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/practice-flow.test.ts
+9 -8
View File
@@ -137,15 +137,16 @@ describe("practice flow (scripted model)", () => {
assert.equal(needsContinuation(afterPose), false);
assert.equal(activeProblemOf(afterPose)?.toolCallId, "call-p1");
// ── Turn 2: the commit arrives as a labeled thinking message ──
// ── Turn 2: the submission arrives as a single-lane thinking message ──
await drain(
await handleChatRequest(
post(chat.id, { type: "thinking", problemCallId: "call-p1", commit: true, text: "|G|=p; ⟨x⟩ has order dividing p" }),
post(chat.id, { type: "thinking", problemCallId: "call-p1", text: "|G|=p; ⟨x⟩ has order dividing p" }),
factory,
),
);
// The model saw the COMMIT label, not the hint label.
assert.ok(lastSaw(calls, "Committed answer:"), "the conclude turn's prompt carries the committed answer");
// The model saw the one label (owner round 2, 2026-10-05): the
// register is read from the text, not declared by a button.
assert.ok(lastSaw(calls, "Your thinking:"), "the conclude turn's prompt carries the thinking");
const afterConclude = chatStore.getChat(chat.id)!.messages;
const concludeAck = afterConclude.find(
(m) => (m as ToolResultMessage).role === "toolResult" && (m as ToolResultMessage).toolName === "conclude",
@@ -157,7 +158,7 @@ describe("practice flow (scripted model)", () => {
assert.equal(activeProblemOf(afterConclude), null);
});
it("the nudge (commit:false) carries the hint label", async () => {
it("a working-trace submission keeps the problem active", async () => {
const chat = createPracticeChat("Test: practice nudge");
const { engine, calls } = scripted([
() => fauxAssistantMessage([fauxToolCall("problem", POSE, { id: "call-p2" })], { stopReason: "toolUse" }),
@@ -167,12 +168,12 @@ describe("practice flow (scripted model)", () => {
await drain(await handleChatRequest(post(chat.id, { type: "text", text: "go" }), factory));
await drain(
await handleChatRequest(
post(chat.id, { type: "thinking", problemCallId: "call-p2", commit: false, text: "stuck at the second step" }),
post(chat.id, { type: "thinking", problemCallId: "call-p2", text: "stuck at the second step" }),
factory,
),
);
assert.ok(lastSaw(calls, "Thinking trace (working, not committed):"), "the nudge turn's prompt carries the hint label");
// The problem is still active — a nudge never concludes.
assert.ok(lastSaw(calls, "Your thinking:"), "the turn's prompt carries the thinking");
// The problem is still active — a trace never concludes.
assert.equal(activeProblemOf(chatStore.getChat(chat.id)!.messages)?.toolCallId, "call-p2");
});