gsd-pi 2.47.0 → 2.48.0-dev.2e7390c

package/src/resources/extensions/gsd/tests/discuss-queued-milestones.test.ts

@@ -0,0 +1,241 @@
+/**
+ * discuss-queued-milestones.test.ts — Tests for #2307.
+ *
+ * /gsd discuss was previously gated on state.activeMilestone, which prevented
+ * users from discussing queued (pending) milestones during roadmap grooming.
+ *
+ * These tests verify:
+ *   1. deriveState correctly identifies pending milestones (the set the picker
+ *      will show when no active milestone is present)
+ *   2. resolveMilestoneFile correctly resolves context artifacts for pending
+ *      milestones so the picker can report their discussion state
+ *   3. The guided-flow.ts source code no longer hard-exits when no active
+ *      milestone exists but pending milestones are present
+ *   4. The helper functions for queued discuss exist in the source
+ */
+
+import { describe, test, afterEach } from "node:test";
+import assert from "node:assert/strict";
+import { mkdtempSync, mkdirSync, rmSync, writeFileSync, readFileSync } from "node:fs";
+import { join } from "node:path";
+import { tmpdir } from "node:os";
+import { fileURLToPath } from "node:url";
+import { dirname } from "node:path";
+
+import { deriveState } from "../state.ts";
+import { invalidateAllCaches } from "../cache.ts";
+import { resolveMilestoneFile } from "../paths.ts";
+
+// ─── Fixture Helpers ──────────────────────────────────────────────────────────
+
+function createBase(): string {
+  const base = mkdtempSync(join(tmpdir(), "gsd-discuss-queued-"));
+  mkdirSync(join(base, ".gsd", "milestones"), { recursive: true });
+  return base;
+}
+
+function cleanup(base: string): void {
+  rmSync(base, { recursive: true, force: true });
+}
+
+function writeMilestoneDir(base: string, mid: string): void {
+  mkdirSync(join(base, ".gsd", "milestones", mid), { recursive: true });
+}
+
+function writeContext(base: string, mid: string, content: string): void {
+  writeMilestoneDir(base, mid);
+  writeFileSync(join(base, ".gsd", "milestones", mid, `${mid}-CONTEXT.md`), content);
+}
+
+function writeContextDraft(base: string, mid: string, content: string): void {
+  writeMilestoneDir(base, mid);
+  writeFileSync(join(base, ".gsd", "milestones", mid, `${mid}-CONTEXT-DRAFT.md`), content);
+}
+
+function writeRoadmap(base: string, mid: string, content: string): void {
+  writeMilestoneDir(base, mid);
+  writeFileSync(join(base, ".gsd", "milestones", mid, `${mid}-ROADMAP.md`), content);
+}
+
+function readGuidedFlowSource(): string {
+  const thisFile = fileURLToPath(import.meta.url);
+  const thisDir = dirname(thisFile);
+  return readFileSync(join(thisDir, "..", "guided-flow.ts"), "utf-8");
+}
+
+// ─── Tests ────────────────────────────────────────────────────────────────────
+
+describe("discuss-queued-milestones (#2307)", () => {
+
+  test("1. pending milestones appear in registry when active milestone exists", async () => {
+    const base = createBase();
+    try {
+      // M001: active — has context + roadmap with a slice
+      writeContext(base, "M001", "# M001: Active\nContext here.");
+      writeRoadmap(base, "M001",
+        "# M001: Active\n\n## Slices\n- [ ] **S01: Do work** `risk:low` `depends:[]`\n  > After this: works\n");
+
+      // M002: pending — context only, no roadmap
+      writeContext(base, "M002", "# M002: Queued\nFuture work.");
+
+      // M003: pending — draft context only
+      writeContextDraft(base, "M003", "# M003: Draft\nSeed material.");
+
+      invalidateAllCaches();
+      const state = await deriveState(base);
+
+      assert.ok(!!state.activeMilestone, "M001 should be the active milestone");
+      assert.strictEqual(state.activeMilestone?.id, "M001");
+
+      const pendingIds = state.registry
+        .filter(m => m.status === "pending")
+        .map(m => m.id);
+
+      assert.ok(pendingIds.includes("M002"), "M002 should be pending");
+      assert.ok(pendingIds.includes("M003"), "M003 should be pending");
+    } finally {
+      cleanup(base);
+    }
+  });
+
+  test("2. first context-only milestone is active, subsequent ones are pending", async () => {
+    const base = createBase();
+    try {
+      // M001: first milestone with context but no roadmap — deriveState marks it active
+      writeContext(base, "M001", "# M001: First\nContext here.");
+      // M002: will be pending since M001 is active
+      writeContext(base, "M002", "# M002: Second\nMore future work.");
+
+      invalidateAllCaches();
+      const state = await deriveState(base);
+
+      // deriveState makes the first unfinished milestone "active" even without a roadmap
+      assert.ok(!!state.activeMilestone, "first milestone should be active");
+      assert.strictEqual(state.activeMilestone?.id, "M001", "M001 is the active milestone");
+
+      const pendingIds = state.registry
+        .filter(m => m.status === "pending")
+        .map(m => m.id);
+
+      assert.ok(pendingIds.includes("M002"),
+        "M002 should be pending — it comes after the active M001");
+    } finally {
+      cleanup(base);
+    }
+  });
+
+  test("3. resolveMilestoneFile finds CONTEXT.md for pending milestone", (t) => {
+    const base = createBase();
+    try {
+      writeContext(base, "M002", "# M002: Queued\nContent.");
+
+      const contextFile = resolveMilestoneFile(base, "M002", "CONTEXT");
+      assert.ok(contextFile !== null, "resolveMilestoneFile should find CONTEXT.md for M002");
+      assert.ok(contextFile!.endsWith("M002-CONTEXT.md"),
+        "resolved path should point to M002-CONTEXT.md");
+    } finally {
+      cleanup(base);
+    }
+  });
+
+  test("4. resolveMilestoneFile finds CONTEXT-DRAFT.md for pending milestone", (t) => {
+    const base = createBase();
+    try {
+      writeContextDraft(base, "M003", "# M003: Draft\nSeed content.");
+
+      const draftFile = resolveMilestoneFile(base, "M003", "CONTEXT-DRAFT");
+      assert.ok(draftFile !== null, "resolveMilestoneFile should find CONTEXT-DRAFT.md for M003");
+      assert.ok(draftFile!.endsWith("M003-CONTEXT-DRAFT.md"),
+        "resolved path should point to M003-CONTEXT-DRAFT.md");
+    } finally {
+      cleanup(base);
+    }
+  });
+
+  test("5. resolveMilestoneFile returns null when pending milestone has no context", (t) => {
+    const base = createBase();
+    try {
+      writeMilestoneDir(base, "M004");
+
+      const contextFile = resolveMilestoneFile(base, "M004", "CONTEXT");
+      assert.strictEqual(contextFile, null,
+        "resolveMilestoneFile should return null when no CONTEXT.md exists");
+
+      const draftFile = resolveMilestoneFile(base, "M004", "CONTEXT-DRAFT");
+      assert.strictEqual(draftFile, null,
+        "resolveMilestoneFile should return null when no CONTEXT-DRAFT.md exists");
+    } finally {
+      cleanup(base);
+    }
+  });
+
+  test("6. guided-flow no longer hard-exits when no active milestone but pending exist", () => {
+    const source = readGuidedFlowSource();
+
+    // The old guard was a simple early-exit:
+    //   if (!state.activeMilestone) {
+    //     ctx.ui.notify("No active milestone. Run /gsd to create one first.", "warning");
+    //     return;
+    //   }
+    //
+    // The new guard should check for pending milestones and route instead.
+    const oldGuardPattern = /if\s*\(!state\.activeMilestone\)\s*\{\s*ctx\.ui\.notify\("No active milestone/;
+    assert.ok(
+      !oldGuardPattern.test(source),
+      "guided-flow must not unconditionally exit when activeMilestone is null",
+    );
+  });
+
+  test("7. showDiscussQueuedMilestone helper exists in guided-flow", () => {
+    const source = readGuidedFlowSource();
+    assert.ok(
+      source.includes("showDiscussQueuedMilestone"),
+      "guided-flow must export showDiscussQueuedMilestone helper",
+    );
+  });
+
+  test("8. dispatchDiscussForMilestone helper exists in guided-flow", () => {
+    const source = readGuidedFlowSource();
+    assert.ok(
+      source.includes("dispatchDiscussForMilestone"),
+      "guided-flow must export dispatchDiscussForMilestone helper",
+    );
+  });
+
+  test("9. dispatchDiscussForMilestone does not set pendingAutoStart", () => {
+    const source = readGuidedFlowSource();
+
+    // Extract the dispatchDiscussForMilestone function body
+    const fnMatch = source.match(
+      /async function dispatchDiscussForMilestone\s*\([^)]*\)[^{]*\{([\s\S]*?)\n\}/,
+    );
+    assert.ok(!!fnMatch, "dispatchDiscussForMilestone function body must be present");
+
+    if (fnMatch) {
+      assert.ok(
+        !fnMatch[1].includes("pendingAutoStart"),
+        "dispatchDiscussForMilestone must NOT set pendingAutoStart — discussing a queued milestone must not activate it",
+      );
+    }
+  });
+
+  test("10. slice picker includes queued milestone option when pending milestones exist", () => {
+    const source = readGuidedFlowSource();
+    assert.ok(
+      source.includes("discuss_queued_milestone"),
+      "slice picker must include a 'discuss_queued_milestone' action id for queued milestones",
+    );
+    assert.ok(
+      source.includes("Discuss a queued milestone"),
+      "slice picker must label the queued milestone action clearly",
+    );
+  });
+
+  test("11. queued milestone picker labels entries with [queued]", () => {
+    const source = readGuidedFlowSource();
+    assert.ok(
+      source.includes("[queued]"),
+      "queued milestone picker must label entries with [queued] to distinguish from active",
+    );
+  });
+});

package/src/resources/extensions/gsd/tests/forensics-error-filter.test.ts

@@ -0,0 +1,121 @@
+/**
+ * Regression test for #2539: extractTrace should not count benign bash
+ * exit-code-1 (grep no-match) or user skips as errors.
+ */
+import { describe, test } from "node:test";
+import assert from "node:assert/strict";
+
+import { extractTrace } from "../session-forensics.ts";
+
+/**
+ * Build a minimal JSONL entry pair: assistant tool_use → toolResult.
+ * This is the shape extractTrace() expects from session activity files.
+ */
+function makeToolPair(
+  toolName: string,
+  input: Record<string, unknown>,
+  resultText: string,
+  isError: boolean,
+): unknown[] {
+  const toolCallId = `toolu_${Math.random().toString(36).slice(2, 10)}`;
+  return [
+    {
+      type: "message",
+      message: {
+        role: "assistant",
+        content: [
+          {
+            type: "toolCall",
+            id: toolCallId,
+            name: toolName,
+            arguments: input,
+          },
+        ],
+      },
+    },
+    {
+      type: "message",
+      message: {
+        role: "toolResult",
+        toolCallId,
+        toolName,
+        isError,
+        content: [{ type: "text", text: resultText }],
+      },
+    },
+  ];
+}
+
+describe("extractTrace error filtering (#2539)", () => {
+  test("grep exit-code-1 (no matches) is not counted as an error", () => {
+    const entries = makeToolPair(
+      "bash",
+      { command: "grep -rn 'nonexistent' src/" },
+      "(no output)\nCommand exited with code 1",
+      true,
+    );
+    const trace = extractTrace(entries);
+    assert.equal(trace.errors.length, 0, "grep no-match should not be an error");
+  });
+
+  test("user skip is not counted as an error", () => {
+    const entries = makeToolPair(
+      "bash",
+      { command: "npm run test" },
+      "Skipped due to queued user message",
+      true,
+    );
+    const trace = extractTrace(entries);
+    assert.equal(trace.errors.length, 0, "user skip should not be an error");
+  });
+
+  test("real bash error is still counted", () => {
+    const entries = makeToolPair(
+      "bash",
+      { command: "cat /nonexistent" },
+      "cat: /nonexistent: No such file or directory\nCommand exited with code 1",
+      true,
+    );
+    const trace = extractTrace(entries);
+    assert.equal(trace.errors.length, 1, "real error should still be counted");
+    assert.match(trace.errors[0], /No such file or directory/);
+  });
+
+  test("non-bash tool error is still counted", () => {
+    const entries = makeToolPair(
+      "edit",
+      { path: "foo.ts", oldText: "x", newText: "y" },
+      "oldText not found in file",
+      true,
+    );
+    const trace = extractTrace(entries);
+    assert.equal(trace.errors.length, 1, "non-bash tool errors should still be counted");
+  });
+
+  test("mixed entries: only real errors are counted", () => {
+    const entries = [
+      // benign grep no-match
+      ...makeToolPair("bash", { command: "grep -rn 'pattern' src/" }, "(no output)\nCommand exited with code 1", true),
+      // user skip
+      ...makeToolPair("bash", { command: "npm test" }, "Skipped due to queued user message", true),
+      // real error
+      ...makeToolPair("bash", { command: "node broken.js" }, "SyntaxError: Unexpected token\nCommand exited with code 1", true),
+      // successful command (not an error)
+      ...makeToolPair("bash", { command: "echo hello" }, "hello", false),
+    ];
+    const trace = extractTrace(entries);
+    assert.equal(trace.errors.length, 1, "only the real error should be counted");
+    assert.match(trace.errors[0], /SyntaxError/);
+  });
+
+  test("exit code 1 with actual output is still an error", () => {
+    const entries = makeToolPair(
+      "bash",
+      { command: "npm run lint" },
+      "src/foo.ts:10:5 - error TS2304: Cannot find name 'x'\nCommand exited with code 1",
+      true,
+    );
+    const trace = extractTrace(entries);
+    assert.equal(trace.errors.length, 1, "lint error with output should be counted");
+  });
+});

package/src/resources/extensions/gsd/tests/forensics-journal.test.ts

@@ -0,0 +1,162 @@
+import { describe, it } from "node:test";
+import assert from "node:assert/strict";
+import { readFileSync } from "node:fs";
+import { join, dirname } from "node:path";
+import { fileURLToPath } from "node:url";
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const gsdDir = join(__dirname, "..");
+
+describe("forensics journal & activity log awareness", () => {
+  const forensicsSrc = readFileSync(join(gsdDir, "forensics.ts"), "utf-8");
+  const promptSrc = readFileSync(join(gsdDir, "prompts", "forensics.md"), "utf-8");
+
+  it("scanJournalForForensics reads journal files directly (no full queryJournal load)", () => {
+    // Must NOT use queryJournal which loads ALL entries into memory
+    assert.ok(
+      !forensicsSrc.includes('queryJournal('),
+      "forensics.ts must NOT call queryJournal() which loads all entries at once",
+    );
+    // Must have its own journal scanning with file-level limits
+    assert.ok(
+      forensicsSrc.includes("scanJournalForForensics"),
+      "forensics.ts must have scanJournalForForensics function",
+    );
+  });
+
+  it("journal scanning limits files parsed to avoid memory bloat", () => {
+    assert.ok(
+      forensicsSrc.includes("MAX_JOURNAL_RECENT_FILES"),
+      "must have MAX_JOURNAL_RECENT_FILES constant to limit parsed files",
+    );
+    assert.ok(
+      forensicsSrc.includes("MAX_JOURNAL_RECENT_EVENTS"),
+      "must have MAX_JOURNAL_RECENT_EVENTS constant to limit events extracted",
+    );
+  });
+
+  it("older journal files are line-counted without full JSON parse", () => {
+    assert.ok(
+      forensicsSrc.includes("olderEntryCount") || forensicsSrc.includes("olderFiles"),
+      "must handle older files separately from recent files",
+    );
+  });
+
+  it("ForensicReport includes journalSummary field", () => {
+    assert.ok(
+      forensicsSrc.includes("journalSummary"),
+      "ForensicReport must include journalSummary field",
+    );
+  });
+
+  it("ForensicReport includes activityLogMeta field", () => {
+    assert.ok(
+      forensicsSrc.includes("activityLogMeta"),
+      "ForensicReport must include activityLogMeta field",
+    );
+  });
+
+  it("buildForensicReport calls scanJournalForForensics", () => {
+    assert.ok(
+      forensicsSrc.includes("scanJournalForForensics"),
+      "buildForensicReport must call scanJournalForForensics",
+    );
+  });
+
+  it("buildForensicReport calls gatherActivityLogMeta", () => {
+    assert.ok(
+      forensicsSrc.includes("gatherActivityLogMeta"),
+      "buildForensicReport must call gatherActivityLogMeta",
+    );
+  });
+
+  it("forensics detects journal-based anomalies", () => {
+    assert.ok(
+      forensicsSrc.includes("detectJournalAnomalies"),
+      "forensics.ts must have detectJournalAnomalies function",
+    );
+    // Check for specific journal anomaly types
+    assert.ok(forensicsSrc.includes('"journal-stuck"'), "must detect journal-stuck anomalies");
+    assert.ok(forensicsSrc.includes('"journal-guard-block"'), "must detect journal-guard-block anomalies");
+    assert.ok(forensicsSrc.includes('"journal-rapid-iterations"'), "must detect journal-rapid-iterations anomalies");
+    assert.ok(forensicsSrc.includes('"journal-worktree-failure"'), "must detect journal-worktree-failure anomalies");
+  });
+
+  it("formatReportForPrompt includes journal summary section", () => {
+    assert.ok(
+      forensicsSrc.includes("Journal Summary"),
+      "prompt formatter must include a Journal Summary section",
+    );
+  });
+
+  it("formatReportForPrompt includes activity log overview section", () => {
+    assert.ok(
+      forensicsSrc.includes("Activity Log Overview"),
+      "prompt formatter must include an Activity Log Overview section",
+    );
+  });
+
+  it("activity log scanning uses tail-read with byte cap (not full file load)", () => {
+    // scanActivityLogs uses nativeParseJsonlTail + MAX_JSONL_BYTES for efficient reading
+    assert.ok(
+      forensicsSrc.includes("nativeParseJsonlTail"),
+      "activity log scanning must use nativeParseJsonlTail for tail-reading",
+    );
+    assert.ok(
+      forensicsSrc.includes("MAX_JSONL_BYTES"),
+      "activity log scanning must respect MAX_JSONL_BYTES cap",
+    );
+    // Only reads last 5 files
+    assert.ok(
+      forensicsSrc.includes("slice(-5)"),
+      "activity log scanning must limit to last 5 files",
+    );
+  });
+
+  it("activity log entries are distilled through extractTrace, not sent raw", () => {
+    assert.ok(
+      forensicsSrc.includes("extractTrace("),
+      "activity log entries must be distilled through extractTrace before reporting",
+    );
+  });
+
+  it("prompt output is hard-capped at 30KB", () => {
+    assert.ok(
+      forensicsSrc.includes("MAX_BYTES") && forensicsSrc.includes("30 * 1024"),
+      "formatReportForPrompt must have a 30KB hard cap",
+    );
+    assert.ok(
+      forensicsSrc.includes("truncated at 30KB"),
+      "prompt must show truncation message when capped",
+    );
+  });
+
+  it("forensics prompt documents journal format", () => {
+    assert.ok(
+      promptSrc.includes("### Journal Format"),
+      "forensics.md must document the journal format",
+    );
+    assert.ok(
+      promptSrc.includes("flowId"),
+      "forensics.md must reference flowId concept",
+    );
+    assert.ok(
+      promptSrc.includes("causedBy"),
+      "forensics.md must reference causedBy for causal chains",
+    );
+  });
+
+  it("forensics prompt includes journal directory in runtime path reference", () => {
+    assert.ok(
+      promptSrc.includes("journal/"),
+      "forensics.md runtime path reference must include journal/",
+    );
+  });
+
+  it("investigation protocol references journal data", () => {
+    assert.ok(
+      promptSrc.includes("journal timeline") || promptSrc.includes("journal events"),
+      "investigation protocol must reference journal data for tracing",
+    );
+  });
+});

package/src/resources/extensions/gsd/tests/preflight-context-draft-filter.test.ts

@@ -0,0 +1,115 @@
+/**
+ * Regression test for #2473: Pre-flight CONTEXT-DRAFT warning should skip
+ * completed and parked milestones.
+ *
+ * The pre-flight loop in auto-start.ts warns about CONTEXT-DRAFT.md files
+ * so the user knows which milestones will pause for discussion. But completed
+ * milestones with leftover CONTEXT-DRAFT.md files are not actionable — the
+ * warning is noise.
+ *
+ * This test exercises the filtering logic directly: given a set of milestones
+ * with CONTEXT-DRAFT files, only active/pending ones should produce warnings.
+ */
+import { describe, test, beforeEach, afterEach } from "node:test";
+import assert from "node:assert/strict";
+import { mkdtempSync, mkdirSync, rmSync, writeFileSync } from "node:fs";
+import { join } from "node:path";
+import { tmpdir } from "node:os";
+
+import {
+  openDatabase,
+  closeDatabase,
+  isDbAvailable,
+  insertMilestone,
+  getMilestone,
+} from "../gsd-db.ts";
+import { resolveMilestoneFile } from "../paths.ts";
+
+describe("pre-flight CONTEXT-DRAFT filter (#2473)", () => {
+  let tmpBase: string;
+  let gsd: string;
+
+  beforeEach(() => {
+    tmpBase = mkdtempSync(join(tmpdir(), "gsd-preflight-draft-"));
+    gsd = join(tmpBase, ".gsd");
+
+    // Create milestone directories with CONTEXT-DRAFT files
+    for (const id of ["M001", "M002", "M003"]) {
+      const msDir = join(gsd, "milestones", id);
+      mkdirSync(msDir, { recursive: true });
+      writeFileSync(join(msDir, `${id}-CONTEXT-DRAFT.md`), `# ${id}: Draft\n`);
+    }
+
+    // Open DB and insert milestones with different statuses
+    const dbPath = join(gsd, "gsd.db");
+    openDatabase(dbPath);
+    insertMilestone({ id: "M001", title: "Complete milestone", status: "complete" });
+    insertMilestone({ id: "M002", title: "Active milestone", status: "active" });
+    insertMilestone({ id: "M003", title: "Parked milestone", status: "parked" });
+  });
+
+  afterEach(() => {
+    closeDatabase();
+    rmSync(tmpBase, { recursive: true, force: true });
+  });
+
+  test("completed milestone is skipped — no warning emitted", () => {
+    assert.ok(isDbAvailable(), "DB should be available");
+    const ms = getMilestone("M001");
+    assert.equal(ms?.status, "complete");
+  });
+
+  test("parked milestone is skipped — no warning emitted", () => {
+    const ms = getMilestone("M003");
+    assert.equal(ms?.status, "parked");
+  });
+
+  test("active milestone with CONTEXT-DRAFT produces warning", () => {
+    const ms = getMilestone("M002");
+    assert.equal(ms?.status, "active");
+
+    const draft = resolveMilestoneFile(tmpBase, "M002", "CONTEXT-DRAFT");
+    assert.ok(draft, "CONTEXT-DRAFT file should be found for active milestone");
+  });
+
+  test("full pre-flight filter produces warnings only for active milestones", () => {
+    const milestoneIds = ["M001", "M002", "M003"];
+    const issues: string[] = [];
+
+    for (const id of milestoneIds) {
+      // Replicate the fixed pre-flight logic from auto-start.ts
+      if (isDbAvailable()) {
+        const ms = getMilestone(id);
+        if (ms?.status === "complete" || ms?.status === "parked") continue;
+      }
+      const draft = resolveMilestoneFile(tmpBase, id, "CONTEXT-DRAFT");
+      if (draft) {
+        issues.push(`${id}: has CONTEXT-DRAFT.md (will pause for discussion)`);
+      }
+    }
+
+    assert.equal(issues.length, 1, "only one warning should be emitted");
+    assert.match(issues[0], /M002/, "warning should be for the active milestone only");
+  });
+
+  test("when DB is unavailable, all milestones with CONTEXT-DRAFT produce warnings (safe fallback)", () => {
+    closeDatabase();
+    assert.ok(!isDbAvailable(), "DB should be unavailable after close");
+
+    const milestoneIds = ["M001", "M002", "M003"];
+    const issues: string[] = [];
+
+    for (const id of milestoneIds) {
+      if (isDbAvailable()) {
+        const ms = getMilestone(id);
+        if (ms?.status === "complete" || ms?.status === "parked") continue;
+      }
+      const draft = resolveMilestoneFile(tmpBase, id, "CONTEXT-DRAFT");
+      if (draft) {
+        issues.push(`${id}: has CONTEXT-DRAFT.md (will pause for discussion)`);
+      }
+    }
+
+    assert.equal(issues.length, 3, "all milestones should warn when DB is unavailable");
+  });
+});