sorokin.oleg.edu/learnharness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 6 Wiki Activity

practice: ONE send lane — the register is read from the text

Browse Source 
Owner round 2 (2026-10-05), Slice H: the Nudge/Commit two-lane buttons
are gone. One Send; Enter sends, Shift+Enter is a newline. Whether a
submission claims an answer is apparent from the thinking itself — the
agent reads it (hint ladder for traces, moves+verdict for claims).

- The thinking input loses the commit field end-to-end: ChatInput,
  parseChatInput, newThinkingMessage (details and the model-facing label
  both collapse to the single 'Your thinking'), the optimistic marker.
- Legacy records' commit flags are deliberately unread.
- Doctrine: the two labeled channels become one read instruction; 'answer
  commits' section retitled 'Claimed answers' — a claim is not a
  progress signal.
- ThinkingCard: one label, collapsed arrival (the unified row).
- The dock placeholder no longer explains lanes that no longer exist.
This commit is contained in:
amosblomqvist
2026-10-05 03:46:32 +03:00
parent f460a90be9
commit 0bd3236af0
13 changed files with 101 additions and 130 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/session-core.test.ts
+6 -6
View File
@@ -1012,11 +1012,11 @@ describe("the live-view law — live is positive evidence, not session busyness"
describe("absorbed — the custom role (the doubled thinking card)", () => {
it("absorbs the optimistic thinking marker by customType + details", () => {
const optimistic = { ...newThinkingMessage("call-p1", false, "my trace"), id: "c_opt1" } as LhMessage;
const serverCopy = { ...newThinkingMessage("call-p1", false, "my trace"), id: "srv_1" } as LhMessage;
const optimistic = { ...newThinkingMessage("call-p1", "my trace"), id: "c_opt1" } as LhMessage;
const serverCopy = { ...newThinkingMessage("call-p1", "my trace"), id: "srv_1" } as LhMessage;
assert.ok(absorbed(optimistic, [...poseRecord(), serverCopy]), "same submission → the local copy drops");
// A DIFFERENT submission (commit flag, text) is its own marker.
const other = { ...newThinkingMessage("call-p1", true, "my trace"), id: "srv_2" } as LhMessage;
// A DIFFERENT submission (different text) is its own marker.
const other = { ...newThinkingMessage("call-p1", "a later, different trace"), id: "srv_2" } as LhMessage;
assert.ok(!absorbed(optimistic, [...poseRecord(), other]), "different details → not the same submission");
// The c_ gate: server stamps are never absorbed, whatever the role.
assert.ok(!absorbed(serverCopy, [optimistic]), "server ids never drop");
@@ -1027,14 +1027,14 @@ describe("absorbed — the custom role (the doubled thinking card)", () => {
// the tail (client id), then the +400ms adoption fetch swaps in the
// write-ahead record (server id). ONE card must remain.
const m = model({ chatId: "chat_x", initialMessages: poseRecord() });
const optimistic = { ...newThinkingMessage("call-p1", false, "my trace"), id: "c_opt1" } as LhMessage;
const optimistic = { ...newThinkingMessage("call-p1", "my trace"), id: "c_opt1" } as LhMessage;
m.applyWireEvent({ type: "message_start", message: optimistic });
assert.equal(
m.getState().nativeMessages.filter((x) => (x as { role?: string }).role === "custom").length,
1,
"the optimistic marker is in the tail",
);
const serverCopy = { ...newThinkingMessage("call-p1", false, "my trace"), id: "srv_1" } as LhMessage;
const serverCopy = { ...newThinkingMessage("call-p1", "my trace"), id: "srv_1" } as LhMessage;
m.adoptRecord([...poseRecord(), serverCopy]);
const customs = m.getState().nativeMessages.filter((x) => (x as { role?: string }).role === "custom");
assert.equal(customs.length, 1, "the server copy replaced the optimistic one — no doubling");