Make Coreutils runtime aliases match built programs

apps/browser-demos/pages/test-runner/main.ts

@@ -17,6 +17,7 @@ import coreutilsWasmUrl from "@binaries/programs/wasm32/coreutils.wasm?url";
 import grepWasmUrl from "@binaries/programs/wasm32/grep.wasm?url";
 import sedWasmUrl from "@binaries/programs/wasm32/sed.wasm?url";
 import genCatWasmUrl from "@binaries/programs/wasm32/posix-utils-lite/gencat.wasm?url";
+import { COREUTILS_NAMES } from "../../lib/init/shell-binaries";
 
 interface DataFile {
   path: string;
@@ -50,23 +51,6 @@ let grepBytes: ArrayBuffer | null = null;
 let sedBytes: ArrayBuffer | null = null;
 let genCatBytes: ArrayBuffer | null = null;
 
-const COREUTILS_NAMES = [
-  "arch", "b2sum", "base32", "base64", "basename", "basenc", "cat",
-  "chcon", "chgrp", "chmod", "chown", "chroot", "cksum", "comm", "cp",
-  "csplit", "cut", "date", "dd", "df", "dir", "dircolors", "dirname",
-  "du", "echo", "env", "expand", "expr", "factor", "false", "fmt",
-  "fold", "groups", "head", "hostid", "id", "install", "join", "link",
-  "ln", "logname", "ls", "md5sum", "mkdir", "mkfifo", "mknod", "mktemp",
-  "mv", "nice", "nl", "nohup", "nproc", "numfmt", "od", "paste",
-  "pathchk", "pr", "printenv", "printf", "ptx", "pwd", "readlink",
-  "realpath", "rm", "rmdir", "runcon", "seq", "sha1sum", "sha224sum",
-  "sha256sum", "sha384sum", "sha512sum", "shred", "shuf", "sleep",
-  "sort", "split", "stat", "stty", "sum", "sync", "tac", "tail",
-  "tee", "test", "timeout", "touch", "tr", "true", "truncate", "tsort",
-  "tty", "uname", "unexpand", "uniq", "unlink", "vdir", "wc", "whoami",
-  "yes",
-];
-
 /** Write a binary file to the virtual filesystem. */
 function writeFileToFs(fs: import("@host/browser-kernel-host").BrowserKernel["fs"], path: string, data: ArrayBuffer): void {
   const bytes = new Uint8Array(data);

examples/run-example.ts

@@ -16,6 +16,7 @@ import { readFileSync, existsSync, statSync } from "fs";
 import { resolve, dirname, isAbsolute, relative } from "path";
 import { NodeKernelHost } from "../host/src/node-kernel-host";
 import { tryResolveBinary } from "../host/src/binary-resolver";
+import { COREUTILS_NAMES } from "../images/vfs/lib/init/shell-binaries";
 
 const repoRoot = resolve(dirname(new URL(import.meta.url).pathname), "..");
 
@@ -61,15 +62,7 @@ const testfixtureWasm = tryResolveBinary("programs/sqlite/testfixture.wasm");
 const mysqltestWasm = tryResolveBinary("programs/mariadb/mysqltest.wasm");
 const echoWasm = tryResolveBinary("programs/echo.wasm") ?? resolve(repoRoot, "examples/echo.wasm");
 
-// GNU coreutils multi-call binary supports all of these as argv[0]
-const coreutilsNames = [
-    "cat", "ls", "cp", "mv", "rm", "mkdir", "rmdir", "ln", "chmod", "chown",
-    "head", "tail", "wc", "sort", "uniq", "tr", "cut", "paste", "tee",
-    "true", "false", "yes", "env", "printenv", "printf", "expr", "test", "[",
-    "basename", "dirname", "readlink", "realpath", "stat", "touch", "date",
-    "sleep", "id", "whoami", "uname", "hostname", "pwd", "dd", "od", "md5sum",
-    "sha256sum", "base64", "seq", "factor", "nproc", "du", "df",
-];
+const coreutilsNames = [...COREUTILS_NAMES, "["];
 
 // Values may be null when a program isn't fetched/built locally.
 // Consumers filter out null entries before use.

images/rootfs/PACKAGES.toml

@@ -130,6 +130,7 @@ aliases = [
   "/usr/bin/id", "/bin/id",
   "/usr/bin/install", "/bin/install",
   "/usr/bin/join", "/bin/join",
+  "/usr/bin/kill", "/bin/kill",
   "/usr/bin/link", "/bin/link",
   "/usr/bin/ln", "/bin/ln",
   "/usr/bin/logname", "/bin/logname",

images/vfs/lib/init/shell-binaries.ts

@@ -9,7 +9,7 @@ export const COREUTILS_NAMES = [
   "chcon", "chgrp", "chmod", "chown", "chroot", "cksum", "comm", "cp",
   "csplit", "cut", "date", "dd", "df", "dir", "dircolors", "dirname",
   "du", "echo", "env", "expand", "expr", "factor", "false", "fmt",
-  "fold", "groups", "head", "hostid", "id", "install", "join", "link",
+  "fold", "groups", "head", "hostid", "id", "install", "join", "kill", "link",
   "ln", "logname", "ls", "md5sum", "mkdir", "mkfifo", "mknod", "mktemp",
   "mv", "nice", "nl", "nohup", "nproc", "numfmt", "od", "paste",
   "pathchk", "pr", "printenv", "printf", "ptx", "pwd", "readlink",

images/vfs/scripts/build-mariadb-test-vfs-image.ts

@@ -25,10 +25,11 @@ import {
   writeVfsBinary,
   symlink,
 } from "../../../host/src/vfs/image-helpers";
-import { resolveBinary, tryResolveBinary, findRepoRoot } from "../../../host/src/binary-resolver";
+import { resolveBinary, findRepoRoot } from "../../../host/src/binary-resolver";
 import { saveImage, walkAndWrite } from "./vfs-image-helpers";
 import { addDinitInit, type DinitService } from "./dinit-image-helpers";
 import { ensureSourceExtract } from "./source-extract-helper";
+import { COREUTILS_NAMES } from "../lib/init/shell-binaries";
 
 const REPO_ROOT = findRepoRoot();
 // Local source-build keeps everything under mariadb-install/; for
@@ -47,22 +48,14 @@ const SYSTEM_DATA_PATH = existsSync(join(MARIADB_LEGACY_INSTALL, "share/mysql/my
   ? join(MARIADB_LEGACY_INSTALL, "share/mysql/mysql_system_tables_data.sql")
   : join(MARIADB_SOURCE, "scripts/mysql_system_tables_data.sql");
 const DASH_PATH = resolveBinary("programs/dash.wasm");
-const COREUTILS_PATH = tryResolveBinary("programs/coreutils.wasm");
+const COREUTILS_PATH = resolveBinary("programs/coreutils.wasm");
 
 const OUT_FILE = process.env.MARIADB_TEST_VFS_OUT
   ?? join(REPO_ROOT, "apps/browser-demos/public/mariadb-test.vfs.zst");
 
 const includeAll = process.argv.includes("--all");
 
-const COREUTILS_SYMLINK_NAMES = [
-  "ls", "cat", "cp", "mv", "rm", "echo", "mkdir", "rmdir", "touch", "pwd",
-  "head", "tail", "wc", "sort", "uniq", "cut", "tr", "date", "basename",
-  "dirname", "chmod", "chown", "ln", "readlink", "true", "false", "yes",
-  "sleep", "env", "printenv", "id", "whoami", "hostname", "uname", "stat",
-  "df", "du", "tee", "nl", "paste", "tac", "rev", "expand", "unexpand",
-  "fold", "fmt", "pr", "od", "hexdump", "xxd", "sha256sum", "sha512sum",
-  "md5sum", "seq", "test", "[",
-];
+const COREUTILS_SYMLINK_NAMES = [...COREUTILS_NAMES, "["];
 
 // 185 tests verified to pass in headless Chromium with MariaDB on kandelo.
 const CURATED_TESTS = [
@@ -234,11 +227,7 @@ async function main() {
     symlink(fs, "/bin/dash", "/usr/bin/dash");
     symlink(fs, "/bin/dash", "/usr/bin/sh");
   }
-  if (COREUTILS_PATH && existsSync(COREUTILS_PATH)) {
-    writeVfsBinary(fs, "/bin/coreutils", new Uint8Array(readFileSync(COREUTILS_PATH)));
-  } else {
-    console.warn("  Warning: coreutils.wasm not found — bootstrap wrapper will fail at sleep");
-  }
+  writeVfsBinary(fs, "/bin/coreutils", new Uint8Array(readFileSync(COREUTILS_PATH)));
   for (const name of COREUTILS_SYMLINK_NAMES) {
     symlink(fs, "/bin/coreutils", `/bin/${name}`);
     symlink(fs, "/bin/coreutils", `/usr/bin/${name}`);

images/vfs/scripts/build-mariadb-vfs-image.ts

@@ -24,10 +24,11 @@ import {
   writeVfsBinary,
   symlink,
 } from "../../../host/src/vfs/image-helpers";
-import { resolveBinary, tryResolveBinary, findRepoRoot } from "../../../host/src/binary-resolver";
+import { resolveBinary, findRepoRoot } from "../../../host/src/binary-resolver";
 import { saveImage } from "./vfs-image-helpers";
 import { addDinitInit, type DinitService } from "./dinit-image-helpers";
 import { ensureSourceExtract } from "./source-extract-helper";
+import { COREUTILS_NAMES } from "../lib/init/shell-binaries";
 
 const REPO_ROOT = findRepoRoot();
 const useWasm64 = process.argv.includes("--wasm64");
@@ -52,24 +53,15 @@ const SYSTEM_DATA_PATH = existsSync(join(MARIADB_LEGACY_INSTALL, "share/mysql/my
   ? join(MARIADB_LEGACY_INSTALL, "share/mysql/mysql_system_tables_data.sql")
   : join(MARIADB_SOURCE, "scripts/mysql_system_tables_data.sql");
 const DASH_PATH = resolveBinary("programs/dash.wasm");
-// Coreutils is baked into the VFS so users get an immediate `ls`/`cat`
-// in the shell demo without waiting for a lazy-load fetch. Not strictly
-// required by the bootstrap wrapper (which is just `exec mariadbd < sql`).
-const COREUTILS_PATH = tryResolveBinary("programs/coreutils.wasm");
+// Coreutils is a declared package dependency and is baked into the VFS so
+// every emitted symlink has a real multicall target.
+const COREUTILS_PATH = resolveBinary("programs/coreutils.wasm");
 
 const OUT_FILE = useWasm64
   ? join(REPO_ROOT, "apps/browser-demos/public/mariadb-64.vfs.zst")
   : join(REPO_ROOT, "apps/browser-demos/public/mariadb.vfs.zst");
 
-const COREUTILS_SYMLINK_NAMES = [
-  "ls", "cat", "cp", "mv", "rm", "echo", "mkdir", "rmdir", "touch", "pwd",
-  "head", "tail", "wc", "sort", "uniq", "cut", "tr", "date", "basename",
-  "dirname", "chmod", "chown", "ln", "readlink", "true", "false", "yes",
-  "sleep", "env", "printenv", "id", "whoami", "hostname", "uname", "stat",
-  "df", "du", "tee", "nl", "paste", "tac", "rev", "expand", "unexpand",
-  "fold", "fmt", "pr", "od", "hexdump", "xxd", "sha256sum", "sha512sum",
-  "md5sum", "seq", "test", "[",
-];
+const COREUTILS_SYMLINK_NAMES = [...COREUTILS_NAMES, "["];
 
 function commonMariadbArgs(engine: string): string[] {
   return [
@@ -175,13 +167,7 @@ async function main() {
     symlink(fs, "/bin/dash", "/usr/bin/dash");
     symlink(fs, "/bin/dash", "/usr/bin/sh");
   }
-  // Bake coreutils.wasm so the shell demo has `ls`/`cat`/etc. without
-  // waiting for a lazy-load fetch on first invocation.
-  if (COREUTILS_PATH && existsSync(COREUTILS_PATH)) {
-    writeVfsBinary(fs, "/bin/coreutils", new Uint8Array(readFileSync(COREUTILS_PATH)));
-  } else {
-    console.warn("  Warning: coreutils.wasm not found — shell demo will lack /bin/ls etc.");
-  }
+  writeVfsBinary(fs, "/bin/coreutils", new Uint8Array(readFileSync(COREUTILS_PATH)));
   for (const name of COREUTILS_SYMLINK_NAMES) {
     symlink(fs, "/bin/coreutils", `/bin/${name}`);
     symlink(fs, "/bin/coreutils", `/usr/bin/${name}`);

images/vfs/scripts/build-sqlite-test-vfs-image.ts

@@ -18,6 +18,7 @@ import {
 } from "../../../host/src/vfs/image-helpers";
 import { findRepoRoot, tryResolveBinary } from "../../../host/src/binary-resolver";
 import { saveImage, walkAndWrite } from "./vfs-image-helpers";
+import { COREUTILS_NAMES } from "../lib/init/shell-binaries";
 
 const REPO_ROOT = findRepoRoot();
 const SQLITE_DIR = join(REPO_ROOT, "packages/registry/sqlite");
@@ -31,11 +32,7 @@ const COREUTILS_PATH = tryResolveBinary("programs/coreutils.wasm");
 const OUT_FILE = process.env.SQLITE_TEST_VFS_OUT
   ?? join(REPO_ROOT, "apps/browser-demos/public/sqlite-test.vfs.zst");
 
-const COREUTILS_SYMLINK_NAMES = [
-  "cat", "chmod", "cp", "date", "dirname", "echo", "env", "expr", "false",
-  "head", "ln", "ls", "mkdir", "mv", "pwd", "rm", "rmdir", "sed", "sleep",
-  "sort", "tail", "tee", "test", "touch", "tr", "true", "uname", "wc", "[",
-];
+const COREUTILS_SYMLINK_NAMES = [...COREUTILS_NAMES, "["];
 
 function checkPrereqs(): void {
   const missing: string[] = [];

packages/registry/coreutils/.gitignore

@@ -1,2 +1,3 @@
 coreutils-src/
+coreutils-*-src/
 bin/

packages/registry/coreutils/build-coreutils.sh

@@ -1,15 +1,15 @@
 #!/usr/bin/env bash
 set -euo pipefail
 
-# Build GNU coreutils 9.5 for wasm32-posix-kernel.
+# Build GNU coreutils 9.6 for wasm32-posix-kernel.
 #
 # Uses the SDK's wasm32posix-configure wrapper for cross-compilation.
 # Output: packages/registry/coreutils/bin/*.wasm
 
-COREUTILS_VERSION="${COREUTILS_VERSION:-9.5}"
+COREUTILS_VERSION="${COREUTILS_VERSION:-9.6}"
 SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
 REPO_ROOT="$(cd "$SCRIPT_DIR/../../.." && pwd)"
-SRC_DIR="$SCRIPT_DIR/coreutils-src"
+SRC_DIR="$SCRIPT_DIR/coreutils-${COREUTILS_VERSION}-src"
 BIN_DIR="$SCRIPT_DIR/bin"
 SYSROOT="$REPO_ROOT/sysroot"
 
@@ -41,8 +41,10 @@ fi
 cd "$SRC_DIR"
 
 # --- Configure ---
-if [ ! -f Makefile ]; then
-    echo "==> Configuring coreutils for wasm32..."
+# Rerun configure even when this source tree was built before. Package recipe
+# changes can alter the multicall member set without changing the upstream
+# version, and an old Makefile must never define the published command surface.
+echo "==> Configuring coreutils for wasm32..."
 
     # GNU coreutils + gnulib does hundreds of configure tests.
     # Cross-compilation link-based tests fail, so we override with correct
@@ -228,15 +230,15 @@ if [ ! -f Makefile ]; then
     wasm32posix-configure \
         --disable-nls \
         --without-selinux \
-        --without-gmp \
+        --without-libgmp \
         --disable-acl \
         --disable-xattr \
         --enable-single-binary=symlinks \
+        --enable-install-program=arch \
         --enable-no-install-program=stdbuf,pinky,who,users,uptime \
         2>&1 | tail -30
 
     echo "==> Configure complete."
-fi
 
 # --- Build ---
 echo "==> Building coreutils..."

packages/registry/coreutils/build.toml

@@ -1,7 +1,7 @@
 script_path = "packages/registry/coreutils/build-coreutils.sh"
 repo_url    = "https://github.com/brandonpayton/kandelo.git"
 commit      = "8c53383229fab78f97b098c3207a655159c03041"
-revision    = 2
+revision    = 3
 
 [binary]
 index_url = "https://github.com/Automattic/kandelo/releases/download/binaries-abi-v{abi}/index.toml"

packages/registry/coreutils/test/coreutils.test.ts

@@ -7,8 +7,10 @@
 import { describe, it, expect } from "vitest";
 import { runCentralizedProgram } from "../../../../host/test/centralized-test-helper";
 import { tryResolveBinary } from "../../../../host/src/binary-resolver";
+import { COREUTILS_NAMES } from "../../../../images/vfs/lib/init/shell-binaries";
 
 const coreutilsBinary = tryResolveBinary("programs/coreutils.wasm");
+const dashBinary = tryResolveBinary("programs/dash.wasm");
 
 const hasCoreutils = !!coreutilsBinary;
 
@@ -41,6 +43,46 @@ describe.skipIf(!hasCoreutils)("GNU coreutils", () => {
     expect(result.exitCode).toBe(1);
   });
 
+  it("arch reports the wasm32 machine", async () => {
+    const result = await runCentralizedProgram({
+      programPath: coreutilsBinary!,
+      argv: ["arch"],
+      timeout: 10_000,
+    });
+    expect(result.exitCode).toBe(0);
+    expect(result.stdout.trim()).toBe("wasm32");
+  });
+
+  it("kill is included in the multicall binary", async () => {
+    const result = await runCentralizedProgram({
+      programPath: coreutilsBinary!,
+      argv: ["kill", "--version"],
+      timeout: 10_000,
+    });
+    expect(result.exitCode).toBe(0);
+    expect(result.stdout).toContain("kill (GNU coreutils) 9.6");
+  });
+
+  it.skipIf(!dashBinary)(
+    "executes arch and kill through /bin aliases",
+    async () => {
+      const execPrograms = new Map(
+        COREUTILS_NAMES.map(
+          (name) => [`/bin/${name}`, coreutilsBinary!] as const,
+        ),
+      );
+      const result = await runCentralizedProgram({
+        programPath: dashBinary!,
+        argv: ["dash", "-c", "/bin/arch; /bin/kill --version"],
+        execPrograms,
+        timeout: 10_000,
+      });
+      expect(result.exitCode).toBe(0);
+      expect(result.stdout).toContain("wasm32\n");
+      expect(result.stdout).toContain("kill (GNU coreutils) 9.6");
+    },
+  );
+
   it("printf formats output", async () => {
     const result = await runCentralizedProgram({
       programPath: coreutilsBinary!,

packages/registry/lamp/build.toml

@@ -31,7 +31,7 @@ inputs = [
 ]
 repo_url    = "https://github.com/brandonpayton/kandelo.git"
 commit      = "8c53383229fab78f97b098c3207a655159c03041"
-revision    = 7
+revision    = 8
 
 [binary]
 index_url = "https://github.com/Automattic/kandelo/releases/download/binaries-abi-v{abi}/index.toml"

packages/registry/mariadb-test/build.toml

@@ -1,7 +1,25 @@
 script_path = "packages/registry/mariadb-test/build-mariadb-test.sh"
+inputs = [
+  "packages/registry/mariadb-test/build-mariadb-test.sh",
+  "packages/registry/mariadb-test/package.toml",
+  "images/vfs/scripts/build-mariadb-test-vfs-image.ts",
+  "images/vfs/scripts/dinit-image-helpers.ts",
+  "images/vfs/scripts/source-extract-helper.ts",
+  "images/vfs/scripts/vfs-image-helpers.ts",
+  "images/vfs/lib/init/shell-binaries.ts",
+  "host/src/binary-resolver.ts",
+  "host/src/constants.ts",
+  "host/src/statfs.ts",
+  "host/src/types.ts",
+  "host/src/vfs/image-helpers.ts",
+  "host/src/vfs/memory-fs.ts",
+  "host/src/vfs/sharedfs-vendor.ts",
+  "host/src/vfs/types.ts",
+  "host/src/vfs/zip.ts",
+]
 repo_url    = "https://github.com/brandonpayton/kandelo.git"
 commit      = "8c53383229fab78f97b098c3207a655159c03041"
-revision    = 1
+revision    = 2
 
 [binary]
 index_url = "https://github.com/Automattic/kandelo/releases/download/binaries-abi-v{abi}/index.toml"

packages/registry/mariadb-test/package.toml

@@ -10,12 +10,14 @@ kernel_abi = 7
 #     on demand.
 #   - calls `resolveBinary("programs/dash.wasm")` (strict — throws if missing)
 #     to bake `/bin/dash` + `/bin/sh` symlink
+#   - calls `resolveBinary("programs/coreutils.wasm")` (strict — throws if
+#     missing) before creating the multicall command symlinks
 #   - calls `addDinitInit(...)` which calls `resolveDinitBinaries()` (strict —
 #     throws if neither tryResolveBinary nor the source-tree path resolves)
 #     to bake the init system that boots the test runner.
 # Without these declarations a clean source-build (e.g., during manual
 # force-rebuild) fails. Folds each prereq's cache_key into ours.
-depends_on = ["mariadb@10.5.28", "dash@0.5.12", "dinit@0.19.4"]
+depends_on = ["mariadb@10.5.28", "dash@0.5.12", "coreutils@9.6", "dinit@0.19.4"]
 
 # Pre-built VFS image for the MariaDB test runner: mariadbd binary,
 # mysql-test test files (curated 185 tests), system-table SQL,