sea: support embedding assets

With this patch:

Users can now include assets by adding a key-path dictionary
to the configuration as the `assets` field. At build time, Node.js
would read the assets from the specified paths and bundle them into
the preparation blob. In the generated executable, users can retrieve
the assets using the `sea.getAsset()` and `sea.getAssetAsBlob()` API.

```json
{
  "main": "/path/to/bundled/script.js",
  "output": "/path/to/write/the/generated/blob.blob",
  "assets": {
    "a.jpg": "/path/to/a.jpg",
    "b.txt": "/path/to/b.txt"
  }
}
```

The single-executable application can access the assets as follows:

```cjs
const { getAsset } = require('node:sea');
// Returns a copy of the data in an ArrayBuffer
const image = getAsset('a.jpg');
// Returns a string decoded from the asset as UTF8.
const text = getAsset('b.txt', 'utf8');
// Returns a Blob containing the asset without copying.
const blob = getAssetAsBlob('a.jpg');
```

Drive-by: update the  documentation to include a section dedicated
to the injected main script and refer to it as "injected main
script" instead of "injected module" because it's a script, not
a module.

Author: joyeecheung

doc/api/errors.md

@@ -2360,6 +2360,13 @@ error indicates that the idle loop has failed to stop.
 An attempt was made to use operations that can only be used when building
 V8 startup snapshot even though Node.js isn't building one.
 
+<a id="ERR_NOT_IN_SEA"></a>
+
+### `ERR_NOT_IN_SEA`
+
+The operation cannot be performed when it's not in a single-executable
+application.
+
 <a id="ERR_NOT_SUPPORTED_IN_SNAPSHOT"></a>
 
 ### `ERR_NOT_SUPPORTED_IN_SNAPSHOT`
@@ -2490,6 +2497,13 @@ example, <kbd>Ctrl</kbd>+<kbd>C</kbd> was pressed.)
 
 Script execution timed out, possibly due to bugs in the script being executed.
 
+<a id="ERR_SEA_ASSET_NOT_FOUND"></a>
+
+### `ERR_SEA_ASSET_NOT_FOUND`
+
+A key was passed to single exectuable application APIs to identify an asset,
+but no match could be found.
+
 <a id="ERR_SERVER_ALREADY_LISTEN"></a>
 
 ### `ERR_SERVER_ALREADY_LISTEN`

doc/api/single-executable-applications.md

@@ -178,14 +178,52 @@ The configuration currently reads the following top-level fields:
   "output": "/path/to/write/the/generated/blob.blob",
   "disableExperimentalSEAWarning": true, // Default: false
   "useSnapshot": false,  // Default: false
-  "useCodeCache": true // Default: false
+  "useCodeCache": true, // Default: false
+  "assets": {  // Optional
+    "a.dat": "/path/to/a.dat",
+    "b.txt": "/path/to/b.txt"
+  }
 }
 ```
 
 If the paths are not absolute, Node.js will use the path relative to the
 current working directory. The version of the Node.js binary used to produce
 the blob must be the same as the one to which the blob will be injected.
 
+### Assets
+
+Users can include assets by adding a key-path dictionary to the configuration
+as the `assets` field. At build time, Node.js would read the assets from the
+specified paths and bundle them into the preparation blob. In the generated
+executable, users can retrieve the assets using the [`sea.getAsset()`][] and
+[`sea.getAssetAsBlob()`][] APIs.
+
+```json
+{
+  "main": "/path/to/bundled/script.js",
+  "output": "/path/to/write/the/generated/blob.blob",
+  "assets": {
+    "a.jpg": "/path/to/a.jpg",
+    "b.txt": "/path/to/b.txt"
+  }
+}
+```
+
+The single-executable application can access the assets as follows:
+
+```cjs
+const { getAsset } = require('node:sea');
+// Returns a copy of the data in an ArrayBuffer.
+const image = getAsset('a.jpg');
+// Returns a string decoded from the asset as UTF8.
+const text = getAsset('b.txt', 'utf8');
+// Returns a Blob containing the asset without copying.
+const blob = getAssetAsBlob('a.jpg');
+```
+
+See documentation of the [`sea.getAsset()`][] and [`sea.getAssetAsBlob()`][]
+APIs for more information.
+
 ### Startup snapshot support
 
 The `useSnapshot` field can be used to enable startup snapshot support. In this
@@ -229,11 +267,58 @@ execute the script, which would improve the startup performance.
 
 **Note:** `import()` does not work when `useCodeCache` is `true`.
 
-## Notes
+## In the injected main script
 
-### `require(id)` in the injected module is not file based
+### Single-executable application API
 
-`require()` in the injected module is not the same as the [`require()`][]
+The `node:sea` builtin allows interaction with the single-executable application
+from the JavaScript main script embedded into the executable.
+
+#### `sea.isSea()`
+
+<!-- YAML
+added: REPLACEME
+-->
+
+* Returns: {boolean} Whether this script is running inside a single-executable
+  application.
+
+### `sea.getAsset(key[, encoding])`
+
+<!-- YAML
+added: REPLACEME
+-->
+
+This method can be used to retrieve the assets configured to be bundled into the
+single-executable application at build time.
+
+* `key`  {string} the key for the asset in the dictionary specified by the
+  `assets` field in the single-executable applciation configuratoin.
+* `encoding` {undefined|string} If specified, the asset will be decoded as
+  a string. Any encoding supported by the `TextDecoder` is accpeted.
+  If unspecified, an `ArrayBuffer` containing a copy of the asset would be
+  returned instead.
+* Returns: {string|ArrayBuffer}
+
+### `sea.getAssetAsBlob(key[, options])`
+
+<!-- YAML
+added: REPLACEME
+-->
+
+Similar to [`sea.getAsset()`][], but returns the result in a [`Blob`][]
+which allows users to read the data from the embedded asset without
+copying it first.
+
+* `key`  {string} the key for the asset in the dictionary specified by the
+  `assets` field in the single-executable applciation configuratoin.
+* `options` {Object}
+  * `type` {string} An optional mime type for the blob.
+* Returns: {Blob}
+
+### `require(id)` in the injected main script is not file based
+
+`require()` in the injected main script is not the same as the [`require()`][]
 available to modules that are not injected. It also does not have any of the
 properties that non-injected [`require()`][] has except [`require.main`][]. It
 can only be used to load built-in modules. Attempting to load a module that can
@@ -250,15 +335,17 @@ const { createRequire } = require('node:module');
 require = createRequire(__filename);
 ```
 
-### `__filename` and `module.filename` in the injected module
+### `__filename` and `module.filename` in the injected main script
 
-The values of `__filename` and `module.filename` in the injected module are
-equal to [`process.execPath`][].
+The values of `__filename` and `module.filename` in the injected main script
+are equal to [`process.execPath`][].
 
-### `__dirname` in the injected module
+### `__dirname` in the injected main script
 
-The value of `__dirname` in the injected module is equal to the directory name
-of [`process.execPath`][].
+The value of `__dirname` in the injected main script is equal to the directory
+name of [`process.execPath`][].
+
+## Notes
 
 ### Single executable application creation process
 
@@ -298,9 +385,12 @@ to help us document them.
 [Mach-O]: https://en.wikipedia.org/wiki/Mach-O
 [PE]: https://en.wikipedia.org/wiki/Portable_Executable
 [Windows SDK]: https://developer.microsoft.com/en-us/windows/downloads/windows-sdk/
+[`Blob`]: https://developer.mozilla.org/en-US/docs/Web/API/Blob
 [`process.execPath`]: process.md#processexecpath
 [`require()`]: modules.md#requireid
 [`require.main`]: modules.md#accessing-the-main-module
+[`sea.getAsset()`]: #seagetassetkey-encoding
+[`sea.getAssetAsBlob()`]: #seagetassetasblobkey-options
 [`v8.startupSnapshot.setDeserializeMainFunction()`]: v8.md#v8startupsnapshotsetdeserializemainfunctioncallback-data
 [`v8.startupSnapshot` API]: v8.md#startup-snapshot-api
 [documentation about startup snapshot support in Node.js]: cli.md#--build-snapshot

lib/internal/bootstrap/realm.js

@@ -128,6 +128,7 @@ const legacyWrapperList = new SafeSet([
 // beginning with "internal/".
 // Modules that can only be imported via the node: scheme.
 const schemelessBlockList = new SafeSet([
+  'sea',
   'test',
   'test/reporters',
 ]);

lib/internal/errors.js

@@ -1601,6 +1601,8 @@ E('ERR_NETWORK_IMPORT_DISALLOWED',
   "import of '%s' by %s is not supported: %s", Error);
 E('ERR_NOT_BUILDING_SNAPSHOT',
   'Operation cannot be invoked when not building startup snapshot', Error);
+E('ERR_NOT_IN_SEA',
+  'Operation cannot be invoked when not in a single-executable application', Error);
 E('ERR_NOT_SUPPORTED_IN_SNAPSHOT', '%s is not supported in startup snapshot', Error);
 E('ERR_NO_CRYPTO',
   'Node.js is not compiled with OpenSSL crypto support', Error);
@@ -1681,6 +1683,8 @@ E('ERR_REQUIRE_ESM',
   }, Error);
 E('ERR_SCRIPT_EXECUTION_INTERRUPTED',
   'Script execution was interrupted by `SIGINT`', Error);
+E('ERR_SEA_ASSET_NOT_FOUND',
+  'Cannot find asset %s for the single executable application', Error);
 E('ERR_SERVER_ALREADY_LISTEN',
   'Listen method has been called more than once without closing.', Error);
 E('ERR_SERVER_NOT_RUNNING', 'Server is not running.', Error);

lib/sea.js

@@ -0,0 +1,47 @@
+'use strict';
+
+const { isSea, getAsset: getAssetInternal } = internalBinding('sea');
+const { TextDecoder } = require('internal/encoding');
+const { validateString } = require('internal/validators');
+const {
+  ERR_NOT_IN_SEA,
+  ERR_SEA_ASSET_NOT_FOUND,
+} = require('internal/errors').codes;
+const { Blob } = require('internal/blob');
+
+function getRawAsset(key) {
+  validateString(key, 'key');
+
+  if (!isSea()) {
+    throw new ERR_NOT_IN_SEA();
+  }
+
+  const asset = getAssetInternal(key);
+  if (asset === undefined) {
+    throw new ERR_SEA_ASSET_NOT_FOUND(key);
+  }
+  return asset;
+}
+
+function getAsset(key, encoding) {
+  if (encoding !== undefined) {
+    validateString(encoding, 'encoding');
+  }
+  const asset = getRawAsset(key);
+  if (encoding === undefined) {
+    return asset.slice();
+  }
+  const decoder = new TextDecoder(encoding);
+  return decoder.decode(asset);
+}
+
+function getAssetAsBlob(key, options) {
+  const asset = getRawAsset(key);
+  return new Blob([asset], options);
+}
+
+module.exports = {
+  isSea,
+  getAsset,
+  getAssetAsBlob,
+};

src/json_parser.cc

@@ -4,6 +4,7 @@
 #include "util-inl.h"
 
 namespace node {
+using v8::Array;
 using v8::Context;
 using v8::Isolate;
 using v8::Local;
@@ -101,4 +102,50 @@ std::optional<bool> JSONParser::GetTopLevelBoolField(std::string_view field) {
   return value->BooleanValue(isolate);
 }
 
+std::optional<std::map<std::string, std::string>>
+JSONParser::GetTopLevelDictOfStrings(std::string_view field) {
+  Isolate* isolate = isolate_.get();
+  v8::HandleScope handle_scope(isolate);
+  Local<Context> context = context_.Get(isolate);
+  Local<Object> content_object = content_.Get(isolate);
+  Local<Value> value;
+  bool has_field;
+  // It's not a real script, so don't print the source line.
+  errors::PrinterTryCatch bootstrapCatch(
+      isolate, errors::PrinterTryCatch::kDontPrintSourceLine);
+  Local<Value> field_local;
+  if (!ToV8Value(context, field, isolate).ToLocal(&field_local)) {
+    return std::nullopt;
+  }
+  if (!content_object->Has(context, field_local).To(&has_field)) {
+    return std::nullopt;
+  }
+  if (!has_field) {
+    return std::map<std::string, std::string>{};
+  }
+  if (!content_object->Get(context, field_local).ToLocal(&value) ||
+      !value->IsObject()) {
+    return std::nullopt;
+  }
+  Local<Object> dict = value.As<Object>();
+  Local<Array> keys;
+  if (!dict->GetOwnPropertyNames(context).ToLocal(&keys)) {
+    return std::nullopt;
+  }
+  std::map<std::string, std::string> result;
+  uint32_t length = keys->Length();
+  for (uint32_t i = 0; i < length; ++i) {
+    Local<Value> key;
+    Local<Value> value;
+    if (!keys->Get(context, i).ToLocal(&key) || !key->IsString()) return {};
+    if (!dict->Get(context, key).ToLocal(&value) || !value->IsString())
+      return {};
+
+    Utf8Value key_utf8(isolate, key);
+    Utf8Value value_utf8(isolate, value);
+    result.emplace(*key_utf8, *value_utf8);
+  }
+  return result;
+}
+
 }  // namespace node

src/json_parser.h

@@ -3,6 +3,7 @@
 
 #if defined(NODE_WANT_INTERNALS) && NODE_WANT_INTERNALS
 
+#include <map>
 #include <memory>
 #include <optional>
 #include <string>
@@ -20,6 +21,8 @@ class JSONParser {
   bool Parse(const std::string& content);
   std::optional<std::string> GetTopLevelStringField(std::string_view field);
   std::optional<bool> GetTopLevelBoolField(std::string_view field);
+  std::optional<std::map<std::string, std::string>> GetTopLevelDictOfStrings(
+      std::string_view field);
 
  private:
   // We might want a lighter-weight JSON parser for this use case. But for now