playground: ship findutils locate and updatedb in WASM

Build the locate and updatedb binaries from uutils/findutils alongside
find and advertise them in the playground. The standalone WASM matrix now
takes a list of bins per module, emitting one .wasm per binary.

xargs is intentionally omitted: it must spawn child processes, which the
browser WASI sandbox does not support.

Author: sylvestre

.github/workflows/website.yml

@@ -118,11 +118,13 @@ jobs:
         name: wasm-coreutils
         path: wasm-out
 
-  # grep, find, diffutils and sed each ship a single standalone binary built the
-  # same way, so they share one matrix job. grep/find pull in the Oniguruma C
-  # library (onig_sys) and need a WASI sysroot to compile its bundled C sources;
-  # diffutils and sed are pure Rust and set needs_wasi: false. The `order` field
-  # keeps the merged command/version lists in a stable sequence in `build`.
+  # grep, findutils, diffutils and sed ship standalone binaries built the same
+  # way, so they share one matrix job. Each entry lists its `bins` (findutils
+  # provides three: find, locate, updatedb); every bin emits its own .wasm.
+  # grep/findutils pull in the Oniguruma C library (onig_sys) and need a WASI
+  # sysroot to compile its bundled C sources; diffutils and sed are pure Rust and
+  # set needs_wasi: false. The `order` field keeps the merged command/version
+  # lists in a stable sequence in `build`.
   wasm-standalone:
     name: Build ${{ matrix.name }} WASM
     runs-on: ubuntu-latest
@@ -132,28 +134,31 @@ jobs:
         include:
         - name: grep
           repo: uutils/grep
-          bin: grep
+          bins: grep
           needs_wasi: true
           commands: '"grep"'
           version_const: UUTILS_GREP_VERSION
           order: 10
+        # findutils ships several standalone binaries; build find, locate and
+        # updatedb so the playground can offer them. (xargs is omitted: it must
+        # spawn child processes, which the browser WASI sandbox can't do.)
         - name: find
           repo: uutils/findutils
-          bin: find
+          bins: find locate updatedb
           needs_wasi: true
-          commands: '"find"'
+          commands: '"find","locate","updatedb"'
           version_const: UUTILS_FINDUTILS_VERSION
           order: 20
         - name: diffutils
           repo: uutils/diffutils
-          bin: diffutils
+          bins: diffutils
           needs_wasi: false
           commands: '"diff","cmp"'
           version_const: UUTILS_DIFFUTILS_VERSION
           order: 30
         - name: sed
           repo: uutils/sed
-          bin: sed
+          bins: sed
           needs_wasi: false
           commands: '"sed"'
           version_const: UUTILS_SED_VERSION
@@ -187,17 +192,21 @@ jobs:
           export CC_wasm32_wasip1="$WASI_SDK_PATH/bin/clang"
           export CFLAGS_wasm32_wasip1="--sysroot=$WASI_SDK_PATH/share/wasi-sysroot"
         fi
-        wasm="${{ matrix.bin }}.wasm"
         cd ${{ matrix.name }}
-        cargo build --release --target wasm32-wasip1 --bin ${{ matrix.bin }}
-        test -f "target/wasm32-wasip1/release/${wasm}"
         mkdir -p ../wasm-out
-        cp "target/wasm32-wasip1/release/${wasm}" "../wasm-out/${wasm}"
-        # Optimize WASM size if wasm-opt is available
-        if command -v wasm-opt &> /dev/null; then
-          wasm-opt -Oz "../wasm-out/${wasm}" -o "../wasm-out/${wasm}"
-        fi
-        echo "${{ matrix.name }} WASM binary size: $(du -h ../wasm-out/${wasm} | cut -f1)"
+        # A module may ship several standalone binaries (findutils → find,
+        # locate, updatedb); build each and emit its own .wasm file.
+        for bin in ${{ matrix.bins }}; do
+          wasm="${bin}.wasm"
+          cargo build --release --target wasm32-wasip1 --bin "${bin}"
+          test -f "target/wasm32-wasip1/release/${wasm}"
+          cp "target/wasm32-wasip1/release/${wasm}" "../wasm-out/${wasm}"
+          # Optimize WASM size if wasm-opt is available
+          if command -v wasm-opt &> /dev/null; then
+            wasm-opt -Oz "../wasm-out/${wasm}" -o "../wasm-out/${wasm}"
+          fi
+          echo "${bin} WASM binary size: $(du -h ../wasm-out/${wasm} | cut -f1)"
+        done
         # Advertise this module's commands in the playground's command list.
         echo '${{ matrix.commands }}' > ../wasm-out/commands-${{ matrix.order }}-${{ matrix.name }}.txt
         # Record the commit used to build this WASM module.

content/playground.md

@@ -29,7 +29,8 @@ template = "page.html"
   document.addEventListener("DOMContentLoaded", function() {
     initPlayground("wasm-playground");
     // Build a "Load" button per optional standalone module (grep, find,
-    // diffutils). These ship as their own WASM modules and load on demand to
+    // locate, updatedb, diffutils, sed). These ship as their own WASM
+    // modules and load on demand to
     // keep the initial page download light; running a command auto-loads its
     // module too (e.g. diff/cmp both come from the diffutils module).
     var loaderBar = document.getElementById("playground-loaders");

static/js/wasm-terminal.js

@@ -15,15 +15,19 @@ if (typeof SharedArrayBuffer === "undefined") {
 
 const WASM_URL = "/wasm/uutils.wasm";
 // Some utilities ship as their own standalone WASM modules rather than as part
-// of the coreutils multicall binary (grep lives in uutils/grep, find in
-// uutils/findutils, diff and cmp in uutils/diffutils, sed in uutils/sed). Each
-// module is loaded on demand and is optional — see loadStandalone. A single
-// module can provide several commands (diffutils → diff, cmp), which the
-// diffutils binary dispatches on argv[0], so each command is invoked directly
-// by its own name.
+// of the coreutils multicall binary (grep lives in uutils/grep, find/locate/
+// updatedb in uutils/findutils, diff and cmp in uutils/diffutils, sed in
+// uutils/sed). Each module is loaded on demand and is optional — see
+// loadStandalone. A single module can provide several commands (diffutils →
+// diff, cmp), which the binary dispatches on argv[0], so each command is invoked
+// directly by its own name. findutils ships separate binaries, so each is its
+// own module. (xargs is intentionally absent: it must spawn child processes,
+// which the browser WASI sandbox can't do.)
 const STANDALONE_MODULES = {
   grep: { url: "/wasm/grep.wasm", commands: ["grep"] },
   find: { url: "/wasm/find.wasm", commands: ["find"] },
+  locate: { url: "/wasm/locate.wasm", commands: ["locate"] },
+  updatedb: { url: "/wasm/updatedb.wasm", commands: ["updatedb"] },
   diffutils: { url: "/wasm/diffutils.wasm", commands: ["diff", "cmp"] },
   sed: { url: "/wasm/sed.wasm", commands: ["sed"] },
 };
@@ -75,7 +79,7 @@ const FALLBACK_COMMANDS = [
   "sha1sum", "sha224sum", "sha256sum", "sha384sum", "sha512sum",
   "shred", "shuf", "sleep", "sum", "tee", "true", "truncate",
   "uname", "unexpand", "uniq", "unlink", "vdir", "wc",
-  "grep", "find", "diff", "cmp", "sed",
+  "grep", "find", "locate", "updatedb", "diff", "cmp", "sed",
 ];
 const AVAILABLE_COMMANDS =
   (typeof WASM_COMMANDS !== "undefined" && Array.isArray(WASM_COMMANDS) && WASM_COMMANDS.length > 0)
@@ -233,7 +237,8 @@ async function initWasm() {
   if (wasmReady) return;
   try {
     // Only the coreutils multicall binary loads eagerly; the standalone
-    // modules (grep, find, diffutils) are fetched on demand — see loadStandalone.
+    // modules (grep, find, locate, updatedb, diffutils, sed) are fetched
+    // on demand — see loadStandalone.
     await Promise.all([loadWasiShim(), loadWasm()]);
     wasmReady = true;
   } catch (e) {
@@ -626,8 +631,9 @@ async function executeCommandLine(line) {
       return `uutils: command not found: ${cmd}\nType 'help' for available commands.\n`;
     }
 
-    // Some utilities (grep, find, diff, cmp) are separate WASM modules rather
-    // than part of the coreutils multicall binary, and are loaded on demand.
+    // Some utilities (grep, find, locate, updatedb, diff, cmp, sed) are
+    // separate WASM modules rather than part of the coreutils multicall binary,
+    // and are loaded on demand.
     // Fetch the module the first time one of its commands is used (no-op once
     // cached; diff and cmp share the single diffutils module).
     const moduleName = STANDALONE_COMMAND_MODULE[cmd];
@@ -955,7 +961,7 @@ async function initPlayground(containerId) {
     terminal.writeln("");
     terminal.writeln("Type \x1b[1;32mhelp\x1b[0m for available commands.");
     terminal.writeln("Sample data files: names.txt, numbers.txt, fruits.txt, csv.txt, words.txt");
-    terminal.writeln("\x1b[2mgrep, find, sed and diff/cmp load on demand — just run them, or use the buttons below.\x1b[0m");
+    terminal.writeln("\x1b[2mgrep, find/locate/updatedb, sed and diff/cmp load on demand — just run them, or use the buttons below.\x1b[0m");
   } catch (e) {
     terminal.writeln(" \x1b[1;31mfailed\x1b[0m");
     terminal.writeln("Failed to load WASM binary. Commands are not available.");
@@ -1003,8 +1009,9 @@ window.uutilsExecute = executeCommandLine;
 window.runInTerminal = runInTerminal;
 window.setLocale = setLocale;
 
-// On-demand loading of the optional standalone modules (grep, find, diffutils),
-// used by the "Load" buttons on the playground page. Keyed by module name; one
+// On-demand loading of the optional standalone modules (grep, find, locate,
+// updatedb, diffutils, sed), used by the "Load" buttons on the
+// playground page. Keyed by module name; one
 // module may back several commands (diffutils → diff, cmp).
 window.uutilsPrograms = Object.keys(STANDALONE_MODULES);
 window.loadProgram = (mod) => loadStandalone(mod);