itself — `createMemo(() => [held(), V()])` with `V = createMemo(() => isPending(x) && latest(e))` shows `V`'s verdict now, beside the screen's `held()`, as the inline probe and a render effect reading both do (#4050: a Provider's children memo over a router shell). **The probe decides the pass's posture in either read order (clarified 2026-10-28):** a pass that probes is a verdict reader whether the probe came before or after a plain read of the same held node — `[x(), isPending(x)]` and `[isPending(x), x()]` both show `[committed, true]` now and re-derive at the landing (A10) — unless the pass is **that** hold's own work: a fetch it waits on (#3884), born in its world (D-27, #4341), or dirtied by its flush having probed nothing of it (D-2 in-flush). A pass dirtied by T's flush that _also_ probes T is T's verdict reader — the probe decides (F-6), and only a fetch T waits on or birth overrides it. A node holding a mere T staging that probes T is T's verdict reader too: the probe wins, the staging is dropped, and T's landing re-derives it as a stale reader (ruled 2026-11-01; `unread-derive-midflight-mount-4491`, `midflight-memo-over-reask-4504`, `latest-first-one-pass-4402`). Per hold, the seat of a pass toward a hold T is decided once, when the pass starts, as exactly one of _T's work_, _T's verdict reader_ (it probed a T-held node or inherited that from a verdict memo of T, #4050, and is not T's work), or _mainline toward T_. A pass may hold different seats toward different holds (U's verdict reader and T's work: #4544 timing edge). A probe of a node no hold holds earns no posture toward any hold (#4544). A pass that read two holds' display-ahead values is one reveal unit of both (A15 lanes corollary, #3335): it waits for either lane's flight, never showing one hold's `latest()` beside a half-computed derivation of the other's.
208
208
209
-
**A verdict posture outlives the pass that asked (ruled 2026-10-28):** a pass that asked (`isPending`, `latest`) is the hold's verdict reader on every later run of its node until one completes having asked nothing — that run still reads as a verdict reader; the next is an ordinary derivation. On each such run its plain read of a node the hold holds is the screen's (a frame reader, A10; re-derived at the landing as A15's stale reader) and it joins nothing, whoever dirtied it — a sync write, a boundary redraw, a landing. The hold's own work is not in this: a node whose fetch the hold waits on, or a pass born in its world, reads plainly whatever it asked before (the #3884 clause and D-27 #4341 above, unchanged). A boundary's display-ahead swap is this posture handed on: an `on` pass that read `latest`/`isPending` is a verdict reader, and the fallback swap it asks for runs as that reader's work — the output's pass reads nothing of the held frame (a pending tree is linked, not read) and its answer, the fallback, is shown now beside the held frame and stays the screen until the landing, instead of being staged with the write's frame as a plain `on`'s swap is (D-15): `[A] → [A + spinner] → [B + spinner] → [B + comments]`, never `[A] → [B + spinner]`. Only the `on` pass's asking counts: `latest()` inside the fallback is the fallback's own display-ahead, not the swap's — a plain `on` over an action-held input lands the swap with the frame whatever the fallback reads (#4337). The posture is remembered per (node, hold): the node carries which holds it is a verdict reader of, and the node-level memory (`CONFIG_VERDICT`) remains only for what is not per hold — the probed node is read either-order in the same pass, and a dependency going pending re-derives the reader instead of marking it. Membership in a verdict lane follows the posture; it does not define it. Pins: `tests/loading-on-rearm-3540.test.ts`, `tests/loading-on-frame-following-3540.test.ts` §4–5, `tests/loading-on-latest-fallback-4337.test.ts`.
209
+
**A verdict posture outlives the pass that asked (ruled 2026-10-28):** a pass that asked (`isPending`, `latest`) is the hold's verdict reader on every later run of its node until one completes having asked nothing — that run still reads as a verdict reader; the next is an ordinary derivation. On each such run its plain read of a node the hold holds is the screen's (a frame reader, A10; re-derived at the landing as A15's stale reader) and it joins nothing, whoever dirtied it — a sync write, a boundary redraw, a landing. The hold's own work is not in this: a node whose fetch the hold waits on, or a pass born in its world, reads plainly whatever it asked before (the #3884 clause and D-27 #4341 above, unchanged). A boundary's display-ahead swap is this posture handed on: an `on` pass that read `latest`/`isPending` is a verdict reader, and the fallback swap it asks for runs as that reader's work — the output's pass reads nothing of the held frame (a pending tree is linked, not read) and its answer, the fallback, is shown now beside the held frame and stays the screen until the landing, instead of being staged with the write's frame as a plain `on`'s swap is (D-15): `[A] → [A + spinner] → [B + spinner] → [B + comments]`, never `[A] → [B + spinner]`. Only the `on` pass's asking counts: `latest()` inside the fallback is the fallback's own display-ahead, not the swap's — a plain `on` over an action-held input lands the swap with the frame whatever the fallback reads (#4337). The posture outlives the _pass_, not the _hold_ (ruled 2026-11-01): a verdict posture, earned by a probe or inherited (#4050), is toward the hold it was earned against and is released when that hold lands; toward a later, _separate_ hold U the node is whatever its reads make it — U's work if U's write dirtied it (D-2 in-flush), mainline otherwise. _Separate_ means isolated: T fully landed before U's write. If T was still open when U's write touched anything T held, D-4 merges them and the posture resolves through the merge to the one hold — the node stays its verdict reader exactly as before. The posture is remembered per (node, hold): the node carries which holds it is a verdict reader of, and the node-level memory (`CONFIG_VERDICT`) remains only for what is not per hold — the probed node is read either-order in the same pass, and a dependency going pending re-derives the reader instead of marking it. Membership in a verdict lane follows the posture; it does not define it. Pins of the release: `tests/latest-two-readers-one-ask-4393.test.ts`, `tests/unread-derive-midflight-mount-4491.test.ts`, `tests/midflight-memo-over-reask-4504.test.ts`, `tests/verdict-memo-flight-no-reask.test.ts` (the rows listed in PR #TBD). Pins: `tests/loading-on-rearm-3540.test.ts`, `tests/loading-on-frame-following-3540.test.ts` §4–5, `tests/loading-on-latest-fallback-4337.test.ts`.
210
210
211
211
**Collapses:** [A31](./SPEC-ASYNC-SEMANTICS.md#a31-a-memo-computes-under-its-own-lane-posture-never-its-pullers) (#3442), §20.1 question 3 ("Yeah definitely it is held. Consistent with any optimistic update"), OL-R24/R25/R34, recovery 6.2. **Status:** standing; sharpened per hold 2026-11-01 (#4544, `documentation/plans/seat-per-hold.md`): a probe of an unheld node earns no posture toward any hold, so a memo over `isPending(plan)` with `plan` unheld is T's plain derivation when T's write dirties it (#4544 Shape 1: the branch reads `signedIn: true`, T lands); a pass that probes T's own node is T's verdict reader in either read order (Shape 2, A10, unchanged); a pass that is U's verdict reader and dirtied by T's write is T's work (the timing edge: `{plan: false, signedIn: true}`, T lands first). Pins: `tests/verdict-memo-held-branch-4544.test.ts`, `packages/web/test/verdict-memo-held-branch-4544.spec.tsx`. A memo over a verdict memo: `tests/verdict-memo-under-held-memo-4050.test.ts`, `packages/web/test/held-shell-pending-feedback-4050.spec.tsx` (U-10).
| V4 | fixed |`docs/SPEC-ASYNC-SEMANTICS.md:503`| seat.ts×4| spec-async-semantics.test.ts×5 | - **V4 (violated the old A20's three-form algebra) — FIXED, then the rule it |
89
+
| V4 | fixed |`docs/SPEC-ASYNC-SEMANTICS.md:503`| seat.ts×6| spec-async-semantics.test.ts×5 | - **V4 (violated the old A20's three-form algebra) — FIXED, then the rule it |
90
90
| V5 | live |`docs/SPEC-ASYNC-SEMANTICS.md:515`| scheduler.ts×1 | spec-async-semantics.test.ts×3 | - **V5 (A17 corollary — found and fixed with the revert-target elimination, |
91
91
## B — tier B
92
92
@@ -380,7 +380,7 @@ Status legend: **live** stated and standing · **ruled** carries an explicit rul
380
380
|---|---|---|---|---|---|
381
381
| §1 | live |`docs/INTERNALS-STORE-STATE.md:27`| seat.ts×1 target.ts×1 | reveal-gating-contract.test.ts×1 | Storage model (the single-home rule) |
0 commit comments