can1357
fceff5e6b6
- Added section-based plan parsing with per-section delete, undo, and annotate. - Added Tab/Shift+Tab focus regions and per-line scroll navigation. - Added Refine feedback loop emitting annotations back to the model. - Added `OverlayOptions.fullscreen` borrowing the terminal's alt screen buffer.
945 lines
37 KiB
TypeScript
945 lines
37 KiB
TypeScript
import { afterEach, beforeAll, beforeEach, describe, expect, it, vi } from "bun:test";
|
|
import * as path from "node:path";
|
|
import { Agent } from "@oh-my-pi/pi-agent-core";
|
|
import type { AssistantMessage, Usage } from "@oh-my-pi/pi-ai";
|
|
import { resetSettingsForTest, Settings } from "@oh-my-pi/pi-coding-agent/config/settings";
|
|
import { resolveLocalUrlToPath } from "@oh-my-pi/pi-coding-agent/internal-urls";
|
|
import { AssistantMessageComponent } from "@oh-my-pi/pi-coding-agent/modes/components/assistant-message";
|
|
import { initTheme } from "@oh-my-pi/pi-coding-agent/modes/theme/theme";
|
|
import { SILENT_ABORT_MARKER } from "@oh-my-pi/pi-coding-agent/session/messages";
|
|
import { formatNumber, TempDir } from "@oh-my-pi/pi-utils";
|
|
import { ModelRegistry } from "../src/config/model-registry";
|
|
import type { HookSelectorSlider } from "../src/modes/components/hook-selector";
|
|
import { InteractiveMode } from "../src/modes/interactive-mode";
|
|
import { AgentSession } from "../src/session/agent-session";
|
|
import { AuthStorage } from "../src/session/auth-storage";
|
|
import { SessionManager } from "../src/session/session-manager";
|
|
|
|
/**
|
|
* Matches the plan-approved synthetic-prompt dispatch. `#approvePlan` calls
|
|
* `session.prompt(rendered, { synthetic: true })` exclusively for that case,
|
|
* so the `synthetic: true` option flag is the unique discriminator.
|
|
*/
|
|
const isPlanApprovedCall = (args: unknown[]): boolean =>
|
|
args.length >= 2 &&
|
|
typeof args[0] === "string" &&
|
|
typeof args[1] === "object" &&
|
|
args[1] !== null &&
|
|
(args[1] as { synthetic?: boolean }).synthetic === true;
|
|
|
|
function usageWithInput(input: number): Usage {
|
|
return {
|
|
input,
|
|
output: 0,
|
|
cacheRead: 0,
|
|
cacheWrite: 0,
|
|
totalTokens: input,
|
|
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },
|
|
};
|
|
}
|
|
|
|
function assistantWithUsage(overrides: Partial<AssistantMessage> = {}): AssistantMessage {
|
|
return {
|
|
role: "assistant",
|
|
content: [],
|
|
api: "anthropic-messages",
|
|
provider: "anthropic",
|
|
model: "test",
|
|
usage: usageWithInput(0),
|
|
stopReason: "stop",
|
|
timestamp: Date.now(),
|
|
...overrides,
|
|
};
|
|
}
|
|
|
|
function compactNumber(value: number): string {
|
|
return formatNumber(value).toLowerCase();
|
|
}
|
|
|
|
describe("InteractiveMode plan review rendering", () => {
|
|
let tempDir: TempDir;
|
|
let authStorage: AuthStorage;
|
|
let session: AgentSession;
|
|
let mode: InteractiveMode;
|
|
|
|
beforeAll(() => {
|
|
initTheme();
|
|
});
|
|
|
|
beforeEach(async () => {
|
|
resetSettingsForTest();
|
|
tempDir = TempDir.createSync("@pi-plan-review-");
|
|
await Settings.init({ inMemory: true, cwd: tempDir.path() });
|
|
authStorage = await AuthStorage.create(path.join(tempDir.path(), "testauth.db"));
|
|
authStorage.setRuntimeApiKey("anthropic", "test-key");
|
|
const modelRegistry = new ModelRegistry(authStorage);
|
|
const model = modelRegistry.find("anthropic", "claude-sonnet-4-5");
|
|
if (!model) {
|
|
throw new Error("Expected claude-sonnet-4-5 to exist in registry");
|
|
}
|
|
|
|
session = new AgentSession({
|
|
agent: new Agent({
|
|
initialState: {
|
|
model,
|
|
systemPrompt: ["Test"],
|
|
tools: [],
|
|
messages: [],
|
|
},
|
|
}),
|
|
sessionManager: SessionManager.create(tempDir.path(), tempDir.path()),
|
|
settings: Settings.isolated(),
|
|
modelRegistry,
|
|
});
|
|
mode = new InteractiveMode(session, "test");
|
|
});
|
|
|
|
afterEach(async () => {
|
|
vi.restoreAllMocks();
|
|
const currentMode = mode;
|
|
const currentSession = session;
|
|
const currentAuthStorage = authStorage;
|
|
const currentTempDir = tempDir;
|
|
mode = undefined as unknown as InteractiveMode;
|
|
session = undefined as unknown as AgentSession;
|
|
authStorage = undefined as unknown as AuthStorage;
|
|
tempDir = undefined as unknown as TempDir;
|
|
currentMode?.stop();
|
|
await currentSession?.dispose();
|
|
currentAuthStorage?.close();
|
|
currentTempDir?.removeSync();
|
|
resetSettingsForTest();
|
|
});
|
|
|
|
it("forwards each submitted plan to the review overlay", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# First plan\n\nalpha");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
const review = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://PLAN.md",
|
|
});
|
|
|
|
expect(review.mock.calls[0]?.[0]).toContain("First plan");
|
|
|
|
await Bun.write(resolvedPlanPath, "# Second plan\n\nbeta");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://PLAN.md",
|
|
});
|
|
|
|
// Each approval shows the current plan in the overlay, not a stale one.
|
|
expect(review.mock.calls[1]?.[0]).toContain("Second plan");
|
|
expect(review.mock.calls[1]?.[0]).not.toContain("First plan");
|
|
});
|
|
|
|
it("re-prompts the model with annotation feedback when Refine is chosen", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nbody");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
const feedback = "Refinement feedback on the plan:\n\n## Goal\n- needs detail\n";
|
|
// The overlay reports annotation feedback through onFeedbackChange before the
|
|
// operator picks "Refine plan".
|
|
vi.spyOn(mode, "showPlanReview").mockImplementation(async (_plan, _title, _options, dialogOptions) => {
|
|
dialogOptions?.onFeedbackChange?.(feedback);
|
|
return "Refine plan";
|
|
});
|
|
const startSpy = vi
|
|
.spyOn(mode, "startPendingSubmission")
|
|
.mockReturnValue({ text: feedback, cancelled: false, started: false });
|
|
const onInput = vi.fn();
|
|
mode.onInputCallback = onInput;
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://PLAN.md",
|
|
});
|
|
|
|
expect(startSpy).toHaveBeenCalledWith(expect.objectContaining({ text: expect.stringContaining("needs detail") }));
|
|
expect(onInput).toHaveBeenCalledTimes(1);
|
|
});
|
|
|
|
it("Refine with no annotations does not re-prompt the model", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nbody");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
const startSpy = vi.spyOn(mode, "startPendingSubmission");
|
|
const onInput = vi.fn();
|
|
mode.onInputCallback = onInput;
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://PLAN.md",
|
|
});
|
|
|
|
expect(startSpy).not.toHaveBeenCalled();
|
|
expect(onInput).not.toHaveBeenCalled();
|
|
});
|
|
|
|
it("approves with in-overlay edits and mirrors them to the plan file", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\noriginal body\n");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
const edited = "# Plan\n\nedited body\n";
|
|
vi.spyOn(mode, "showPlanReview").mockImplementation(async (_plan, _title, _options, dialogOptions) => {
|
|
dialogOptions?.onPlanEdited?.(edited);
|
|
return "Approve and execute";
|
|
});
|
|
vi.spyOn(mode, "handleClearCommand").mockResolvedValue();
|
|
const promptSpy = vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://PLAN.md",
|
|
});
|
|
|
|
// The synthetic plan-approved prompt carries the in-overlay edit, not the
|
|
// stale on-disk content (preferring editedContent avoids the write race).
|
|
const call = promptSpy.mock.calls.find(isPlanApprovedCall);
|
|
expect(call).toBeDefined();
|
|
expect(call?.[0] as string).toContain("edited body");
|
|
expect(call?.[0] as string).not.toContain("original body");
|
|
// onPlanEdited mirrored the edit to the plan file.
|
|
expect(await Bun.file(resolvedPlanPath).text()).toContain("edited body");
|
|
});
|
|
|
|
it("offers approve-and-keep-context as a distinct plan approval path", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nDo the thing.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: 7320, contextWindow: 10000, percent: 73.2 });
|
|
const selector = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://APPROVED.md",
|
|
});
|
|
|
|
expect(selector).toHaveBeenCalledWith(
|
|
expect.any(String),
|
|
"Plan mode - next step",
|
|
[
|
|
"Approve and execute",
|
|
"Approve and compact context",
|
|
"Approve and keep context (~7.3k / 10k)",
|
|
"Refine plan",
|
|
],
|
|
expect.any(Object),
|
|
expect.any(Object),
|
|
);
|
|
});
|
|
|
|
it("ignores aborted zero-usage assistant messages when estimating context usage", () => {
|
|
session.agent.appendMessage(assistantWithUsage({ usage: usageWithInput(7320), stopReason: "stop" }));
|
|
session.agent.appendMessage(assistantWithUsage({ usage: usageWithInput(0), stopReason: "aborted" }));
|
|
|
|
expect(session.getContextUsage({ contextWindow: 10000 })).toMatchObject({
|
|
tokens: 7320,
|
|
contextWindow: 10000,
|
|
percent: 73.2,
|
|
});
|
|
});
|
|
|
|
it("measures keep-context approval against the execution model restored after plan mode", async () => {
|
|
mode.stop();
|
|
await session.dispose();
|
|
|
|
const modelRegistry = new ModelRegistry(authStorage);
|
|
const executionModel = modelRegistry.find("anthropic", "claude-sonnet-4-5");
|
|
const planModel = modelRegistry.find("anthropic", "claude-opus-4-6");
|
|
if (!executionModel?.contextWindow || !planModel?.contextWindow) {
|
|
throw new Error("Expected test models with context windows");
|
|
}
|
|
session = new AgentSession({
|
|
agent: new Agent({
|
|
initialState: {
|
|
model: executionModel,
|
|
systemPrompt: ["Test"],
|
|
tools: [],
|
|
messages: [],
|
|
},
|
|
}),
|
|
sessionManager: SessionManager.create(tempDir.path(), tempDir.path()),
|
|
settings: Settings.isolated({ modelRoles: { plan: `anthropic/${planModel.id}` } }),
|
|
modelRegistry,
|
|
});
|
|
mode = new InteractiveMode(session, "test");
|
|
|
|
await mode.handlePlanModeCommand();
|
|
expect(session.model?.id).toBe(planModel.id);
|
|
|
|
const planFilePath = mode.planModePlanFilePath ?? "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nUse execution context.");
|
|
|
|
const tokens = 180000;
|
|
const contextSpy = vi.spyOn(session, "getContextUsage").mockImplementation(options => {
|
|
const contextWindow = options?.contextWindow ?? 0;
|
|
return {
|
|
tokens,
|
|
contextWindow,
|
|
percent: (tokens / contextWindow) * 100,
|
|
};
|
|
});
|
|
const selector = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://APPROVED.md",
|
|
});
|
|
|
|
expect(contextSpy).toHaveBeenCalledWith({ contextWindow: executionModel.contextWindow });
|
|
expect(selector.mock.calls[0]?.[2]).toEqual([
|
|
"Approve and execute",
|
|
"Approve and compact context",
|
|
`Approve and keep context (~${compactNumber(tokens)} / ${compactNumber(executionModel.contextWindow)})`,
|
|
"Refine plan",
|
|
]);
|
|
});
|
|
|
|
it("disables keep-context approval when execution context usage is above ninety-five percent", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nToo much context.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: 9600, contextWindow: 10000, percent: 96 });
|
|
const selector = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://APPROVED.md",
|
|
});
|
|
|
|
expect(selector.mock.calls[0]?.[3]).toEqual(
|
|
expect.objectContaining({
|
|
disabledIndices: [2],
|
|
}),
|
|
);
|
|
});
|
|
|
|
it("keeps keep-context approval enabled at exactly ninety-five percent", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nAt the threshold.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: 9500, contextWindow: 10000, percent: 95 });
|
|
const selector = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://APPROVED.md",
|
|
});
|
|
|
|
expect(selector.mock.calls[0]?.[3]).toEqual(
|
|
expect.objectContaining({
|
|
disabledIndices: undefined,
|
|
}),
|
|
);
|
|
});
|
|
|
|
it("keeps the keep-context label plain when context usage is unknown", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nDo the thing.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
// Post-compaction: tokens unknown until the next LLM response.
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: null, contextWindow: 200000, percent: null });
|
|
const selector = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Refine plan");
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath: "local://APPROVED.md",
|
|
});
|
|
|
|
expect(selector).toHaveBeenCalledWith(
|
|
expect.any(String),
|
|
"Plan mode - next step",
|
|
["Approve and execute", "Approve and compact context", "Approve and keep context", "Refine plan"],
|
|
expect.any(Object),
|
|
expect.any(Object),
|
|
);
|
|
});
|
|
|
|
it("approves a plan without clearing the session when keeping context", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
const resolvedFinalPlanPath = resolveLocalUrlToPath(finalPlanFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nKeep context.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: null, contextWindow: 200000, percent: null });
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and keep context");
|
|
const clear = vi.spyOn(mode, "handleClearCommand").mockResolvedValue();
|
|
const prompt = vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(clear).not.toHaveBeenCalled();
|
|
expect(await Bun.file(resolvedFinalPlanPath).text()).toBe("# Plan\n\nKeep context.");
|
|
expect(prompt).toHaveBeenCalledWith(expect.any(String), {
|
|
synthetic: true,
|
|
});
|
|
});
|
|
|
|
it("keeps the existing approve-and-execute path clearing the session", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nClear context.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and execute");
|
|
const clear = vi.spyOn(mode, "handleClearCommand").mockResolvedValue();
|
|
const prompt = vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(clear).toHaveBeenCalledTimes(1);
|
|
expect(prompt).toHaveBeenCalledWith(expect.any(String), {
|
|
synthetic: true,
|
|
});
|
|
});
|
|
|
|
it("executes on the slider-selected tier, surviving #exitPlanMode's model restore", async () => {
|
|
// Regression: the model-tier slider's choice used to be applied BEFORE
|
|
// #approvePlan ran. #approvePlan → #exitPlanMode restores the model that
|
|
// was active before plan mode (#planModePreviousModelState), which silently
|
|
// reverted the operator's pick — sliding to "slow" still executed on the
|
|
// default model. The fix defers application until after the plan-mode exit.
|
|
authStorage.setRuntimeApiKey("anthropic", "test-key");
|
|
const slow = session.modelRegistry.find("anthropic", "claude-opus-4-5");
|
|
const def = session.modelRegistry.find("anthropic", "claude-sonnet-4-5");
|
|
if (!slow || !def) throw new Error("Expected sonnet + opus to exist in registry");
|
|
|
|
// plan === default === the session model: this is what makes plan-mode entry
|
|
// record a previous-model state for #exitPlanMode to restore. slow differs,
|
|
// so an early application would be clobbered by that restore.
|
|
session.settings.setModelRole("default", "anthropic/claude-sonnet-4-5");
|
|
session.settings.setModelRole("slow", "anthropic/claude-opus-4-5");
|
|
session.settings.setModelRole("plan", "anthropic/claude-sonnet-4-5");
|
|
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nRun this on the slow tier.");
|
|
|
|
await mode.handlePlanModeCommand();
|
|
expect(session.getPlanModeState()?.enabled).toBe(true);
|
|
expect(session.model?.id).toBe(def.id);
|
|
|
|
// Keep-context path avoids newSession() so the assertion isolates the
|
|
// exit-plan-mode restore from session-clear effects.
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: null, contextWindow: 200000, percent: null });
|
|
vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
let observedSegments: string[] = [];
|
|
vi.spyOn(mode, "showPlanReview").mockImplementation(
|
|
async (_planContent, _title, _options, _dialogOptions, extra?: { slider?: HookSelectorSlider }) => {
|
|
const slider = extra?.slider;
|
|
expect(slider).toBeDefined();
|
|
observedSegments = slider!.segments.map(segment => segment.label);
|
|
const slowIndex = slider!.segments.findIndex(segment => segment.label === "slow");
|
|
expect(slowIndex).toBeGreaterThanOrEqual(0);
|
|
// Simulate the operator sliding the tier to "slow" before approving.
|
|
slider!.onChange?.(slowIndex);
|
|
return "Approve and keep context";
|
|
},
|
|
);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(observedSegments).toEqual(["default", "slow"]);
|
|
// The load-bearing assertion: the approved plan executes on the operator's
|
|
// selected tier, not the restored default.
|
|
expect(session.model?.id).toBe(slow.id);
|
|
});
|
|
|
|
it("re-enters plan mode on the approved titled artifact after approve-and-execute", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nExecute then edit.");
|
|
|
|
await mode.handlePlanModeCommand();
|
|
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and execute");
|
|
vi.spyOn(mode, "handleClearCommand").mockResolvedValue();
|
|
vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "APPROVED",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(mode.planModeEnabled).toBe(false);
|
|
expect(session.getPlanReferencePath()).toBe(finalPlanFilePath);
|
|
|
|
await mode.handlePlanModeCommand();
|
|
expect(session.getPlanModeState()).toMatchObject({
|
|
enabled: true,
|
|
planFilePath: finalPlanFilePath,
|
|
reentry: true,
|
|
});
|
|
});
|
|
|
|
it("Approve and compact context: ok outcome dispatches plan-approved after compaction", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nCompact and execute.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and compact context");
|
|
const compactSpy = vi.spyOn(mode, "handleCompactCommand").mockResolvedValue("ok");
|
|
const markSentSpy = vi.spyOn(session, "markPlanReferenceSent");
|
|
const promptSpy = vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
// Compaction was run with the rendered planning-specific custom instruction.
|
|
expect(compactSpy).toHaveBeenCalledTimes(1);
|
|
const [compactInstruction] = compactSpy.mock.calls[0]!;
|
|
expect(typeof compactInstruction).toBe("string");
|
|
expect(compactInstruction as string).toContain(finalPlanFilePath);
|
|
|
|
// Plan-approved synthetic prompt was dispatched.
|
|
const planApprovedIdx = promptSpy.mock.calls.findIndex(isPlanApprovedCall);
|
|
expect(planApprovedIdx).toBeGreaterThanOrEqual(0);
|
|
|
|
// markPlanReferenceSent fires on the dispatch path so the executor's first
|
|
// turn doesn't double-inject the plan reference (it was just dispatched
|
|
// inside the synthetic prompt).
|
|
expect(markSentSpy).toHaveBeenCalledTimes(1);
|
|
});
|
|
|
|
it("Approve and compact context: cancelled outcome skips plan-approved dispatch", async () => {
|
|
// Mock `handleCompactCommand` to surface the "cancelled" outcome directly.
|
|
// (Testing the consumer — `#approvePlan`'s outcome handling — at the
|
|
// CompactionOutcome boundary; the underlying executeCompaction → sentinel
|
|
// classification path is producer-layer and not under T3's contract.)
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nCancel mid-compact.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and compact context");
|
|
vi.spyOn(mode, "handleCompactCommand").mockResolvedValue("cancelled");
|
|
const showWarningSpy = vi.spyOn(mode, "showWarning");
|
|
const setPlanRefSpy = vi.spyOn(session, "setPlanReferencePath");
|
|
const markSentSpy = vi.spyOn(session, "markPlanReferenceSent");
|
|
const promptSpy = vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
// Operator was told the dispatch was deferred.
|
|
expect(showWarningSpy).toHaveBeenCalledWith(
|
|
expect.stringContaining("Plan approved, but compaction was cancelled"),
|
|
);
|
|
// Plan reference path was recorded so the session knows about the approved
|
|
// plan at its final destination …
|
|
expect(setPlanRefSpy).toHaveBeenCalledWith(finalPlanFilePath);
|
|
// … but markPlanReferenceSent was NOT called, so the next operator turn
|
|
// will inject the reference fresh via #buildPlanReferenceMessage. This is
|
|
// the load-bearing assertion that the cancel path leaves the executor
|
|
// with the plan in its first turn.
|
|
expect(markSentSpy).not.toHaveBeenCalled();
|
|
// And — the contract — the plan-approved synthetic prompt was NOT dispatched.
|
|
expect(promptSpy.mock.calls.some(isPlanApprovedCall)).toBe(false);
|
|
});
|
|
|
|
it("Approve and compact context: failed outcome still dispatches plan-approved (best-effort)", async () => {
|
|
// Mock `handleCompactCommand` to surface the "failed" outcome directly.
|
|
// Failure → approval intent stands → synthetic dispatch fires.
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nFail mid-compact.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and compact context");
|
|
vi.spyOn(mode, "handleCompactCommand").mockResolvedValue("failed");
|
|
const markSentSpy = vi.spyOn(session, "markPlanReferenceSent");
|
|
const promptSpy = vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
// Plan-approved synthetic prompt WAS dispatched despite the failure.
|
|
expect(promptSpy.mock.calls.some(isPlanApprovedCall)).toBe(true);
|
|
// markPlanReferenceSent fires on this dispatch path.
|
|
expect(markSentSpy).toHaveBeenCalledTimes(1);
|
|
});
|
|
it("Approve and compact context: setPlanReferencePath is pinned BEFORE compaction flushes the queue", async () => {
|
|
// Regression: handleCompactCommand internally awaits flushCompactionQueue,
|
|
// which can deliver a user-queued message back to the session. If
|
|
// setPlanReferencePath had not been called yet, that queued turn would
|
|
// hit #buildPlanReferenceMessage with the stale plan-mode path. Pin it
|
|
// before the compaction await.
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nQueue race.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and compact context");
|
|
vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
const setPlanRefSpy = vi.spyOn(session, "setPlanReferencePath");
|
|
let planRefSetWhenCompactionRan = false;
|
|
vi.spyOn(mode, "handleCompactCommand").mockImplementation(async () => {
|
|
planRefSetWhenCompactionRan = setPlanRefSpy.mock.calls.some(call => call[0] === finalPlanFilePath);
|
|
return "ok";
|
|
});
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
// The contract: by the time handleCompactCommand runs (and flushes the
|
|
// compaction queue inside), setPlanReferencePath has already pinned the
|
|
// approved plan path, so any user message queued during compaction is
|
|
// dispatched against the approved plan, not the plan-mode draft.
|
|
expect(planRefSetWhenCompactionRan).toBe(true);
|
|
});
|
|
|
|
// ==========================================================================
|
|
// Phase 6 — B layer: #approvePlan flag lifecycle via try/finally.
|
|
//
|
|
// Drives `handlePlanApproval` with each CompactionOutcome variant and
|
|
// asserts `session.isPlanCompactAbortPending === false` after `#approvePlan`
|
|
// resolves/rejects. The flag is the only state that can leak into later
|
|
// unrelated aborts; the `try/finally` in `#approvePlan` is what protects it.
|
|
// ==========================================================================
|
|
|
|
/**
|
|
* Drives `handlePlanApproval` with the "Approve and compact context"
|
|
* picker outcome and the given compaction-outcome mock. Returns the promise
|
|
* the harness produces so the caller decides between `await` (B1-B3 happy
|
|
* paths) and `expect(...).rejects` (B4 throw path). Does NOT swallow errors.
|
|
*/
|
|
async function approveWithCompact(
|
|
compactOutcome: "ok" | "cancelled" | "failed" | "throw",
|
|
throwError?: Error,
|
|
): Promise<void> {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nBody.");
|
|
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and compact context");
|
|
if (compactOutcome === "throw") {
|
|
vi.spyOn(mode, "handleCompactCommand").mockRejectedValue(throwError ?? new Error("compact boom"));
|
|
} else {
|
|
vi.spyOn(mode, "handleCompactCommand").mockResolvedValue(compactOutcome);
|
|
}
|
|
vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
}
|
|
|
|
it("B1: Approve and compact context + ok outcome → flag cleared by finally", async () => {
|
|
await approveWithCompact("ok");
|
|
expect(session.isPlanCompactAbortPending).toBe(false);
|
|
});
|
|
|
|
it("B2: Approve and compact context + cancelled outcome → flag cleared by finally even without aborted message_end", async () => {
|
|
await approveWithCompact("cancelled");
|
|
// The leak-guard contract: no aborted message_end consumed the flag,
|
|
// but `finally` still cleared it so the next real abort cannot be
|
|
// silenced.
|
|
expect(session.isPlanCompactAbortPending).toBe(false);
|
|
});
|
|
|
|
it("B3: Approve and compact context + failed outcome → flag cleared by finally", async () => {
|
|
await approveWithCompact("failed");
|
|
expect(session.isPlanCompactAbortPending).toBe(false);
|
|
});
|
|
|
|
it("B4: Approve and compact context + handleCompactCommand throws → showError surfaces the failure AND flag cleared by finally before the outer catch", async () => {
|
|
// `handlePlanApproval` wraps `#approvePlan` in a try/catch
|
|
// in `InteractiveMode` that consumes the throw and reports via
|
|
// `showError`. The contract under test is:
|
|
// 1. `#approvePlan`'s own `try/finally` clears the flag BEFORE the
|
|
// throw bubbles up to that outer catch.
|
|
// 2. The outer catch surfaces the failure via `showError` (not
|
|
// silenced).
|
|
const showErrorSpy = vi.spyOn(mode, "showError");
|
|
await approveWithCompact("throw", new Error("synthetic compaction failure"));
|
|
expect(session.isPlanCompactAbortPending).toBe(false);
|
|
expect(showErrorSpy).toHaveBeenCalledWith(expect.stringContaining("synthetic compaction failure"));
|
|
});
|
|
|
|
it("B5: Approve and execute (no compact) → markPlanCompactAbortPending never called; flag stays false", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nBody.");
|
|
mode.planModeEnabled = true;
|
|
mode.planModePlanFilePath = planFilePath;
|
|
vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and execute");
|
|
const markSpy = vi.spyOn(session, "markPlanCompactAbortPending");
|
|
vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "PLAN",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(markSpy).not.toHaveBeenCalled();
|
|
expect(session.isPlanCompactAbortPending).toBe(false);
|
|
});
|
|
|
|
it("re-enters plan mode on the approved titled artifact after approval", async () => {
|
|
const planFilePath = "local://PLAN.md";
|
|
const finalPlanFilePath = "local://APPROVED.md";
|
|
const resolvedPlanPath = resolveLocalUrlToPath(planFilePath, {
|
|
getArtifactsDir: () => session.sessionManager.getArtifactsDir(),
|
|
getSessionId: () => session.sessionManager.getSessionId(),
|
|
});
|
|
await Bun.write(resolvedPlanPath, "# Plan\n\nKeep editing this artifact.");
|
|
|
|
await mode.handlePlanModeCommand();
|
|
expect(session.getPlanModeState()?.planFilePath).toBe(planFilePath);
|
|
|
|
vi.spyOn(session, "getContextUsage").mockReturnValue({ tokens: null, contextWindow: 200000, percent: null });
|
|
const selector = vi.spyOn(mode, "showPlanReview").mockResolvedValue("Approve and keep context");
|
|
const showError = vi.spyOn(mode, "showError");
|
|
vi.spyOn(session, "prompt").mockResolvedValue(undefined as never);
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath,
|
|
planExists: true,
|
|
title: "APPROVED",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(mode.planModeEnabled).toBe(false);
|
|
expect(session.getPlanReferencePath()).toBe(finalPlanFilePath);
|
|
|
|
await mode.handlePlanModeCommand();
|
|
expect(session.getPlanModeState()).toMatchObject({
|
|
enabled: true,
|
|
planFilePath: finalPlanFilePath,
|
|
reentry: true,
|
|
});
|
|
|
|
await mode.handlePlanApproval({
|
|
planFilePath: finalPlanFilePath,
|
|
planExists: true,
|
|
title: "APPROVED",
|
|
finalPlanFilePath,
|
|
});
|
|
|
|
expect(selector).toHaveBeenCalledTimes(2);
|
|
expect(showError).not.toHaveBeenCalled();
|
|
});
|
|
|
|
// ==========================================================================
|
|
// Phase 6 — D layer: replay-side render branches in AssistantMessageComponent.
|
|
//
|
|
// D1 asserts that the persisted `SILENT_ABORT_MARKER` suppresses the red
|
|
// "Operation aborted" line. D2 is the over-suppression regression guard —
|
|
// an aborted message with NO marker must still render the line.
|
|
// ==========================================================================
|
|
|
|
function renderAssistant(message: AssistantMessage, width = 120): string {
|
|
const component = new AssistantMessageComponent(message);
|
|
return Bun.stripANSI(component.render(width).join("\n"));
|
|
}
|
|
|
|
/** Build an aborted assistant message with the minimum required fields. */
|
|
function buildAbortedAssistantMessage(overrides: Partial<AssistantMessage> = {}): AssistantMessage {
|
|
return {
|
|
role: "assistant",
|
|
content: [{ type: "text", text: "Approved plan; transitioning to compaction." }],
|
|
api: "openai-completions",
|
|
provider: "github-copilot",
|
|
model: "gpt-4o",
|
|
usage: {
|
|
input: 0,
|
|
output: 0,
|
|
cacheRead: 0,
|
|
cacheWrite: 0,
|
|
totalTokens: 0,
|
|
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },
|
|
},
|
|
stopReason: "aborted",
|
|
timestamp: Date.now(),
|
|
...overrides,
|
|
};
|
|
}
|
|
|
|
it("D1: Replay of an assistant message with SILENT_ABORT_MARKER + aborted: rendered component contains no /Operation aborted/", () => {
|
|
const message = buildAbortedAssistantMessage({ errorMessage: SILENT_ABORT_MARKER });
|
|
const rendered = renderAssistant(message);
|
|
expect(rendered).not.toMatch(/Operation aborted/);
|
|
// The marker itself MUST NOT leak into rendered output either.
|
|
expect(rendered).not.toContain(SILENT_ABORT_MARKER);
|
|
});
|
|
|
|
it("D2: Replay of an aborted message with no marker + empty content: rendered component DOES contain 'Operation aborted'", () => {
|
|
// Over-suppression regression guard: silent path is opt-in via the
|
|
// persisted marker. A user-cancel abort with no marker and no content
|
|
// still surfaces the standard label.
|
|
const message = buildAbortedAssistantMessage({ content: [], errorMessage: undefined });
|
|
const rendered = renderAssistant(message);
|
|
expect(rendered).toContain("Operation aborted");
|
|
});
|
|
});
|