Files
kuber/lib/workspace.ts
T
2026-10-06 15:31:51 +00:00

732 lines
22 KiB
TypeScript

import { execFile, spawn } from "node:child_process";
import { createHash } from "node:crypto";
import { constants } from "node:fs";
import { access } from "node:fs/promises";
import {
chmod,
lstat,
mkdir,
open,
readdir,
readlink,
realpath,
symlink,
readFile,
} from "node:fs/promises";
import type { Stats } from "node:fs";
import {
delimiter,
dirname,
isAbsolute,
relative,
resolve,
sep,
} from "node:path";
import { promisify } from "node:util";
import {
BUILD_PROTOCOL_VERSION,
assertSha256Digest,
type Sha256Digest,
type WorkspaceFile,
type WorkspaceManifest,
} from "../shared/build-protocol";
const execFileAsync = promisify(execFile);
export async function gitAvailable(): Promise<boolean> {
for (const directory of (process.env.PATH ?? "").split(delimiter)) {
if (!directory) continue;
try {
await access(
resolve(directory, process.platform === "win32" ? "git.exe" : "git"),
constants.X_OK,
);
return true;
} catch {
/* Continue searching PATH. */
}
}
return false;
}
export type WorkspaceBlob = {
digest: Sha256Digest;
data: Uint8Array;
};
export type WorkspaceSnapshot = {
manifest: WorkspaceManifest;
digest: Sha256Digest;
blobs: WorkspaceBlob[];
};
export type BlobReader =
| ((digest: Sha256Digest) => Promise<Uint8Array>)
| { get(digest: Sha256Digest): Promise<Uint8Array> };
function digest(data: Uint8Array | string): Sha256Digest {
return `sha256:${createHash("sha256").update(data).digest("hex")}`;
}
export function validateWorkspacePath(path: string): void {
if (
!path ||
path.includes("\0") ||
path.includes("\\") ||
isAbsolute(path) ||
path.split("/").some((part) => !part || part === "." || part === "..")
) {
throw new Error(`Unsafe workspace path: ${JSON.stringify(path)}`);
}
}
function isWithin(root: string, candidate: string): boolean {
const path = relative(root, candidate);
return (
path === "" ||
(!path.startsWith(`..${sep}`) && path !== ".." && !isAbsolute(path))
);
}
function assertSafeSymlinkTarget(
root: string,
path: string,
target: string,
symlinks: Set<string>,
): void {
const source = resolve(root, path);
if (
target.includes("\0") ||
isAbsolute(target) ||
!isWithin(root, resolve(dirname(source), target))
)
throw new Error(`Symlink escapes workspace: ${path} -> ${target}`);
// Inspect raw components before normalization: `sub/..` still traverses
// `sub` when it is a symlink, even though it resolves lexically to `.`.
const components = path.split("/").slice(0, -1);
for (const part of target.split(sep === "\\" ? /[\\/]/ : "/")) {
if (!part || part === ".") continue;
if (part === "..") components.pop();
else {
components.push(part);
if (symlinks.has(components.join("/")))
throw new Error(
`Symlink traverses snapshot symlink: ${path} -> ${target}`,
);
}
}
}
async function gitFiles(root: string, args: string[]): Promise<string[]> {
const { stdout } = await execFileAsync(
"git",
["-C", root, "ls-files", "-z", ...args],
{
encoding: "buffer",
maxBuffer: 64 * 1024 * 1024,
},
);
const decoder = new TextDecoder("utf-8", { fatal: true });
const files: string[] = [];
let start = 0;
for (
let end = stdout.indexOf(0);
end !== -1;
end = stdout.indexOf(0, start)
) {
if (end > start) files.push(decoder.decode(stdout.subarray(start, end)));
start = end + 1;
}
return files;
}
async function selectedFiles(root: string): Promise<string[]> {
const [normal, dotenv] = await Promise.all([
gitFiles(root, ["--cached", "--others", "--exclude-standard"]),
gitFiles(root, [
"--others",
"--ignored",
"--exclude-standard",
"--",
".env*",
"**/.env*",
]),
]);
return [...new Set([...normal, ...dotenv])].sort((a, b) =>
Buffer.from(a).compare(Buffer.from(b)),
);
}
async function selectedAutoFiles(root: string): Promise<string[]> {
// ls-files includes tracked files even when they now match .gitignore.
// --no-index makes check-ignore apply the same rules to those paths too.
const candidates = await gitFiles(root, [
"--cached",
"--others",
"--exclude-standard",
]);
if (candidates.length === 0) return [];
const child = spawn("git", [
"-C",
root,
"check-ignore",
"--no-index",
"--stdin",
"-z",
]);
const output: Buffer[] = [];
const errors: Buffer[] = [];
child.stdout.on("data", (chunk: Buffer) => output.push(chunk));
child.stderr.on("data", (chunk: Buffer) => errors.push(chunk));
const completed = new Promise<void>((resolveExit, rejectExit) => {
child.once("error", rejectExit);
child.once("close", (code) => {
if (code === 0 || code === 1) resolveExit();
else
rejectExit(
new Error(
Buffer.concat(errors).toString("utf8") ||
`git check-ignore exited with ${code}`,
),
);
});
});
child.stdin.end(Buffer.from(`${candidates.join("\0")}\0`));
await completed;
const decoder = new TextDecoder("utf-8", { fatal: true });
// Decode strictly, like ls-files, so an invalid path cannot silently alias another.
const ignored = new Set<string>();
const bytes = Buffer.concat(output);
let start = 0;
for (let end = bytes.indexOf(0); end !== -1; end = bytes.indexOf(0, start)) {
if (end > start) ignored.add(decoder.decode(bytes.subarray(start, end)));
start = end + 1;
}
const selected = candidates
.filter((path) => !ignored.has(path))
.sort((a, b) => Buffer.from(a).compare(Buffer.from(b)));
rejectSelectedAutoSecrets(selected);
return selected;
}
function rejectSelectedAutoSecrets(paths: string[]): void {
const unsafe = paths.find((path) => {
const parts = path.split("/");
const name = parts.at(-1)!.toLowerCase();
const directories = parts.slice(0, -1).map((part) => part.toLowerCase());
const dotenv =
/^\.env(?:$|[._-]|rc$)/.test(name) &&
!/^\.env[._-](?:example|sample|template)(?:$|[._-])/.test(name);
const credentialDirectory = directories.some((part) =>
[
".aws",
".azure",
".gnupg",
".ssh",
".kube",
".docker",
"credentials",
"secrets",
].includes(part),
);
const conventionalCredential =
[".npmrc", ".pypirc", ".netrc"].includes(name) ||
/^(?:id_(?:rsa|dsa|ecdsa|ed25519)|identity)(?:$|\.)/.test(name) ||
/^(?:credentials?|.*_credentials?)(?:\.|$)/.test(name) ||
/\.(?:pem|key|p12|pfx|keystore)$/.test(name);
return dotenv || credentialDirectory || conventionalCredential;
});
if (unsafe) {
throw new Error(
`build: auto refuses to snapshot potential credentials at ${unsafe}; add it to .gitignore and retry`,
);
}
}
async function filesystemFiles(root: string): Promise<string[]> {
const files: string[] = [];
const ignoredDirectories = new Set([
".git",
".hg",
".svn",
"node_modules",
"vendor",
"bower_components",
".venv",
"venv",
"__pycache__",
".tox",
".mypy_cache",
".pytest_cache",
".next",
".nuxt",
".svelte-kit",
".cache",
".turbo",
"dist",
"build",
"coverage",
"target",
"out",
"tmp",
"temp",
]);
const sensitiveDirectory =
/(?:^|[-_.])(?:secrets?|credentials?|configs?)(?:$|[-_.])/i;
const ignoreRules: Array<{
base: string;
pattern: string;
directory: boolean;
}> = [];
const loadIgnore = async (directory: string): Promise<void> => {
try {
const content = await readFile(resolve(directory, ".gitignore"), "utf8");
for (const raw of content.split(/\r?\n/)) {
const line = raw.trim();
if (!line || line.startsWith("#")) continue;
// Negations are skipped: safely re-including descendants requires
// Git's parent-directory semantics, so fallback stays fail-closed.
if (line.startsWith("!")) continue;
const rule = line.replace(/^\//, "");
if (rule)
ignoreRules.push({
base: relative(root, directory).split(sep).join("/"),
pattern: rule.replace(/\/$/, ""),
directory: line.endsWith("/"),
});
}
} catch (error) {
if ((error as NodeJS.ErrnoException).code !== "ENOENT") throw error;
}
};
const ignoredByRules = (path: string, isDirectory: boolean): boolean => {
let ignored = false;
for (const rule of ignoreRules) {
const prefix = rule.base ? `${rule.base}/` : "";
if (rule.base && path !== rule.base && !path.startsWith(prefix)) continue;
const local =
rule.base && path.startsWith(prefix) ? path.slice(prefix.length) : path;
const glob = rule.pattern
.replace(/[.+^${}()|[\]\\]/g, "\\$&")
.replace(/\*\*/g, "__DOUBLESTAR__")
.replace(/\*/g, "[^/]*")
.replace(/\?/g, "[^/]")
.replace(/__DOUBLESTAR__/g, ".*");
const matcher = new RegExp(`^(?:${glob})(?:/.*)?$`);
const basenameMatcher = new RegExp(`^(?:${glob})$`);
if (
(matcher.test(local) ||
local.split("/").some((part) => basenameMatcher.test(part))) &&
(!rule.directory || isDirectory || local.includes("/"))
)
ignored = true;
}
return ignored;
};
const visit = async (directory: string): Promise<void> => {
await loadIgnore(directory);
for (const entry of await readdir(directory, { withFileTypes: true })) {
const source = resolve(directory, entry.name);
const path = relative(root, source).split(sep).join("/");
if (entry.isDirectory()) {
if (
ignoredDirectories.has(entry.name) ||
sensitiveDirectory.test(entry.name) ||
ignoredByRules(path, true)
)
continue;
await visit(source);
} else if (entry.isFile() || entry.isSymbolicLink()) {
if (ignoredByRules(path, false)) continue;
if (
/^\.env/i.test(entry.name) ||
/(?:secret|credential|password|token|private[-_.]?key)/i.test(
entry.name,
) ||
/^(?:id_rsa|id_ed25519|known_hosts|config\.json|\.npmrc|\.pypirc|\.netrc)$/i.test(
entry.name,
)
)
continue;
files.push(path);
} else {
throw new Error(
`Special files are not allowed in workspaces: ${relative(root, source).split(sep).join("/")}`,
);
}
}
};
await visit(root);
return files.sort((a, b) => Buffer.from(a).compare(Buffer.from(b)));
}
async function isGitRepository(root: string): Promise<boolean> {
if (!(await gitAvailable())) return false;
try {
await execFileAsync("git", ["-C", root, "rev-parse", "--show-toplevel"]);
return true;
} catch (error) {
const code = (error as { code?: unknown }).code;
const exitCode = (error as { exitCode?: number }).exitCode;
const message = error instanceof Error ? error.message : String(error);
if (code === "ENOENT") return false;
if (
exitCode === 128 ||
code === 128 ||
/not a git repository/i.test(message)
)
return false;
throw error;
}
}
async function rejectSelectedSpecialFiles(root: string): Promise<void> {
// Let Git identify ignored directories using its own ignore engine. With
// --directory, ignored trees are returned as directory entries, so the
// filesystem walk below can prune them without visiting their contents.
const ignoredResult = await execFileAsync(
"git",
[
"-C",
root,
"ls-files",
"--others",
"--ignored",
"--exclude-standard",
"--directory",
"-z",
],
{ encoding: "buffer", maxBuffer: 64 * 1024 * 1024 },
);
const decoder = new TextDecoder("utf-8", { fatal: true });
const ignoredDirectories = new Set<string>();
let start = 0;
for (
let end = ignoredResult.stdout.indexOf(0);
end !== -1;
end = ignoredResult.stdout.indexOf(0, start)
) {
if (end > start) {
const path = decoder.decode(ignoredResult.stdout.subarray(start, end));
if (path.endsWith("/")) ignoredDirectories.add(path.slice(0, -1));
}
start = end + 1;
}
const specialPaths: string[] = [];
const visit = async (directory: string): Promise<void> => {
for (const entry of await readdir(directory, { withFileTypes: true })) {
if (directory === root && entry.name === ".git") continue;
const source = resolve(directory, entry.name);
const path = relative(root, source).split(sep).join("/");
if (entry.isDirectory()) {
if (!ignoredDirectories.has(path)) await visit(source);
} else if (!entry.isFile() && !entry.isSymbolicLink()) {
specialPaths.push(path);
}
}
};
await visit(root);
if (specialPaths.length === 0) return;
const input = Buffer.from(`${specialPaths.join("\0")}\0`);
const child = spawn("git", ["-C", root, "check-ignore", "--stdin", "-z"]);
const output: Buffer[] = [];
const errors: Buffer[] = [];
child.stdout.on("data", (chunk: Buffer) => output.push(chunk));
child.stderr.on("data", (chunk: Buffer) => errors.push(chunk));
const completed = new Promise<void>((resolveExit, rejectExit) => {
child.once("error", rejectExit);
child.once("close", (code) => {
if (code === 0 || code === 1) resolveExit();
else
rejectExit(
new Error(
Buffer.concat(errors).toString("utf8") ||
`git check-ignore exited with ${code}`,
),
);
});
});
child.stdin.end(input);
await completed;
const stdout = Buffer.concat(output);
const ignored = new Set<string>();
start = 0;
for (
let end = stdout.indexOf(0);
end !== -1;
end = stdout.indexOf(0, start)
) {
if (end > start) ignored.add(decoder.decode(stdout.subarray(start, end)));
start = end + 1;
}
for (const path of specialPaths) {
const name = path.slice(path.lastIndexOf("/") + 1);
if (!ignored.has(path) || name.startsWith(".env"))
throw new Error(`Special files are not allowed in workspaces: ${path}`);
}
}
function canonicalManifest(manifest: WorkspaceManifest): string {
return JSON.stringify({
version: manifest.version,
files: manifest.files.map((file) => ({
path: file.path,
type: file.type,
digest: file.digest,
size: file.size,
mode: file.mode,
})),
});
}
export function serializeWorkspaceManifest(
manifest: WorkspaceManifest,
): Uint8Array {
validateWorkspaceManifest(manifest);
return Buffer.from(canonicalManifest(manifest));
}
export function workspaceManifestDigest(
manifest: WorkspaceManifest,
): Sha256Digest {
return digest(serializeWorkspaceManifest(manifest));
}
export function validateWorkspaceManifest(manifest: WorkspaceManifest): void {
if (
manifest.version !== BUILD_PROTOCOL_VERSION ||
!Array.isArray(manifest.files)
) {
throw new Error("Unsupported workspace manifest");
}
let previous = "";
const seen = new Set<string>();
for (const file of manifest.files) {
validateWorkspacePath(file.path);
assertSha256Digest(file.digest);
if (!Number.isSafeInteger(file.size) || file.size < 0)
throw new Error(`Invalid size for ${file.path}`);
if (
(file.type === "file" && file.mode !== 0o644 && file.mode !== 0o755) ||
(file.type === "symlink" && file.mode !== 0o777)
) {
throw new Error(`Invalid mode for ${file.path}`);
}
if (file.type !== "file" && file.type !== "symlink")
throw new Error(`Invalid entry type for ${file.path}`);
if (seen.has(file.path))
throw new Error(`Duplicate workspace path: ${file.path}`);
for (const parent of file.path
.split("/")
.slice(0, -1)
.map((_, index, parts) => parts.slice(0, index + 1).join("/"))) {
if (seen.has(parent))
throw new Error(`Workspace entry is used as a directory: ${parent}`);
}
if (previous && Buffer.from(previous).compare(Buffer.from(file.path)) >= 0)
throw new Error("Workspace files must be bytewise sorted");
seen.add(file.path);
previous = file.path;
}
}
export async function enumerateWorkspace(
root: string,
mode: "default" | "auto" = "default",
): Promise<WorkspaceSnapshot> {
const repository = await realpath(root);
const gitBacked = await isGitRepository(repository);
if (mode === "auto" && !gitBacked)
throw new Error(
"build: auto requires Git and a Git repository to apply .gitignore rules safely",
);
const paths =
mode === "auto"
? await selectedAutoFiles(repository)
: gitBacked
? await (async () => {
await rejectSelectedSpecialFiles(repository);
return selectedFiles(repository);
})()
: await filesystemFiles(repository);
const files: WorkspaceFile[] = [];
const blobs = new Map<Sha256Digest, Uint8Array>();
for (const path of paths) {
validateWorkspacePath(path);
const source = resolve(repository, path);
if (!isWithin(repository, source))
throw new Error(`Workspace path escapes root: ${path}`);
// Git's index can retain files beneath a directory replaced by a symlink.
// A leaf O_NOFOLLOW does not protect against following that parent.
let parent = repository;
for (const part of path.split("/").slice(0, -1)) {
parent = resolve(parent, part);
try {
if (!(await lstat(parent)).isDirectory())
throw new Error(`Unsafe workspace parent: ${path}`);
} catch (error) {
if ((error as NodeJS.ErrnoException).code === "ENOENT") break;
throw error;
}
}
let stat: Stats;
try {
stat = await lstat(source);
} catch (error) {
if ((error as NodeJS.ErrnoException).code === "ENOENT") continue;
throw error;
}
let data: Uint8Array;
let entry: WorkspaceFile;
if (stat.isSymbolicLink()) {
const target = await readlink(source);
data = Buffer.from(target);
entry = {
path,
type: "symlink",
digest: digest(data),
size: data.byteLength,
mode: 0o777,
};
} else if (stat.isFile()) {
let mode: 0o644 | 0o755;
const handle = await open(
source,
constants.O_RDONLY | constants.O_NOFOLLOW,
);
try {
const opened = await handle.stat();
if (!opened.isFile()) throw new Error(`Not a regular file: ${path}`);
mode = opened.mode & 0o111 ? 0o755 : 0o644;
data = await handle.readFile();
} finally {
await handle.close();
}
entry = {
path,
type: "file",
digest: digest(data),
size: data.byteLength,
mode,
};
} else {
throw new Error(`Special files are not allowed in workspaces: ${path}`);
}
files.push(entry);
blobs.set(entry.digest, data);
}
const manifest = {
version: BUILD_PROTOCOL_VERSION,
files,
} satisfies WorkspaceManifest;
const symlinks = new Set(
files.filter((file) => file.type === "symlink").map((file) => file.path),
);
for (const file of files) {
if (file.type === "symlink")
assertSafeSymlinkTarget(
repository,
file.path,
Buffer.from(blobs.get(file.digest)!).toString("utf8"),
symlinks,
);
}
return {
manifest,
digest: workspaceManifestDigest(manifest),
blobs: [...blobs].map(([blobDigest, data]) => ({
digest: blobDigest,
data,
})),
};
}
async function ensureParentDirectories(
root: string,
path: string,
): Promise<void> {
let current = root;
for (const part of path.split("/").slice(0, -1)) {
current = resolve(current, part);
try {
const stat = await lstat(current);
if (!stat.isDirectory())
throw new Error(`Workspace parent is not a directory: ${path}`);
} catch (error) {
if ((error as NodeJS.ErrnoException).code !== "ENOENT") throw error;
await mkdir(current, { mode: 0o755 });
}
}
}
async function readBlob(
reader: BlobReader,
blobDigest: Sha256Digest,
): Promise<Uint8Array> {
return typeof reader === "function"
? reader(blobDigest)
: reader.get(blobDigest);
}
export async function materializeWorkspace(
destination: string,
manifest: WorkspaceManifest,
reader: BlobReader,
): Promise<void> {
validateWorkspaceManifest(manifest);
const symlinks = new Set(
manifest.files
.filter((file) => file.type === "symlink")
.map((file) => file.path),
);
const linkTargets = new Map<string, string>();
for (const file of manifest.files) {
if (file.type !== "symlink") continue;
const data = await readBlob(reader, file.digest);
if (data.byteLength !== file.size || digest(data) !== file.digest)
throw new Error(`Blob verification failed for ${file.path}`);
const linkTarget = Buffer.from(data).toString("utf8");
assertSafeSymlinkTarget(destination, file.path, linkTarget, symlinks);
linkTargets.set(file.path, linkTarget);
}
await mkdir(destination, { recursive: false, mode: 0o755 });
const root = await realpath(destination);
for (const file of manifest.files.filter((entry) => entry.type === "file")) {
await ensureParentDirectories(root, file.path);
const data = await readBlob(reader, file.digest);
if (data.byteLength !== file.size || digest(data) !== file.digest)
throw new Error(`Blob verification failed for ${file.path}`);
const target = resolve(root, file.path);
const handle = await open(
target,
constants.O_CREAT | constants.O_EXCL | constants.O_WRONLY,
file.mode,
);
try {
await handle.writeFile(data);
await handle.sync();
} finally {
await handle.close();
}
await chmod(target, file.mode);
}
for (const file of manifest.files.filter(
(entry) => entry.type === "symlink",
)) {
await ensureParentDirectories(root, file.path);
const linkTarget = linkTargets.get(file.path)!;
const target = resolve(root, file.path);
await symlink(linkTarget, target);
}
}