fix: skip binary workspace diff contents

src/workspace/get-workspace-changes.ts

@@ -38,6 +38,7 @@ interface ChangesFromRefInput {
 interface DiffStat {
 	additions: number;
 	deletions: number;
+	isBinary: boolean;
 }
 
 interface FileFingerprint {
@@ -192,40 +193,53 @@ async function readWorkingTreeFile(repoRoot: string, path: string): Promise<stri
 
 function fallbackStats(oldText: string | null, newText: string | null): DiffStat {
 	if (oldText == null && newText == null) {
-		return { additions: 0, deletions: 0 };
+		return { additions: 0, deletions: 0, isBinary: false };
 	}
 	if (oldText == null) {
-		return { additions: toLineCount(newText ?? ""), deletions: 0 };
+		return { additions: toLineCount(newText ?? ""), deletions: 0, isBinary: false };
 	}
 	if (newText == null) {
-		return { additions: 0, deletions: toLineCount(oldText) };
+		return { additions: 0, deletions: toLineCount(oldText), isBinary: false };
 	}
 
 	const oldLines = toLineCount(oldText);
 	const newLines = toLineCount(newText);
 	return {
 		additions: Math.max(newLines - oldLines, 0),
 		deletions: Math.max(oldLines - newLines, 0),
+		isBinary: false,
+	};
+}
+
+function parseDiffStatOutput(output: string): DiffStat | null {
+	const firstLine = output
+		.split("\n")
+		.map((line) => line.trim())
+		.find(Boolean);
+	if (!firstLine) {
+		return null;
+	}
+	const [addedRaw, deletedRaw] = firstLine.split("\t");
+	if (addedRaw === "-" && deletedRaw === "-") {
+		return {
+			additions: 0,
+			deletions: 0,
+			isBinary: true,
+		};
+	}
+	const additions = Number.parseInt(addedRaw ?? "", 10);
+	const deletions = Number.parseInt(deletedRaw ?? "", 10);
+	return {
+		additions: Number.isFinite(additions) ? additions : 0,
+		deletions: Number.isFinite(deletions) ? deletions : 0,
+		isBinary: false,
 	};
 }
 
 async function readDiffStat(repoRoot: string, path: string): Promise<DiffStat | null> {
 	try {
 		const output = await getGitStdout(["diff", "--numstat", "HEAD", "--", path], repoRoot);
-		const firstLine = output
-			.split("\n")
-			.map((line) => line.trim())
-			.find(Boolean);
-		if (!firstLine) {
-			return null;
-		}
-		const [addedRaw, deletedRaw] = firstLine.split("\t");
-		const additions = Number.parseInt(addedRaw ?? "", 10);
-		const deletions = Number.parseInt(deletedRaw ?? "", 10);
-		return {
-			additions: Number.isFinite(additions) ? additions : 0,
-			deletions: Number.isFinite(deletions) ? deletions : 0,
-		};
+		return parseDiffStatOutput(output);
 	} catch {
 		return null;
 	}
@@ -239,20 +253,7 @@ async function readDiffStatBetweenRefs(
 ): Promise<DiffStat | null> {
 	try {
 		const output = await getGitStdout(["diff", "--numstat", fromRef, toRef, "--", path], repoRoot);
-		const firstLine = output
-			.split("\n")
-			.map((line) => line.trim())
-			.find(Boolean);
-		if (!firstLine) {
-			return null;
-		}
-		const [addedRaw, deletedRaw] = firstLine.split("\t");
-		const additions = Number.parseInt(addedRaw ?? "", 10);
-		const deletions = Number.parseInt(deletedRaw ?? "", 10);
-		return {
-			additions: Number.isFinite(additions) ? additions : 0,
-			deletions: Number.isFinite(deletions) ? deletions : 0,
-		};
+		return parseDiffStatOutput(output);
 	} catch {
 		return null;
 	}
@@ -261,34 +262,25 @@ async function readDiffStatBetweenRefs(
 async function readDiffStatFromRef(repoRoot: string, fromRef: string, path: string): Promise<DiffStat | null> {
 	try {
 		const output = await getGitStdout(["diff", "--numstat", fromRef, "--", path], repoRoot);
-		const firstLine = output
-			.split("\n")
-			.map((line) => line.trim())
-			.find(Boolean);
-		if (!firstLine) {
-			return null;
-		}
-		const [addedRaw, deletedRaw] = firstLine.split("\t");
-		const additions = Number.parseInt(addedRaw ?? "", 10);
-		const deletions = Number.parseInt(deletedRaw ?? "", 10);
-		return {
-			additions: Number.isFinite(additions) ? additions : 0,
-			deletions: Number.isFinite(deletions) ? deletions : 0,
-		};
+		return parseDiffStatOutput(output);
 	} catch {
 		return null;
 	}
 }
 
 async function buildFileChange(repoRoot: string, entry: NameStatusEntry): Promise<RuntimeWorkspaceFileChange> {
 	const basePath = entry.previousPath ?? entry.path;
+	const diffStat = entry.status === "untracked" ? null : await readDiffStat(repoRoot, entry.path);
 	const oldText =
-		entry.status === "added" || entry.status === "untracked" ? null : await readHeadFile(repoRoot, basePath);
-	const newText = entry.status === "deleted" ? null : await readWorkingTreeFile(repoRoot, entry.path);
+		entry.status === "added" || entry.status === "untracked" || diffStat?.isBinary
+			? null
+			: await readHeadFile(repoRoot, basePath);
+	const newText =
+		entry.status === "deleted" || diffStat?.isBinary ? null : await readWorkingTreeFile(repoRoot, entry.path);
 	const stats =
 		entry.status === "untracked"
 			? { additions: toLineCount(newText ?? ""), deletions: 0 }
-			: ((await readDiffStat(repoRoot, entry.path)) ?? fallbackStats(oldText, newText));
+			: (diffStat ?? fallbackStats(oldText, newText));
 
 	return {
 		path: entry.path,
@@ -308,10 +300,12 @@ async function buildFileChangeBetweenRefs(
 	toRef: string,
 ): Promise<RuntimeWorkspaceFileChange> {
 	const basePath = entry.previousPath ?? entry.path;
-	const oldText = entry.status === "added" ? null : await readFileAtRef(repoRoot, fromRef, basePath);
-	const newText = entry.status === "deleted" ? null : await readFileAtRef(repoRoot, toRef, entry.path);
-	const stats =
-		(await readDiffStatBetweenRefs(repoRoot, fromRef, toRef, entry.path)) ?? fallbackStats(oldText, newText);
+	const diffStat = await readDiffStatBetweenRefs(repoRoot, fromRef, toRef, entry.path);
+	const oldText =
+		entry.status === "added" || diffStat?.isBinary ? null : await readFileAtRef(repoRoot, fromRef, basePath);
+	const newText =
+		entry.status === "deleted" || diffStat?.isBinary ? null : await readFileAtRef(repoRoot, toRef, entry.path);
+	const stats = diffStat ?? fallbackStats(oldText, newText);
 
 	return {
 		path: entry.path,
@@ -330,15 +324,17 @@ async function buildFileChangeFromRef(
 	fromRef: string,
 ): Promise<RuntimeWorkspaceFileChange> {
 	const basePath = entry.previousPath ?? entry.path;
+	const diffStat = entry.status === "untracked" ? null : await readDiffStatFromRef(repoRoot, fromRef, entry.path);
 	const oldText =
-		entry.status === "added" || entry.status === "untracked"
+		entry.status === "added" || entry.status === "untracked" || diffStat?.isBinary
 			? null
 			: await readFileAtRef(repoRoot, fromRef, basePath);
-	const newText = entry.status === "deleted" ? null : await readWorkingTreeFile(repoRoot, entry.path);
+	const newText =
+		entry.status === "deleted" || diffStat?.isBinary ? null : await readWorkingTreeFile(repoRoot, entry.path);
 	const stats =
 		entry.status === "untracked"
 			? { additions: toLineCount(newText ?? ""), deletions: 0 }
-			: ((await readDiffStatFromRef(repoRoot, fromRef, entry.path)) ?? fallbackStats(oldText, newText));
+			: (diffStat ?? fallbackStats(oldText, newText));
 
 	return {
 		path: entry.path,

test/runtime/get-workspace-changes.test.ts

@@ -0,0 +1,92 @@
+import { execFile } from "node:child_process";
+import { mkdtemp, open, rm, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { promisify } from "node:util";
+
+import { afterEach, beforeEach, describe, expect, it } from "vitest";
+
+import { getWorkspaceChanges, getWorkspaceChangesBetweenRefs } from "../../src/workspace/get-workspace-changes";
+
+const execFileAsync = promisify(execFile);
+const FIFTY_MIB = 50 * 1024 * 1024;
+
+async function runGit(repoRoot: string, args: string[]): Promise<string> {
+	const { stdout } = await execFileAsync("git", ["-c", "core.quotepath=false", ...args], {
+		cwd: repoRoot,
+		encoding: "utf8",
+		env: {
+			...process.env,
+			GIT_CONFIG_GLOBAL: "/dev/null",
+			GIT_CONFIG_NOSYSTEM: "1",
+		},
+	});
+	return stdout.trim();
+}
+
+async function commitAll(repoRoot: string, message: string): Promise<string> {
+	await runGit(repoRoot, ["add", "."]);
+	await runGit(repoRoot, ["commit", "-m", message]);
+	return await runGit(repoRoot, ["rev-parse", "HEAD"]);
+}
+
+async function createSparseFile(path: string, size: number): Promise<void> {
+	const handle = await open(path, "w");
+	try {
+		await handle.truncate(size);
+	} finally {
+		await handle.close();
+	}
+}
+
+describe("getWorkspaceChanges", () => {
+	let repoRoot: string;
+
+	beforeEach(async () => {
+		repoRoot = await mkdtemp(join(tmpdir(), "kanban-workspace-changes-"));
+		await runGit(repoRoot, ["init"]);
+		await runGit(repoRoot, ["config", "user.email", "test@example.com"]);
+		await runGit(repoRoot, ["config", "user.name", "Test User"]);
+	});
+
+	afterEach(async () => {
+		await rm(repoRoot, { recursive: true, force: true });
+	});
+
+	it("skips a 50 MiB working tree binary without loading file text", async () => {
+		const filePath = join(repoRoot, "large.bin");
+		await writeFile(filePath, "initial\n");
+		await commitAll(repoRoot, "Initial commit");
+
+		await createSparseFile(filePath, FIFTY_MIB);
+
+		const response = await getWorkspaceChanges(repoRoot);
+		const file = response.files.find((entry) => entry.path === "large.bin");
+
+		expect(file).toBeDefined();
+		expect(file?.status).toBe("modified");
+		expect(file?.oldText).toBeNull();
+		expect(file?.newText).toBeNull();
+	});
+
+	it("skips a 50 MiB historical binary before reading it from git", async () => {
+		const filePath = join(repoRoot, "large.bin");
+		await writeFile(filePath, "initial\n");
+		const smallRef = await commitAll(repoRoot, "Initial commit");
+
+		await createSparseFile(filePath, FIFTY_MIB);
+		const largeRef = await commitAll(repoRoot, "Large binary commit");
+
+		const response = await getWorkspaceChangesBetweenRefs({
+			cwd: repoRoot,
+			fromRef: smallRef,
+			toRef: largeRef,
+		});
+		const file = response.files.find((entry) => entry.path === "large.bin");
+
+		expect(file).toBeDefined();
+		expect(file?.status).toBe("modified");
+		expect(file?.oldText).toBeNull();
+		expect(file?.newText).toBeNull();
+	});
+});

web-ui/src/components/detail-panels/diff-viewer-panel.test.tsx

@@ -265,6 +265,7 @@ describe("DiffViewerPanel", () => {
 
 		expect(container.textContent).toContain("assets/logo.png");
 		expect(container.textContent).toContain("Binary");
+		expect(container.textContent).toContain("Binary file not shown.");
 		expect(container.querySelector(".kb-diff-row")).toBeNull();
 	});
 

web-ui/src/components/detail-panels/diff-viewer-panel.tsx

@@ -889,7 +889,17 @@ export function DiffViewerPanel({
 										>
 											{group.entries.map((entry) => (
 												<div key={entry.id} className="kb-diff-entry">
-													{entry.isBinary ? null : viewMode === "split" ? (
+													{entry.isBinary ? (
+														<div
+															style={{
+																padding: "12px",
+																fontSize: 12,
+																color: "var(--color-text-tertiary)",
+															}}
+														>
+															Binary file not shown.
+														</div>
+													) : viewMode === "split" ? (
 														<SplitDiff
 															path={group.path}
 															oldText={entry.oldText}

web-ui/src/components/git-history/git-commit-diff-panel.test.tsx

@@ -134,7 +134,7 @@ describe("GitCommitDiffPanel", () => {
 		expect(scrollContainer.scrollTop).toBe(547);
 	});
 
-	it("shows only the file header for binary paths", async () => {
+	it("shows a binary not shown message for binary paths", async () => {
 		const diffSource: GitCommitDiffSource = {
 			type: "commit",
 			files: [
@@ -162,6 +162,7 @@ describe("GitCommitDiffPanel", () => {
 
 		expect(container.textContent).toContain("assets/logo.png");
 		expect(container.textContent).toContain("Binary");
+		expect(container.textContent).toContain("Binary file not shown.");
 		expect(container.textContent).not.toContain("No textual diff available.");
 		expect(container.querySelector(".kb-diff-row")).toBeNull();
 	});

web-ui/src/components/git-history/git-commit-diff-panel.tsx

@@ -412,9 +412,19 @@ export function GitCommitDiffPanel({
 													Renamed from <code className="font-mono">{commitFile.previousPath}</code>
 												</div>
 											) : null}
-											{!isBinaryFile && rows.length > 0 ? (
+											{isBinaryFile ? (
+												<div
+													style={{
+														padding: "12px",
+														fontSize: 12,
+														color: "var(--color-text-tertiary)",
+													}}
+												>
+													Binary file not shown.
+												</div>
+											) : rows.length > 0 ? (
 												<ReadOnlyUnifiedDiff rows={rows} path={path} />
-											) : !isBinaryFile ? (
+											) : (
 												<div
 													style={{
 														padding: "12px",
@@ -424,7 +434,7 @@ export function GitCommitDiffPanel({
 												>
 													No textual diff available.
 												</div>
-											) : null}
+											)}
 										</div>
 									</div>
 								) : null}