Node.js Require source code coarse read

Author: Xiao Lei

Personal homepage: Github

Recently, I have been writing some related tools in Node.js. I don’t know much about how Node.js modules are loaded and the details of the module loading specification that they follow. This file is also seen node.js source code and part of the article summary:

Before the ES2015 standard, JS did not have a mature module system specification. Node.js addresses this shortcoming by adopting the module specification defined in the CommonJS specification, which includes:

1.require

Require is a function that takes a module identifier and refers to apis exposed by other modules.

2.module context

Module context specifies that there is a require variable in a module, which complies with the definition of the require function, an exports object, and that modules add properties to an exports object if they need to expose the API. And a Module Object.

3.module Identifiers

Module Identifiers define the parameters that require functions accept, such as strings that must be small camel names and may not have file suffixes. Or.. Indicates that the file path is relative and so on.

For details about the Module specification defined in commonJS, see the Wiki documentation

In our Node.js program, we use the require method that looks global (we’ll explain why it looks global later) to load other modules.

const util = require('./util')
Copy the code

First let’s take a look at how node.js defines this method internally:

Module.prototype.require = function () {
  assert(path, 'missing path');
  assert(typeof path === 'string'.'path must be a string');
  // It actually calls the module._load method
  return Module._load(path, this./* isMain */ false);
}

Module._load = function (request, parent, isMain) {...// Get the file name
  var filename = Module._resolveFilename(request, parent, isMain);

  // _cache Specifies the cache module
  var cachedModule = Module._cache[filename];
  if (cachedModule) {
    updateChildren(parent, cachedModule, true);
    return cachedModule.exports;
  }

  // For nativeModule
  if (NativeModule.nonInternalExists(filename)) {
    debug('load native module %s', request);
    return NativeModule.require(filename);
  }

  // Don't call updateChildren(), Module constructor already does.
  // Initialize a new module
  var module = new Module(filename, parent);

  if (isMain) {
    process.mainModule = module;
    module.id = '. ';
  }

  // Cache the module before loading it. Notice how node.js's module loading system avoids circular dependencies
  Module._cache[filename] = module;

  / / load the module
  tryModuleLoad(module, filename);

  Module. exports returns the exported contents
  return module.exports;
}
Copy the code

The module. _load method is an internal method that:

Find the module path based on the string you pass to represent the module path;
Cache based on found module path;
Then load the corresponding module.

Let’s look at how Node.js looks for the module based on the module path string passed in:

Module._resolveFilename = function (request, parent, isMain, options) {
  if (NativeModule.nonInternalExists(request)) {
    return request;
  }

  var paths;

  if (typeof options === 'object'&& options ! = =null &&
      Array.isArray(options.paths)) {
    ...
  } else {
    / / modules for roughly the path [parentDir] | [id, [parentDir]]
    paths = Module._resolveLookupPaths(request, parent, true);
  }

  // look up the filename first, since that's the cache key.
  // node index.js
  // request = index.js
  // paths = ['/root/foo/bar/index.js', '/root/foo/bar']
  var filename = Module._findPath(request, paths, isMain);
  if(! filename) {var err = new Error(`Cannot find module '${request}'`);
    err.code = 'MODULE_NOT_FOUND';
    throw err;
  }
  return filename;
}
Copy the code

Inside this method, we need to call an internal method: module. _resolveLookupPaths, which retrieves all possible paths for the Module based on the parent Module’s path:

Module._resolveLookupPaths = function (request, parent, newReturn) {... }Copy the code

Internally, this method handles the following cases:

Is the boot module, that is, throughnode xxxEnabled module

Node.js will retrieve your program execution path directly and return it in this method

require(xxx)Require an existence innode_modulesIn the module

All possible node_modules paths on the execution path are traversed

require(./)Require a relative path or absolute path module

Return directly to the parent path

When you have the desired path, call the module. _findPath method to find the corresponding file path.

Module._findPath = function (request, paths, isMain) {
  if (path.isAbsolute(request)) {
    paths = [' '];
  } else if(! paths || paths.length ===0) {
    return false;
  }

  // \x00 -> null, equivalent to an empty string
  var cacheKey = request + '\x00' +
                (paths.length === 1 ? paths[0] : paths.join('\x00'));
  // Path cache
  var entry = Module._pathCache[cacheKey];
  if (entry)
    return entry;

  var exts;
  // Is there a/on the end
  var trailingSlash = request.length > 0 &&
                      request.charCodeAt(request.length - 1) = = =47/ * / * /;

  // For each path
  for (var i = 0; i < paths.length; i++) {
    // Don't search further if path doesn't exist
    const curPath = paths[i];   // Current path
    if (curPath && stat(curPath) < 1) continue;
    var basePath = path.resolve(curPath, request);
    var filename;

    // Call internalModuleStat to determine the file type
    var rc = stat(basePath);
    // If the path does not end in /, it could be a file or a folder
    if(! trailingSlash) {if (rc === 0) {  / / File. The File
        if(preserveSymlinks && ! isMain) { filename = path.resolve(basePath); }else{ filename = toRealPath(basePath); }}else if (rc === 1) {  // Directory. If the provided path is a folder, this path will be used to find the module entry file corresponding to the main field in package.json
        if (exts === undefined)
          // '.js' '.json' '.node' '.ms'
          exts = Object.keys(Module._extensions);
        // Get the value of the main field inside the PKG
        filename = tryPackage(basePath, exts, isMain);
      }

      if(! filename) {// try it with each of the extensions
        if (exts === undefined)
          exts = Object.keys(Module._extensions);
        filename = tryExtensions(basePath, exts, isMain); / / ${basePath}. (js | json | node) and other documents suffixes, see if the file exists}}// If the path ends in /, it is a folder
    if(! filename && rc ===1) {  // Directory.
      if (exts === undefined)
        exts = Object.keys(Module._extensions);
      filename = tryPackage(basePath, exts, isMain) ||
        // try it with each of the extensions at "index"
        tryExtensions(path.resolve(basePath, 'index'), exts, isMain);
    }

    if (filename) {
      // Warn once if '.' resolved outside the module dir
      if (request === '. ' && i > 0) {
        if(! warned) { warned =true;
          process.emitWarning(
            'warning: require(\'.\') resolved outside the package ' +
            'directory. This functionality is deprecated and will be removed ' +
            'soon.'.'DeprecationWarning'.'DEP0019'); }}// Cache path
      Module._pathCache[cacheKey] = filename;
      returnfilename; }}return false;
}

function tryPackage(requestPath, exts, isMain) {
  var pkg = readPackage(requestPath); // Get the main field in package.json

  if(! pkg)return false;

  var filename = path.resolve(requestPath, pkg);  // Parse the path
  return tryFile(filename, isMain) ||             // Check whether the file exists directly
         tryExtensions(filename, exts, isMain) || // Check the existence of files ending with js,json,node, etc
         tryExtensions(path.resolve(filename, 'index'), exts, isMain);  / / judge the respectively by ${filename} / index (js | json | node) etc at the end of the file exists
}
Copy the code

Let’s review a strategy for querying modules above:

requireModule when the last character of the string passed is not/When:

If it is a file, return the path to the file
If it is a package.json file, and the path of the main field in this file (the corresponding method in the source code is tryPackage) :
- If the path corresponding to the main field is a file and exists, the path is returned
- The path corresponding to the main field does not have a suffix, so try using it.js..json..node..msSuffix to load the corresponding file
- If the preceding two conditions are not met, try the corresponding pathindex.js.index.json.index.nodefile
Js,.json,.node,.ms to load the corresponding file (tryExtensions in the source code).

requireModule, the last character of the string passed is/When, that is,requireIs a folder when:

First, search the path corresponding to the main field in package.json file under this folder, and the specific process method is the same as the search for package.json file mentioned above
Query the value ofindex.js.index.json.index.nodeAnd other documents

When the path to the file is found, we call tryModuleLoad to load the module. Inside this method, we actually call the load method of the module instance:

Module.prototype.load = function () {... this.filename = filename;// Define paths for the Module. Gets all possible node_modules paths on the module path
  this.paths = Module._nodeModulePaths(path.dirname(filename));

  var extension = path.extname(filename) || '.js';
  if(! Module._extensions[extension]) extension ='.js';
  // Start loading the file
  Module._extensions[extension](this, filename);
  this.loaded = true; . }Copy the code

Call the module. _extension method to load files in different formats, such as js files:

Module._extensions['.js'] = function(module, filename) {
  // First read the text content of the file
  var content = fs.readFileSync(filename, 'utf8');
  module._compile(internalModule.stripBOM(content), filename);
};

Copy the code

Internally we call the module.prototype. _compile method:

Module.prototype._compile = function (content, filename)) {
  content = internalModule.stripShebang(content);

  // create wrapper function
  // Wrap a layer around the source text
  var wrapper = Module.wrap(content);

  // vm.runInThisContext executes the wrapper code inside a V8 virtual machine
  var compiledWrapper = vm.runInThisContext(wrapper, {
    filename: filename,
    lineOffset: 0.displayErrors: true
  });

  var inspectorWrapper = null;
  if (process._breakFirstLine && process._eval == null) {
    if(! resolvedArgv) {// we enter the repl if we're not given a filename argument.
      if (process.argv[1]) {
        resolvedArgv = Module._resolveFilename(process.argv[1].null.false);
      } else {
        resolvedArgv = 'repl'; }}// Set breakpoint on module start
    if (filename === resolvedArgv) {
      delete process._breakFirstLine;
      inspectorWrapper = process.binding('inspector').callAndPauseOnStart; }}var dirname = path.dirname(filename);
  // construct the require function
  var require = internalModule.makeRequireFunction(this);
  var depth = internalModule.requireDepth;
  if (depth === 0) stat.cache = new Map(a);var result;
  if (inspectorWrapper) {
    result = inspectorWrapper(compiledWrapper, this.exports, this.exports,
                              require.this, filename, dirname);
  } else {
    // Start the function
    Module. Exports/require/module/filename/dirname
    result = compiledWrapper.call(this.exports, this.exports, require.this,
                                  filename, dirname);
  }
  if (depth === 0) stat.cache = null;
  return result;
}

Module.wrap = function(script) {
  return Module.wrapper[0] + script + Module.wrapper[1];
};

Module.wrapper = [
  '(function (exports, require, module, __filename, __dirname) { '.'\n}); '
];
Copy the code

throughModule.wrapWrap the source code in a layer (followcommonJSSpecification)
By calling thevmThe V8 virtual machine exposes the method to construct a new function
Complete the function call

Module.wrapper passes five arguments to the source text:

exports

Is a reference to the exports property of the third argument, Module

require

The require is not Module. The prototype. The require method, but rather through internalModule. MakeRequireFunction restructuring, This method internally relies on the module.prototype. require method to load modules, and also extends the require method.

module

Module objects, which need to expose apis for other modules to use, need to be defined on the module.exports property

__filename

The absolute path to the current file

__dirname

The absolute path to the parent folder of the current file

A few questions

Exports and Module. exports

Pay special attention to the connection between the first argument and the third argument: the first argument is a reference to the exports property of the third argument. Once a module exports is assigned to a new object, it breaks the reference between the module.exports property and the API exposed through the current module exports. A reference to a module always gets the module.exports property.

A circular reference

Official example:

a.js

console.log('a beginning');
exports.done = false;
const b = require('./b.js');
console.log('在 a 中，b.done = %j', b.done);
exports.done = true;
console.log(End of the 'a');
Copy the code

b.js

console.log('b start');
exports.done = false;
const a = require('./a.js');
console.log('in b, a. tone = %j', a.done);
exports.done = true;
console.log(End of the 'b');
Copy the code

main.js

console.log('the main start');
const a = require('./a.js');
const b = require('./b.js');
console.log('In main, a.tone =%j, b.tone =%j', a.done, b.done);
Copy the code

$node main.js main start A start B Start IN B, a.dot =falseB ends in A, b. tone =trueA ends in main, a. Tone =true, b.d one =true
Copy the code

When Module A is loaded, Module B needs to be loaded, but Module A is cached before it is actually loaded. For details, see module._load:

Module._cache[filename] = module;

tryModuleLoad(module, filename);
Copy the code

When module A is loaded again during module B loading, the API exported by module A is directly fetched from the cache. At this time, the property of exports.done is still false and not set to true.

Node.js Require source code coarse read

A few questions

Exports and Module. exports

A circular reference

Related Posts

I don’t know the deep copy and shallow copy of JS

Node Events (EventEmitter)

Read Chapter 1 of angular source code: Opening and platformBrowserDynamic