| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967 |
- 'use strict';
- var debug = require('debug')('urllib');
- var http = require('http');
- var https = require('https');
- var urlutil = require('url');
- var util = require('util');
- var qs = require('qs');
- var querystring = require('querystring');
- var zlib = require('zlib');
- var ua = require('default-user-agent');
- var digestAuthHeader = require('digest-header');
- var ms = require('humanize-ms');
- var statuses = require('statuses');
- var contentTypeParser = require('content-type');
- var first = require('ee-first');
- var detectProxyAgent = require('./detect_proxy_agent');
- var _Promise;
- var _iconv;
- var pkg = require('../package.json');
- var USER_AGENT = exports.USER_AGENT = ua('node-urllib', pkg.version);
- // change Agent.maxSockets to 1000
- exports.agent = new http.Agent();
- exports.agent.maxSockets = 1000;
- exports.httpsAgent = new https.Agent();
- exports.httpsAgent.maxSockets = 1000;
- /**
- * The default request timeout(in milliseconds).
- * @type {Number}
- * @const
- */
- exports.TIMEOUT = ms('5s');
- exports.TIMEOUTS = [ms('5s'), ms('5s')];
- var REQUEST_ID = 0;
- var MAX_VALUE = Math.pow(2, 31) - 10;
- var isNode010 = /^v0\.10\.\d+$/.test(process.version);
- var isNode012 = /^v0\.12\.\d+$/.test(process.version);
- /**
- * support data types
- * will auto decode response body
- * @type {Array}
- */
- var TEXT_DATA_TYPES = [
- 'json',
- 'text'
- ];
- var PROTO_RE = /^https?:\/\//i;
- /**
- * Handle all http request, both http and https support well.
- *
- * @example
- *
- * // GET http://httptest.cnodejs.net
- * urllib.request('http://httptest.cnodejs.net/test/get', function(err, data, res) {});
- * // POST http://httptest.cnodejs.net
- * var args = { type: 'post', data: { foo: 'bar' } };
- * urllib.request('http://httptest.cnodejs.net/test/post', args, function(err, data, res) {});
- *
- * @param {String|Object} url
- * @param {Object} [args], optional
- * - {Object} [data]: request data, will auto be query stringify.
- * - {Boolean} [dataAsQueryString]: force convert `data` to query string.
- * - {String|Buffer} [content]: optional, if set content, `data` will ignore.
- * - {ReadStream} [stream]: read stream to sent.
- * - {WriteStream} [writeStream]: writable stream to save response data.
- * If you use this, callback's data should be null.
- * We will just `pipe(ws, {end: true})`.
- * - {consumeWriteStream} [true]: consume the writeStream, invoke the callback after writeStream close.
- * - {String} [method]: optional, could be GET | POST | DELETE | PUT, default is GET
- * - {String} [contentType]: optional, request data type, could be `json`, default is undefined
- * - {String} [dataType]: optional, response data type, could be `text` or `json`, default is buffer
- * - {Boolean|Function} [fixJSONCtlChars]: optional, fix the control characters (U+0000 through U+001F)
- * before JSON parse response. Default is `false`.
- * `fixJSONCtlChars` can be a function, will pass data to the first argument. e.g.: `data = fixJSONCtlChars(data)`
- * - {Object} [headers]: optional, request headers
- * - {Number|Array} [timeout]: request timeout(in milliseconds), default is `exports.TIMEOUTS containing connect timeout and response timeout`
- * - {Agent} [agent]: optional, http agent. Set `false` if you does not use agent.
- * - {Agent} [httpsAgent]: optional, https agent. Set `false` if you does not use agent.
- * - {String} [auth]: Basic authentication i.e. 'user:password' to compute an Authorization header.
- * - {String} [digestAuth]: Digest authentication i.e. 'user:password' to compute an Authorization header.
- * - {String|Buffer|Array} [ca]: An array of strings or Buffers of trusted certificates.
- * If this is omitted several well known "root" CAs will be used, like VeriSign.
- * These are used to authorize connections.
- * Notes: This is necessary only if the server uses the self-signed certificate
- * - {Boolean} [rejectUnauthorized]: If true, the server certificate is verified against the list of supplied CAs.
- * An 'error' event is emitted if verification fails. Default: true.
- * - {String|Buffer} [pfx]: A string or Buffer containing the private key,
- * certificate and CA certs of the server in PFX or PKCS12 format.
- * - {String|Buffer} [key]: A string or Buffer containing the private key of the client in PEM format.
- * Notes: This is necessary only if using the client certificate authentication
- * - {String|Buffer} [cert]: A string or Buffer containing the certificate key of the client in PEM format.
- * Notes: This is necessary only if using the client certificate authentication
- * - {String} [passphrase]: A string of passphrase for the private key or pfx.
- * - {String} [ciphers]: A string describing the ciphers to use or exclude.
- * - {String} [secureProtocol]: The SSL method to use, e.g. SSLv3_method to force SSL version 3.
- * The possible values depend on your installation of OpenSSL and are defined in the constant SSL_METHODS.
- * - {Boolean} [followRedirect]: Follow HTTP 3xx responses as redirects. defaults to false.
- * - {Number} [maxRedirects]: The maximum number of redirects to follow, defaults to 10.
- * - {Function(from, to)} [formatRedirectUrl]: Format the redirect url by your self. Default is `url.resolve(from, to)`
- * - {Function(options)} [beforeRequest]: Before request hook, you can change every thing here.
- * - {Boolean} [streaming]: let you get the res object when request connected, default is `false`. alias `customResponse`
- * - {Boolean} [gzip]: Accept gzip response content and auto decode it, default is `false`.
- * - {Boolean} [timing]: Enable timing or not, default is `false`.
- * - {Function} [lookup]: Custom DNS lookup function, default is `dns.lookup`.
- * Require node >= 4.0.0 and only work on `http` protocol.
- * - {Boolean} [enableProxy]: optional, enable proxy request. Default is `false`.
- * - {String|Object} [proxy]: optional proxy agent uri or options. Default is `null`.
- * @param {Function} [callback]: callback(error, data, res). If missing callback, will return a promise object.
- * @return {HttpRequest} req object.
- * @api public
- */
- exports.request = function request(url, args, callback) {
- // request(url, callback)
- if (arguments.length === 2 && typeof args === 'function') {
- callback = args;
- args = null;
- }
- if (typeof callback === 'function') {
- return exports.requestWithCallback(url, args, callback);
- }
- // Promise
- if (!_Promise) {
- _Promise = require('any-promise');
- }
- return new _Promise(function (resolve, reject) {
- exports.requestWithCallback(url, args, makeCallback(resolve, reject));
- });
- };
- // alias to curl
- exports.curl = exports.request;
- function makeCallback(resolve, reject) {
- return function (err, data, res) {
- if (err) {
- return reject(err);
- }
- resolve({
- data: data,
- status: res.statusCode,
- headers: res.headers,
- res: res
- });
- };
- }
- // yield urllib.requestThunk(url, args)
- exports.requestThunk = function requestThunk(url, args) {
- return function (callback) {
- exports.requestWithCallback(url, args, function (err, data, res) {
- if (err) {
- return callback(err);
- }
- callback(null, {
- data: data,
- status: res.statusCode,
- headers: res.headers,
- res: res
- });
- });
- };
- };
- exports.requestWithCallback = function requestWithCallback(url, args, callback) {
- // requestWithCallback(url, callback)
- if (!url || (typeof url !== 'string' && typeof url !== 'object')) {
- var msg = util.format('expect request url to be a string or a http request options, but got %j', url);
- throw new Error(msg);
- }
- if (arguments.length === 2 && typeof args === 'function') {
- callback = args;
- args = null;
- }
- args = args || {};
- if (REQUEST_ID >= MAX_VALUE) {
- REQUEST_ID = 0;
- }
- var reqId = ++REQUEST_ID;
- args.requestUrls = args.requestUrls || [];
- args.timeout = args.timeout || exports.TIMEOUTS;
- args.maxRedirects = args.maxRedirects || 10;
- args.streaming = args.streaming || args.customResponse;
- var requestStartTime = Date.now();
- var parsedUrl;
- if (typeof url === 'string') {
- if (!PROTO_RE.test(url)) {
- // Support `request('www.server.com')`
- url = 'http://' + url;
- }
- parsedUrl = urlutil.parse(url);
- } else {
- parsedUrl = url;
- }
- var reqMeta = {
- requestId: reqId,
- url: parsedUrl.href,
- args: args,
- ctx: args.ctx,
- };
- if (args.emitter) {
- args.emitter.emit('request', reqMeta);
- }
- var method = (args.type || args.method || parsedUrl.method || 'GET').toUpperCase();
- var port = parsedUrl.port || 80;
- var httplib = http;
- var agent = getAgent(args.agent, exports.agent);
- var fixJSONCtlChars = args.fixJSONCtlChars;
- if (parsedUrl.protocol === 'https:') {
- httplib = https;
- agent = getAgent(args.httpsAgent, exports.httpsAgent);
- if (!parsedUrl.port) {
- port = 443;
- }
- }
- // request through proxy tunnel
- var proxyTunnelAgent = detectProxyAgent(parsedUrl, args);
- if (proxyTunnelAgent) {
- agent = proxyTunnelAgent;
- }
- var options = {
- host: parsedUrl.hostname || parsedUrl.host || 'localhost',
- path: parsedUrl.path || '/',
- method: method,
- port: port,
- agent: agent,
- headers: {},
- // default is dns.lookup
- // https://github.com/nodejs/node/blob/master/lib/net.js#L986
- // custom dnslookup require node >= 4.0.0
- // https://github.com/nodejs/node/blob/archived-io.js-v0.12/lib/net.js#L952
- lookup: args.lookup,
- };
- if (args.headers) {
- for (var k in args.headers) {
- options.headers[k] = args.headers[k];
- }
- }
- var sslNames = [
- 'pfx',
- 'key',
- 'passphrase',
- 'cert',
- 'ca',
- 'ciphers',
- 'rejectUnauthorized',
- 'secureProtocol',
- 'secureOptions',
- ];
- for (var i = 0; i < sslNames.length; i++) {
- var name = sslNames[i];
- if (args.hasOwnProperty(name)) {
- options[name] = args[name];
- }
- }
- // don't check ssl
- if (options.rejectUnauthorized === false && !options.hasOwnProperty('secureOptions')) {
- options.secureOptions = require('constants').SSL_OP_NO_TLSv1_2;
- }
- var auth = args.auth || parsedUrl.auth;
- if (auth) {
- options.auth = auth;
- }
- var body = args.content || args.data;
- var dataAsQueryString = method === 'GET' || method === 'HEAD' || args.dataAsQueryString;
- if (!args.content) {
- if (body && !(typeof body === 'string' || Buffer.isBuffer(body))) {
- if (dataAsQueryString) {
- // read: GET, HEAD, use query string
- body = args.nestedQuerystring ? qs.stringify(body) : querystring.stringify(body);
- } else {
- var contentType = options.headers['Content-Type'] || options.headers['content-type'];
- // auto add application/x-www-form-urlencoded when using urlencode form request
- if (!contentType) {
- if (args.contentType === 'json') {
- contentType = 'application/json';
- } else {
- contentType = 'application/x-www-form-urlencoded';
- }
- options.headers['Content-Type'] = contentType;
- }
- if (parseContentType(contentType).type === 'application/json') {
- body = JSON.stringify(body);
- } else {
- // 'application/x-www-form-urlencoded'
- body = args.nestedQuerystring ? qs.stringify(body) : querystring.stringify(body);
- }
- }
- }
- }
- // if it's a GET or HEAD request, data should be sent as query string
- if (dataAsQueryString && body) {
- options.path += (parsedUrl.query ? '&' : '?') + body;
- body = null;
- }
- var requestSize = 0;
- if (body) {
- var length = body.length;
- if (!Buffer.isBuffer(body)) {
- length = Buffer.byteLength(body);
- }
- requestSize = options.headers['Content-Length'] = length;
- }
- if (args.dataType === 'json') {
- options.headers.Accept = 'application/json';
- }
- if (typeof args.beforeRequest === 'function') {
- // you can use this hook to change every thing.
- args.beforeRequest(options);
- }
- var connectTimer = null;
- var responseTimer = null;
- var __err = null;
- var connected = false; // socket connected or not
- var keepAliveSocket = false; // request with keepalive socket
- var responseSize = 0;
- var statusCode = -1;
- var responseAborted = false;
- var remoteAddress = '';
- var remotePort = '';
- var timing = null;
- if (args.timing) {
- timing = {
- // socket assigned
- queuing: 0,
- // dns lookup time
- dnslookup: 0,
- // socket connected
- connected: 0,
- // request sent
- requestSent: 0,
- // Time to first byte (TTFB)
- waiting: 0,
- contentDownload: 0,
- };
- }
- function cancelConnectTimer() {
- if (connectTimer) {
- clearTimeout(connectTimer);
- connectTimer = null;
- }
- }
- function cancelResponseTimer() {
- if (responseTimer) {
- clearTimeout(responseTimer);
- responseTimer = null;
- }
- }
- function done(err, data, res) {
- cancelResponseTimer();
- if (!callback) {
- console.warn('[urllib:warn] [%s] [%s] [worker:%s] %s %s callback twice!!!',
- Date(), reqId, process.pid, options.method, url);
- // https://github.com/node-modules/urllib/pull/30
- if (err) {
- console.warn('[urllib:warn] [%s] [%s] [worker:%s] %s: %s\nstack: %s',
- Date(), reqId, process.pid, err.name, err.message, err.stack);
- }
- return;
- }
- var cb = callback;
- callback = null;
- var headers = {};
- if (res) {
- statusCode = res.statusCode;
- headers = res.headers;
- }
- // handle digest auth
- if (statusCode === 401 && headers['www-authenticate']
- && !options.headers.Authorization && args.digestAuth) {
- var authenticate = headers['www-authenticate'];
- if (authenticate.indexOf('Digest ') >= 0) {
- debug('Request#%d %s: got digest auth header WWW-Authenticate: %s', reqId, url, authenticate);
- options.headers.Authorization = digestAuthHeader(options.method, options.path, authenticate, args.digestAuth);
- debug('Request#%d %s: auth with digest header: %s', reqId, url, options.headers.Authorization);
- if (res.headers['set-cookie']) {
- options.headers.Cookie = res.headers['set-cookie'].join(';');
- }
- args.headers = options.headers;
- return exports.requestWithCallback(url, args, cb);
- }
- }
- var requestUseTime = Date.now() - requestStartTime;
- if (timing) {
- timing.contentDownload = requestUseTime;
- }
- debug('[%sms] done, %s bytes HTTP %s %s %s %s, keepAliveSocket: %s, timing: %j',
- requestUseTime, responseSize, statusCode, options.method, options.host, options.path,
- keepAliveSocket, timing);
- var response = {
- status: statusCode,
- statusCode: statusCode,
- headers: headers,
- size: responseSize,
- aborted: responseAborted,
- rt: requestUseTime,
- keepAliveSocket: keepAliveSocket,
- data: data,
- requestUrls: args.requestUrls,
- timing: timing,
- remoteAddress: remoteAddress,
- remotePort: remotePort,
- };
- if (err) {
- var agentStatus = '';
- if (agent && typeof agent.getCurrentStatus === 'function') {
- // add current agent status to error message for logging and debug
- agentStatus = ', agent status: ' + JSON.stringify(agent.getCurrentStatus());
- }
- err.message += ', ' + options.method + ' ' + url + ' ' + statusCode
- + ' (connected: ' + connected + ', keepalive socket: ' + keepAliveSocket + agentStatus + ')'
- + '\nheaders: ' + JSON.stringify(headers);
- err.data = data;
- err.path = options.path;
- err.status = statusCode;
- err.headers = headers;
- err.res = response;
- }
- cb(err, data, args.streaming ? res : response);
- if (args.emitter) {
- // keep to use the same reqMeta object on request event before
- reqMeta.url = parsedUrl.href;
- reqMeta.socket = req && req.connection;
- reqMeta.options = options;
- reqMeta.size = requestSize;
- args.emitter.emit('response', {
- requestId: reqId,
- error: err,
- ctx: args.ctx,
- req: reqMeta,
- res: response,
- });
- }
- }
- function handleRedirect(res) {
- var err = null;
- if (args.followRedirect && statuses.redirect[res.statusCode]) { // handle redirect
- args._followRedirectCount = (args._followRedirectCount || 0) + 1;
- var location = res.headers.location;
- if (!location) {
- err = new Error('Got statusCode ' + res.statusCode + ' but cannot resolve next location from headers');
- err.name = 'FollowRedirectError';
- } else if (args._followRedirectCount > args.maxRedirects) {
- err = new Error('Exceeded maxRedirects. Probably stuck in a redirect loop ' + url);
- err.name = 'MaxRedirectError';
- } else {
- var newUrl = args.formatRedirectUrl ? args.formatRedirectUrl(url, location) : urlutil.resolve(url, location);
- debug('Request#%d %s: `redirected` from %s to %s', reqId, options.path, url, newUrl);
- // make sure timer stop
- cancelResponseTimer();
- // should clean up headers.Host on `location: http://other-domain/url`
- if (options.headers.Host && PROTO_RE.test(location)) {
- options.headers.Host = null;
- args.headers = options.headers;
- }
- // avoid done will be execute in the future change.
- var cb = callback;
- callback = null;
- exports.requestWithCallback(newUrl, args, cb);
- return {
- redirect: true,
- error: null
- };
- }
- }
- return {
- redirect: false,
- error: err
- };
- }
- // set user-agent
- if (!options.headers['User-Agent'] && !options.headers['user-agent']) {
- options.headers['User-Agent'] = USER_AGENT;
- }
- if (args.gzip) {
- if (!options.headers['Accept-Encoding'] && !options.headers['accept-encoding']) {
- options.headers['Accept-Encoding'] = 'gzip';
- }
- }
- function decodeContent(res, body, cb) {
- var encoding = res.headers['content-encoding'];
- if (body.length === 0) {
- return cb(null, body, encoding);
- }
- if (!encoding || encoding.toLowerCase() !== 'gzip') {
- return cb(null, body, encoding);
- }
- debug('gunzip %d length body', body.length);
- zlib.gunzip(body, cb);
- }
- var writeStream = args.writeStream;
- debug('Request#%d %s %s with headers %j, options.path: %s',
- reqId, method, url, options.headers, options.path);
- args.requestUrls.push(parsedUrl.href);
- function onResponse(res) {
- if (timing) {
- timing.waiting = Date.now() - requestStartTime;
- }
- debug('Request#%d %s `req response` event emit: status %d, headers: %j',
- reqId, url, res.statusCode, res.headers);
- if (args.streaming) {
- var result = handleRedirect(res);
- if (result.redirect) {
- res.resume();
- return;
- }
- if (result.error) {
- res.resume();
- return done(result.error, null, res);
- }
- return done(null, null, res);
- }
- res.on('close', function () {
- debug('Request#%d %s: `res close` event emit, total size %d',
- reqId, url, responseSize);
- });
- res.on('error', function () {
- debug('Request#%d %s: `res error` event emit, total size %d',
- reqId, url, responseSize);
- });
- res.on('aborted', function () {
- responseAborted = true;
- debug('Request#%d %s: `res aborted` event emit, total size %d',
- reqId, url, responseSize);
- });
- if (writeStream) {
- // If there's a writable stream to recieve the response data, just pipe the
- // response stream to that writable stream and call the callback when it has
- // finished writing.
- //
- // NOTE that when the response stream `res` emits an 'end' event it just
- // means that it has finished piping data to another stream. In the
- // meanwhile that writable stream may still writing data to the disk until
- // it emits a 'close' event.
- //
- // That means that we should not apply callback until the 'close' of the
- // writable stream is emited.
- //
- // See also:
- // - https://github.com/TBEDP/urllib/commit/959ac3365821e0e028c231a5e8efca6af410eabb
- // - http://nodejs.org/api/stream.html#stream_event_end
- // - http://nodejs.org/api/stream.html#stream_event_close_1
- var result = handleRedirect(res);
- if (result.redirect) {
- res.resume();
- return;
- }
- if (result.error) {
- res.resume();
- // end ths stream first
- writeStream.end();
- return done(result.error, null, res);
- }
- // you can set consumeWriteStream false that only wait response end
- if (args.consumeWriteStream === false) {
- res.on('end', done.bind(null, null, null, res));
- } else {
- // node 0.10, 0.12: only emit res aborted, writeStream close not fired
- if (isNode010 || isNode012) {
- first([
- [ writeStream, 'close' ],
- [ res, 'aborted' ],
- ], function(_, stream, event) {
- debug('Request#%d %s: writeStream or res %s event emitted', reqId, url, event);
- done(__err || null, null, res);
- });
- } else {
- writeStream.on('close', function() {
- debug('Request#%d %s: writeStream close event emitted', reqId, url);
- done(__err || null, null, res);
- });
- }
- }
- return res.pipe(writeStream);
- }
- // Otherwise, just concat those buffers.
- //
- // NOTE that the `chunk` is not a String but a Buffer. It means that if
- // you simply concat two chunk with `+` you're actually converting both
- // Buffers into Strings before concating them. It'll cause problems when
- // dealing with multi-byte characters.
- //
- // The solution is to store each chunk in an array and concat them with
- // 'buffer-concat' when all chunks is recieved.
- //
- // See also:
- // http://cnodejs.org/topic/4faf65852e8fb5bc65113403
- var chunks = [];
- res.on('data', function (chunk) {
- debug('Request#%d %s: `res data` event emit, size %d', reqId, url, chunk.length);
- responseSize += chunk.length;
- chunks.push(chunk);
- });
- res.on('end', function () {
- var body = Buffer.concat(chunks, responseSize);
- debug('Request#%d %s: `res end` event emit, total size %d, _dumped: %s',
- reqId, url, responseSize, res._dumped);
- if (__err) {
- // req.abort() after `res data` event emit.
- return done(__err, body, res);
- }
- var result = handleRedirect(res);
- if (result.error) {
- return done(result.error, body, res);
- }
- if (result.redirect) {
- return;
- }
- decodeContent(res, body, function (err, data, encoding) {
- if (err) {
- return done(err, body, res);
- }
- // if body not decode, dont touch it
- if (!encoding && TEXT_DATA_TYPES.indexOf(args.dataType) >= 0) {
- // try to decode charset
- try {
- data = decodeBodyByCharset(data, res);
- } catch (e) {
- debug('decodeBodyByCharset error: %s', e);
- // if error, dont touch it
- return done(null, data, res);
- }
- if (args.dataType === 'json') {
- if (responseSize === 0) {
- data = null;
- } else {
- var r = parseJSON(data, fixJSONCtlChars);
- if (r.error) {
- err = r.error;
- } else {
- data = r.data;
- }
- }
- }
- }
- if (responseAborted) {
- // err = new Error('Remote socket was terminated before `response.end()` was called');
- // err.name = 'RemoteSocketClosedError';
- debug('Request#%d %s: Remote socket was terminated before `response.end()` was called', reqId, url);
- }
- done(err, data, res);
- });
- });
- }
- var connectTimeout, responseTimeout;
- if (Array.isArray(args.timeout)) {
- connectTimeout = ms(args.timeout[0]);
- responseTimeout = ms(args.timeout[1]);
- } else { // set both timeout equal
- connectTimeout = responseTimeout = ms(args.timeout);
- }
- debug('ConnectTimeout: %d, ResponseTimeout: %d', connectTimeout, responseTimeout);
- function startConnectTimer() {
- debug('Connect timer ticking, timeout: %d', connectTimeout);
- connectTimer = setTimeout(function () {
- connectTimer = null;
- if (statusCode === -1) {
- statusCode = -2;
- }
- var msg = 'Connect timeout for ' + connectTimeout + 'ms';
- var errorName = 'ConnectionTimeoutError';
- if (!req.socket) {
- errorName = 'SocketAssignTimeoutError';
- msg += ', working sockets is full';
- }
- __err = new Error(msg);
- __err.name = errorName;
- __err.requestId = reqId;
- debug('ConnectTimeout: Request#%d %s %s: %s, connected: %s', reqId, url, __err.name, msg, connected);
- abortRequest();
- }, connectTimeout);
- }
- function startResposneTimer() {
- debug('Response timer ticking, timeout: %d', responseTimeout);
- responseTimer = setTimeout(function () {
- responseTimer = null;
- var msg = 'Response timeout for ' + responseTimeout + 'ms';
- var errorName = 'ResponseTimeoutError';
- __err = new Error(msg);
- __err.name = errorName;
- __err.requestId = reqId;
- debug('ResponseTimeout: Request#%d %s %s: %s, connected: %s', reqId, url, __err.name, msg, connected);
- abortRequest();
- }, responseTimeout);
- }
- var req;
- // request headers checker will throw error
- try {
- req = httplib.request(options, onResponse);
- } catch (err) {
- return done(err);
- }
- // environment detection: browser or nodejs
- if (typeof(window) === 'undefined') {
- // start connect timer just after `request` return, and just in nodejs environment
- startConnectTimer();
- }
- function abortRequest() {
- debug('Request#%d %s abort, connected: %s', reqId, url, connected);
- // it wont case error event when req haven't been assigned a socket yet.
- if (!req.socket) {
- __err.noSocket = true;
- done(__err);
- }
- req.abort();
- }
- if (timing) {
- // request sent
- req.on('finish', function() {
- timing.requestSent = Date.now() - requestStartTime;
- });
- }
- req.once('socket', function (socket) {
- if (timing) {
- // socket queuing time
- timing.queuing = Date.now() - requestStartTime;
- }
- // https://github.com/nodejs/node/blob/master/lib/net.js#L377
- // https://github.com/nodejs/node/blob/v0.10.40-release/lib/net.js#L352
- // should use socket.socket on 0.10.x
- if (isNode010 && socket.socket) {
- socket = socket.socket;
- }
- var readyState = socket.readyState;
- if (readyState === 'opening') {
- socket.once('lookup', function(err, ip, addressType) {
- debug('Request#%d %s lookup: %s, %s, %s', reqId, url, err, ip, addressType);
- if (timing) {
- timing.dnslookup = Date.now() - requestStartTime;
- }
- if (ip) {
- remoteAddress = ip;
- }
- });
- socket.once('connect', function() {
- if (timing) {
- // socket connected
- timing.connected = Date.now() - requestStartTime;
- }
- // cancel socket timer at first and start tick for TTFB
- cancelConnectTimer();
- startResposneTimer();
- debug('Request#%d %s new socket connected', reqId, url);
- connected = true;
- if (!remoteAddress) {
- remoteAddress = socket.remoteAddress;
- }
- remotePort = socket.remotePort;
- });
- return;
- }
- debug('Request#%d %s reuse socket connected, readyState: %s', reqId, url, readyState);
- connected = true;
- keepAliveSocket = true;
- if (!remoteAddress) {
- remoteAddress = socket.remoteAddress;
- }
- remotePort = socket.remotePort;
- // reuse socket, timer should be canceled.
- cancelConnectTimer();
- startResposneTimer();
- });
- req.on('error', function (err) {
- if (err.name === 'Error') {
- err.name = connected ? 'ResponseError' : 'RequestError';
- }
- err.message += ' (req "error")';
- debug('Request#%d %s `req error` event emit, %s: %s', reqId, url, err.name, err.message);
- done(__err || err);
- });
- if (writeStream) {
- writeStream.once('error', function (err) {
- err.message += ' (writeStream "error")';
- __err = err;
- debug('Request#%d %s `writeStream error` event emit, %s: %s', reqId, url, err.name, err.message);
- abortRequest();
- });
- }
- if (args.stream) {
- args.stream.pipe(req);
- args.stream.once('error', function (err) {
- err.message += ' (stream "error")';
- __err = err;
- debug('Request#%d %s `readStream error` event emit, %s: %s', reqId, url, err.name, err.message);
- abortRequest();
- });
- } else {
- req.end(body);
- }
- req.requestId = reqId;
- return req;
- };
- var JSONCtlCharsMap = {
- '"': '\\"', // \u0022
- '\\': '\\\\', // \u005c
- '\b': '\\b', // \u0008
- '\f': '\\f', // \u000c
- '\n': '\\n', // \u000a
- '\r': '\\r', // \u000d
- '\t': '\\t' // \u0009
- };
- var JSONCtlCharsRE = /[\u0000-\u001F\u005C]/g;
- function _replaceOneChar(c) {
- return JSONCtlCharsMap[c] || '\\u' + (c.charCodeAt(0) + 0x10000).toString(16).substr(1);
- }
- function replaceJSONCtlChars(str) {
- return str.replace(JSONCtlCharsRE, _replaceOneChar);
- }
- function parseJSON(data, fixJSONCtlChars) {
- var result = {
- error: null,
- data: null
- };
- if (fixJSONCtlChars) {
- if (typeof fixJSONCtlChars === 'function') {
- data = fixJSONCtlChars(data);
- } else {
- // https://github.com/node-modules/urllib/pull/77
- // remote the control characters (U+0000 through U+001F)
- data = replaceJSONCtlChars(data);
- }
- }
- try {
- result.data = JSON.parse(data);
- } catch (err) {
- if (err.name === 'SyntaxError') {
- err.name = 'JSONResponseFormatError';
- }
- if (data.length > 1024) {
- // show 0~512 ... -512~end data
- err.message += ' (data json format: ' +
- JSON.stringify(data.slice(0, 512)) + ' ...skip... ' + JSON.stringify(data.slice(data.length - 512)) + ')';
- } else {
- err.message += ' (data json format: ' + JSON.stringify(data) + ')';
- }
- result.error = err;
- }
- return result;
- }
- /**
- * decode response body by parse `content-type`'s charset
- * @param {Buffer} data
- * @param {Http(s)Response} res
- * @return {String}
- */
- function decodeBodyByCharset(data, res) {
- var type = res.headers['content-type'];
- if (!type) {
- return data.toString();
- }
- var type = parseContentType(type);
- var charset = type.parameters.charset || 'utf-8';
- if (!Buffer.isEncoding(charset)) {
- if (!_iconv) {
- _iconv = require('iconv-lite');
- }
- return _iconv.decode(data, charset);
- }
- return data.toString(charset);
- }
- function getAgent(agent, defaultAgent) {
- return agent === undefined ? defaultAgent : agent;
- }
- function parseContentType(str) {
- try {
- return contentTypeParser.parse(str);
- } catch (err) {
- // ignore content-type error, tread as default
- return { parameters: {} };
- }
- }
|