Fix checkout diff stalls on untracked binary artifacts

packages/server/src/server/daemon-e2e/checkout-diff-debug.ts

@@ -1,186 +0,0 @@
-#!/usr/bin/env npx tsx
-/**
- * Ad-hoc checkout diff debugger.
- *
- * Usage:
- *   npx tsx packages/server/src/server/daemon-e2e/checkout-diff-debug.ts
- *   npx tsx packages/server/src/server/daemon-e2e/checkout-diff-debug.ts --agent <agentId>
- *   npx tsx packages/server/src/server/daemon-e2e/checkout-diff-debug.ts --cwd <path>
- *   npx tsx packages/server/src/server/daemon-e2e/checkout-diff-debug.ts --limit 3
- *
- * Optional env:
- *   PASEO_LISTEN=127.0.0.1:6767
- */
-
-import os from "node:os";
-import { DaemonClient } from "../../client/daemon-client.js";
-
-type CliArgs = {
-  agentId?: string;
-  cwd?: string;
-  limit: number;
-};
-
-function parseArgs(argv: string[]): CliArgs {
-  const args: CliArgs = { limit: 5 };
-  for (let i = 0; i < argv.length; i += 1) {
-    const token = argv[i];
-    if (token === "--agent") {
-      const value = argv[i + 1];
-      if (value) {
-        args.agentId = value;
-        i += 1;
-      }
-      continue;
-    }
-    if (token === "--cwd") {
-      const value = argv[i + 1];
-      if (value) {
-        args.cwd = value;
-        i += 1;
-      }
-      continue;
-    }
-    if (token === "--limit") {
-      const value = Number.parseInt(argv[i + 1] ?? "", 10);
-      if (!Number.isNaN(value) && value > 0) {
-        args.limit = value;
-        i += 1;
-      }
-    }
-  }
-  return args;
-}
-
-function fmtMs(ms: number): string {
-  return `${ms.toLocaleString()}ms`;
-}
-
-async function main() {
-  const args = parseArgs(process.argv.slice(2));
-  const listen = process.env.PASEO_LISTEN ?? "127.0.0.1:6767";
-  const url = `ws://${listen}/ws`;
-  const home = process.env.PASEO_HOME ?? `${os.homedir()}/.paseo`;
-
-  console.log("Checkout Diff Debugger");
-  console.log(`daemon=${url}`);
-  console.log(`PASEO_HOME=${home}`);
-  console.log(
-    `filters agent=${args.agentId ?? "-"} cwd=${args.cwd ?? "-"} limit=${args.limit}`
-  );
-  console.log("");
-
-  const client = new DaemonClient({
-    url,
-    reconnect: { enabled: false },
-  });
-
-  client.on("checkout_status_response", (message) => {
-    if (message.type !== "checkout_status_response") return;
-    const payload = message.payload;
-    console.log(
-      `[raw] checkout_status_response requestId=${payload.requestId} cwd=${payload.cwd} isGit=${payload.isGit}`
-    );
-  });
-
-  client.on("checkout_diff_response", (message) => {
-    if (message.type !== "checkout_diff_response") return;
-    const payload = message.payload;
-    console.log(
-      `[raw] checkout_diff_response requestId=${payload.requestId} cwd=${payload.cwd} files=${payload.files.length} error=${payload.error ? "yes" : "no"}`
-    );
-  });
-
-  client.on("rpc_error", (message) => {
-    if (message.type !== "rpc_error") return;
-    const payload = message.payload;
-    console.log(
-      `[raw] rpc_error requestId=${payload.requestId} requestType=${payload.requestType} code=${payload.code ?? "none"}`
-    );
-  });
-
-  try {
-    await client.connect();
-    const ping = await client.ping({ timeoutMs: 3000 });
-    console.log(`ping=${fmtMs(ping.rttMs)}`);
-
-    const snapshots = await client.fetchAgents({ filter: { labels: { ui: "true" } } });
-    const candidates = snapshots
-      .filter((snapshot) => !args.agentId || snapshot.id === args.agentId)
-      .map((snapshot) => ({
-        id: snapshot.id,
-        title: snapshot.title ?? "(untitled)",
-        cwd: snapshot.cwd,
-        updatedAt: snapshot.updatedAt,
-      }))
-      .filter((item) => !args.cwd || item.cwd === args.cwd)
-      .sort((a, b) => (a.updatedAt < b.updatedAt ? 1 : -1));
-
-    const targets = (args.cwd
-      ? [{ id: "(manual)", title: "(manual)", cwd: args.cwd, updatedAt: new Date().toISOString() }]
-      : candidates
-    )
-      .slice(0, args.limit)
-      .filter((item, index, list) => list.findIndex((v) => v.cwd === item.cwd) === index);
-
-    if (targets.length === 0) {
-      console.log("No matching agents/cwds found.");
-      return;
-    }
-
-    console.log(`Testing ${targets.length} cwd target(s)\n`);
-
-    for (const target of targets) {
-      console.log(`--- ${target.cwd}`);
-      console.log(`agent=${target.id} title=${target.title}`);
-
-      const statusStart = Date.now();
-      let statusPayload: Awaited<ReturnType<typeof client.getCheckoutStatus>>;
-      try {
-        statusPayload = await client.getCheckoutStatus(target.cwd);
-        console.log(
-          `status: ok ${fmtMs(Date.now() - statusStart)} isGit=${statusPayload.isGit} branch=${statusPayload.currentBranch ?? "-"} dirty=${statusPayload.isDirty ?? "-"} baseRef=${statusPayload.baseRef ?? "-"}`
-        );
-        if (statusPayload.error) {
-          console.log(`status.error=${statusPayload.error.message}`);
-        }
-      } catch (error) {
-        console.log(`status: FAIL ${fmtMs(Date.now() - statusStart)} ${String(error)}`);
-        console.log("");
-        continue;
-      }
-
-      if (!statusPayload.isGit) {
-        console.log("diff: skipped (not a git repo)\n");
-        continue;
-      }
-
-      const compareMode = statusPayload.isDirty ? "uncommitted" : "base";
-      const diffStart = Date.now();
-      try {
-        const diff = await client.getCheckoutDiff(target.cwd, {
-          mode: compareMode,
-          baseRef: statusPayload.baseRef ?? undefined,
-        });
-        const diffDuration = Date.now() - diffStart;
-        console.log(
-          `diff: ok ${fmtMs(diffDuration)} mode=${compareMode} files=${diff.files.length} error=${diff.error ? "yes" : "no"}`
-        );
-        if (diff.error) {
-          console.log(`diff.error=${diff.error.message}`);
-        }
-      } catch (error) {
-        console.log(`diff: FAIL ${fmtMs(Date.now() - diffStart)} ${String(error)}`);
-      }
-      console.log("");
-    }
-  } finally {
-    await client.close();
-  }
-}
-
-void main().catch((error) => {
-  console.error(error);
-  process.exitCode = 1;
-});
-

packages/server/src/utils/checkout-git.test.ts

@@ -132,6 +132,36 @@ describe("checkout git utilities", () => {
     );
   });
 
+  it("short-circuits untracked binary files", async () => {
+    const binaryPath = join(repoDir, "blob.bin");
+    writeFileSync(binaryPath, Buffer.from([0x00, 0xff, 0x10, 0x80, 0x00, 0x7f, 0x00]));
+
+    const diff = await getCheckoutDiff(repoDir, {
+      mode: "uncommitted",
+      includeStructured: true,
+    });
+
+    const entry = diff.structured?.find((file) => file.path === "blob.bin");
+    expect(entry).toBeTruthy();
+    expect(entry?.status).toBe("binary");
+    expect(diff.diff).toContain("# blob.bin: binary diff omitted");
+  });
+
+  it("marks untracked oversized files as too_large", async () => {
+    const large = Array.from({ length: 240_000 }, (_, i) => `line ${i}`).join("\n") + "\n";
+    writeFileSync(join(repoDir, "untracked-large.txt"), large);
+
+    const diff = await getCheckoutDiff(repoDir, {
+      mode: "uncommitted",
+      includeStructured: true,
+    });
+
+    const entry = diff.structured?.find((file) => file.path === "untracked-large.txt");
+    expect(entry).toBeTruthy();
+    expect(entry?.status).toBe("too_large");
+    expect(diff.diff).toContain("# untracked-large.txt: diff too large omitted");
+  });
+
   it("handles status/diff/commit in a .paseo worktree", async () => {
     const result = await createWorktree({
       branchName: "main",

packages/server/src/utils/checkout-git.ts

@@ -2,6 +2,7 @@ import { exec, execFile, spawn } from "child_process";
 import { promisify } from "util";
 import { resolve, dirname, basename } from "path";
 import { realpathSync } from "fs";
+import { open as openFile, stat as statFile } from "fs/promises";
 import type { ParsedDiffFile } from "../server/utils/diff-highlighter.js";
 import { parseAndHighlightDiff } from "../server/utils/diff-highlighter.js";
 import { isPaseoOwnedWorktreeCwd } from "./worktree.js";
@@ -20,13 +21,6 @@ async function execGit(command: string, options: { cwd: string; env?: NodeJS.Pro
   return execAsync(command, { ...options, maxBuffer: SMALL_OUTPUT_MAX_BUFFER });
 }
 
-async function execGitFile(
-  args: string[],
-  options: { cwd: string; env?: NodeJS.ProcessEnv }
-): Promise<{ stdout: string; stderr: string }> {
-  return execFileAsync("git", args, { ...options, maxBuffer: SMALL_OUTPUT_MAX_BUFFER });
-}
-
 type LimitedTextResult = {
   text: string;
   truncated: boolean;
@@ -190,10 +184,20 @@ async function tryResolveMergeBase(cwd: string, baseRef: string): Promise<string
 
 type FileStat = { additions: number; deletions: number; isBinary: boolean } | null;
 
-async function tryGetNumstat(cwd: string, args: string[]): Promise<FileStat> {
+async function tryGetNumstat(
+  cwd: string,
+  args: string[]
+): Promise<FileStat> {
   try {
-    const { stdout } = await execGitFile(args, { cwd, env: READ_ONLY_GIT_ENV });
-    const line = stdout.trim().split("\n").map((l) => l.trim()).filter(Boolean)[0] ?? "";
+    const { text } = await spawnLimitedText({
+      cmd: "git",
+      args,
+      cwd,
+      env: READ_ONLY_GIT_ENV,
+      maxBytes: 64 * 1024,
+      acceptExitCodes: [0],
+    });
+    const line = text.trim().split("\n").map((l) => l.trim()).filter(Boolean)[0] ?? "";
     if (!line) return null;
     const [aRaw, dRaw] = line.split(/\s+/);
     if (!aRaw || !dRaw) return null;
@@ -634,6 +638,65 @@ async function getAheadOfOrigin(cwd: string, currentBranch: string): Promise<num
 
 const PER_FILE_DIFF_MAX_BYTES = 1024 * 1024; // 1MB
 const TOTAL_DIFF_MAX_BYTES = 2 * 1024 * 1024; // 2MB
+const UNTRACKED_BINARY_SNIFF_BYTES = 16 * 1024;
+
+async function isLikelyBinaryFile(absolutePath: string): Promise<boolean> {
+  const handle = await openFile(absolutePath, "r");
+  try {
+    const buffer = Buffer.allocUnsafe(UNTRACKED_BINARY_SNIFF_BYTES);
+    const { bytesRead } = await handle.read(buffer, 0, buffer.length, 0);
+    if (bytesRead === 0) {
+      return false;
+    }
+
+    let suspicious = 0;
+    for (let i = 0; i < bytesRead; i += 1) {
+      const byte = buffer[i];
+      if (byte === 0) {
+        return true;
+      }
+      // Treat control bytes as suspicious while allowing common whitespace.
+      if (byte < 7 || (byte > 14 && byte < 32) || byte === 127) {
+        suspicious += 1;
+      }
+    }
+
+    return suspicious / bytesRead > 0.3;
+  } finally {
+    await handle.close();
+  }
+}
+
+async function inspectUntrackedFile(
+  cwd: string,
+  relativePath: string
+): Promise<{ stat: FileStat; truncated: boolean }> {
+  const absolutePath = resolve(cwd, relativePath);
+  const metadata = await statFile(absolutePath);
+
+  if (!metadata.isFile()) {
+    return { stat: null, truncated: false };
+  }
+
+  if (await isLikelyBinaryFile(absolutePath)) {
+    return {
+      stat: { additions: 0, deletions: 0, isBinary: true },
+      truncated: false,
+    };
+  }
+
+  if (metadata.size > PER_FILE_DIFF_MAX_BYTES) {
+    return {
+      stat: { additions: 0, deletions: 0, isBinary: false },
+      truncated: true,
+    };
+  }
+
+  return {
+    stat: { additions: 0, deletions: 0, isBinary: false },
+    truncated: false,
+  };
+}
 
 function buildPlaceholderParsedDiffFile(
   change: CheckoutFileChange,
@@ -655,16 +718,16 @@ async function getPerFileDiffText(
   ref: string,
   change: CheckoutFileChange
 ): Promise<{ text: string; truncated: boolean; stat: FileStat }> {
-  const stat: FileStat =
-    change.isUntracked
-      ? null
-      : await tryGetNumstat(cwd, ["diff", "--numstat", ref, "--", change.path]);
-
-  if (stat?.isBinary) {
-    return { text: "", truncated: false, stat };
-  }
-
   if (change.isUntracked) {
+    try {
+      const inspected = await inspectUntrackedFile(cwd, change.path);
+      if (inspected.stat?.isBinary || inspected.truncated) {
+        return { text: "", truncated: inspected.truncated, stat: inspected.stat };
+      }
+    } catch {
+      // Fall through to git diff path if metadata probing fails.
+    }
+
     const result = await spawnLimitedText({
       cmd: "git",
       args: ["diff", "--no-index", "/dev/null", "--", change.path],
@@ -673,7 +736,16 @@ async function getPerFileDiffText(
       maxBytes: PER_FILE_DIFF_MAX_BYTES,
       acceptExitCodes: [0, 1],
     });
-    return { text: result.text, truncated: result.truncated, stat };
+    return {
+      text: result.text,
+      truncated: result.truncated,
+      stat: { additions: 0, deletions: 0, isBinary: false },
+    };
+  }
+
+  const stat = await tryGetNumstat(cwd, ["diff", "--numstat", ref, "--", change.path]);
+  if (stat?.isBinary) {
+    return { text: "", truncated: false, stat };
   }
 
   const result = await spawnLimitedText({