1 line
23 KiB
JSON
1 line
23 KiB
JSON
{"ast":null,"code":"// Copyright Joyent, Inc. and other Node contributors.\n//\n// Permission is hereby granted, free of charge, to any person obtaining a\n// copy of this software and associated documentation files (the\n// \"Software\"), to deal in the Software without restriction, including\n// without limitation the rights to use, copy, modify, merge, publish,\n// distribute, sublicense, and/or sell copies of the Software, and to permit\n// persons to whom the Software is furnished to do so, subject to the\n// following conditions:\n//\n// The above copyright notice and this permission notice shall be included\n// in all copies or substantial portions of the Software.\n//\n// THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS\n// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF\n// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN\n// NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,\n// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR\n// OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE\n// USE OR OTHER DEALINGS IN THE SOFTWARE.\n'use strict';\n\nvar punycode = require('punycode');\n\nvar util = require('./util');\n\nexports.parse = urlParse;\nexports.resolve = urlResolve;\nexports.resolveObject = urlResolveObject;\nexports.format = urlFormat;\nexports.Url = Url;\n\nfunction Url() {\n this.protocol = null;\n this.slashes = null;\n this.auth = null;\n this.host = null;\n this.port = null;\n this.hostname = null;\n this.hash = null;\n this.search = null;\n this.query = null;\n this.pathname = null;\n this.path = null;\n this.href = null;\n} // Reference: RFC 3986, RFC 1808, RFC 2396\n// define these here so at least they only have to be\n// compiled once on the first module load.\n\n\nvar protocolPattern = /^([a-z0-9.+-]+:)/i,\n portPattern = /:[0-9]*$/,\n // Special case for a simple path URL\nsimplePathPattern = /^(\\/\\/?(?!\\/)[^\\?\\s]*)(\\?[^\\s]*)?$/,\n // RFC 2396: characters reserved for delimiting URLs.\n// We actually just auto-escape these.\ndelims = ['<', '>', '\"', '`', ' ', '\\r', '\\n', '\\t'],\n // RFC 2396: characters not allowed for various reasons.\nunwise = ['{', '}', '|', '\\\\', '^', '`'].concat(delims),\n // Allowed by RFCs, but cause of XSS attacks. Always escape these.\nautoEscape = ['\\''].concat(unwise),\n // Characters that are never ever allowed in a hostname.\n// Note that any invalid chars are also handled, but these\n// are the ones that are *expected* to be seen, so we fast-path\n// them.\nnonHostChars = ['%', '/', '?', ';', '#'].concat(autoEscape),\n hostEndingChars = ['/', '?', '#'],\n hostnameMaxLen = 255,\n hostnamePartPattern = /^[+a-z0-9A-Z_-]{0,63}$/,\n hostnamePartStart = /^([+a-z0-9A-Z_-]{0,63})(.*)$/,\n // protocols that can allow \"unsafe\" and \"unwise\" chars.\nunsafeProtocol = {\n 'javascript': true,\n 'javascript:': true\n},\n // protocols that never have a hostname.\nhostlessProtocol = {\n 'javascript': true,\n 'javascript:': true\n},\n // protocols that always contain a // bit.\nslashedProtocol = {\n 'http': true,\n 'https': true,\n 'ftp': true,\n 'gopher': true,\n 'file': true,\n 'http:': true,\n 'https:': true,\n 'ftp:': true,\n 'gopher:': true,\n 'file:': true\n},\n querystring = require('querystring');\n\nfunction urlParse(url, parseQueryString, slashesDenoteHost) {\n if (url && util.isObject(url) && url instanceof Url) return url;\n var u = new Url();\n u.parse(url, parseQueryString, slashesDenoteHost);\n return u;\n}\n\nUrl.prototype.parse = function (url, parseQueryString, slashesDenoteHost) {\n if (!util.isString(url)) {\n throw new TypeError(\"Parameter 'url' must be a string, not \" + typeof url);\n } // Copy chrome, IE, opera backslash-handling behavior.\n // Back slashes before the query string get converted to forward slashes\n // See: https://code.google.com/p/chromium/issues/detail?id=25916\n\n\n var queryIndex = url.indexOf('?'),\n splitter = queryIndex !== -1 && queryIndex < url.indexOf('#') ? '?' : '#',\n uSplit = url.split(splitter),\n slashRegex = /\\\\/g;\n uSplit[0] = uSplit[0].replace(slashRegex, '/');\n url = uSplit.join(splitter);\n var rest = url; // trim before proceeding.\n // This is to support parse stuff like \" http://foo.com \\n\"\n\n rest = rest.trim();\n\n if (!slashesDenoteHost && url.split('#').length === 1) {\n // Try fast path regexp\n var simplePath = simplePathPattern.exec(rest);\n\n if (simplePath) {\n this.path = rest;\n this.href = rest;\n this.pathname = simplePath[1];\n\n if (simplePath[2]) {\n this.search = simplePath[2];\n\n if (parseQueryString) {\n this.query = querystring.parse(this.search.substr(1));\n } else {\n this.query = this.search.substr(1);\n }\n } else if (parseQueryString) {\n this.search = '';\n this.query = {};\n }\n\n return this;\n }\n }\n\n var proto = protocolPattern.exec(rest);\n\n if (proto) {\n proto = proto[0];\n var lowerProto = proto.toLowerCase();\n this.protocol = lowerProto;\n rest = rest.substr(proto.length);\n } // figure out if it's got a host\n // user@server is *always* interpreted as a hostname, and url\n // resolution will treat //foo/bar as host=foo,path=bar because that's\n // how the browser resolves relative URLs.\n\n\n if (slashesDenoteHost || proto || rest.match(/^\\/\\/[^@\\/]+@[^@\\/]+/)) {\n var slashes = rest.substr(0, 2) === '//';\n\n if (slashes && !(proto && hostlessProtocol[proto])) {\n rest = rest.substr(2);\n this.slashes = true;\n }\n }\n\n if (!hostlessProtocol[proto] && (slashes || proto && !slashedProtocol[proto])) {\n // there's a hostname.\n // the first instance of /, ?, ;, or # ends the host.\n //\n // If there is an @ in the hostname, then non-host chars *are* allowed\n // to the left of the last @ sign, unless some host-ending character\n // comes *before* the @-sign.\n // URLs are obnoxious.\n //\n // ex:\n // http://a@b@c/ => user:a@b host:c\n // http://a@b?@c => user:a host:c path:/?@c\n // v0.12 TODO(isaacs): This is not quite how Chrome does things.\n // Review our test case against browsers more comprehensively.\n // find the first instance of any hostEndingChars\n var hostEnd = -1;\n\n for (var i = 0; i < hostEndingChars.length; i++) {\n var hec = rest.indexOf(hostEndingChars[i]);\n if (hec !== -1 && (hostEnd === -1 || hec < hostEnd)) hostEnd = hec;\n } // at this point, either we have an explicit point where the\n // auth portion cannot go past, or the last @ char is the decider.\n\n\n var auth, atSign;\n\n if (hostEnd === -1) {\n // atSign can be anywhere.\n atSign = rest.lastIndexOf('@');\n } else {\n // atSign must be in auth portion.\n // http://a@b/c@d => host:b auth:a path:/c@d\n atSign = rest.lastIndexOf('@', hostEnd);\n } // Now we have a portion which is definitely the auth.\n // Pull that off.\n\n\n if (atSign !== -1) {\n auth = rest.slice(0, atSign);\n rest = rest.slice(atSign + 1);\n this.auth = decodeURIComponent(auth);\n } // the host is the remaining to the left of the first non-host char\n\n\n hostEnd = -1;\n\n for (var i = 0; i < nonHostChars.length; i++) {\n var hec = rest.indexOf(nonHostChars[i]);\n if (hec !== -1 && (hostEnd === -1 || hec < hostEnd)) hostEnd = hec;\n } // if we still have not hit it, then the entire thing is a host.\n\n\n if (hostEnd === -1) hostEnd = rest.length;\n this.host = rest.slice(0, hostEnd);\n rest = rest.slice(hostEnd); // pull out port.\n\n this.parseHost(); // we've indicated that there is a hostname,\n // so even if it's empty, it has to be present.\n\n this.hostname = this.hostname || ''; // if hostname begins with [ and ends with ]\n // assume that it's an IPv6 address.\n\n var ipv6Hostname = this.hostname[0] === '[' && this.hostname[this.hostname.length - 1] === ']'; // validate a little.\n\n if (!ipv6Hostname) {\n var hostparts = this.hostname.split(/\\./);\n\n for (var i = 0, l = hostparts.length; i < l; i++) {\n var part = hostparts[i];\n if (!part) continue;\n\n if (!part.match(hostnamePartPattern)) {\n var newpart = '';\n\n for (var j = 0, k = part.length; j < k; j++) {\n if (part.charCodeAt(j) > 127) {\n // we replace non-ASCII char with a temporary placeholder\n // we need this to make sure size of hostname is not\n // broken by replacing non-ASCII by nothing\n newpart += 'x';\n } else {\n newpart += part[j];\n }\n } // we test again with ASCII char only\n\n\n if (!newpart.match(hostnamePartPattern)) {\n var validParts = hostparts.slice(0, i);\n var notHost = hostparts.slice(i + 1);\n var bit = part.match(hostnamePartStart);\n\n if (bit) {\n validParts.push(bit[1]);\n notHost.unshift(bit[2]);\n }\n\n if (notHost.length) {\n rest = '/' + notHost.join('.') + rest;\n }\n\n this.hostname = validParts.join('.');\n break;\n }\n }\n }\n }\n\n if (this.hostname.length > hostnameMaxLen) {\n this.hostname = '';\n } else {\n // hostnames are always lower case.\n this.hostname = this.hostname.toLowerCase();\n }\n\n if (!ipv6Hostname) {\n // IDNA Support: Returns a punycoded representation of \"domain\".\n // It only converts parts of the domain name that\n // have non-ASCII characters, i.e. it doesn't matter if\n // you call it with a domain that already is ASCII-only.\n this.hostname = punycode.toASCII(this.hostname);\n }\n\n var p = this.port ? ':' + this.port : '';\n var h = this.hostname || '';\n this.host = h + p;\n this.href += this.host; // strip [ and ] from the hostname\n // the host field still retains them, though\n\n if (ipv6Hostname) {\n this.hostname = this.hostname.substr(1, this.hostname.length - 2);\n\n if (rest[0] !== '/') {\n rest = '/' + rest;\n }\n }\n } // now rest is set to the post-host stuff.\n // chop off any delim chars.\n\n\n if (!unsafeProtocol[lowerProto]) {\n // First, make 100% sure that any \"autoEscape\" chars get\n // escaped, even if encodeURIComponent doesn't think they\n // need to be.\n for (var i = 0, l = autoEscape.length; i < l; i++) {\n var ae = autoEscape[i];\n if (rest.indexOf(ae) === -1) continue;\n var esc = encodeURIComponent(ae);\n\n if (esc === ae) {\n esc = escape(ae);\n }\n\n rest = rest.split(ae).join(esc);\n }\n } // chop off from the tail first.\n\n\n var hash = rest.indexOf('#');\n\n if (hash !== -1) {\n // got a fragment string.\n this.hash = rest.substr(hash);\n rest = rest.slice(0, hash);\n }\n\n var qm = rest.indexOf('?');\n\n if (qm !== -1) {\n this.search = rest.substr(qm);\n this.query = rest.substr(qm + 1);\n\n if (parseQueryString) {\n this.query = querystring.parse(this.query);\n }\n\n rest = rest.slice(0, qm);\n } else if (parseQueryString) {\n // no query string, but parseQueryString still requested\n this.search = '';\n this.query = {};\n }\n\n if (rest) this.pathname = rest;\n\n if (slashedProtocol[lowerProto] && this.hostname && !this.pathname) {\n this.pathname = '/';\n } //to support http.request\n\n\n if (this.pathname || this.search) {\n var p = this.pathname || '';\n var s = this.search || '';\n this.path = p + s;\n } // finally, reconstruct the href based on what has been validated.\n\n\n this.href = this.format();\n return this;\n}; // format a parsed object into a url string\n\n\nfunction urlFormat(obj) {\n // ensure it's an object, and not a string url.\n // If it's an obj, this is a no-op.\n // this way, you can call url_format() on strings\n // to clean up potentially wonky urls.\n if (util.isString(obj)) obj = urlParse(obj);\n if (!(obj instanceof Url)) return Url.prototype.format.call(obj);\n return obj.format();\n}\n\nUrl.prototype.format = function () {\n var auth = this.auth || '';\n\n if (auth) {\n auth = encodeURIComponent(auth);\n auth = auth.replace(/%3A/i, ':');\n auth += '@';\n }\n\n var protocol = this.protocol || '',\n pathname = this.pathname || '',\n hash = this.hash || '',\n host = false,\n query = '';\n\n if (this.host) {\n host = auth + this.host;\n } else if (this.hostname) {\n host = auth + (this.hostname.indexOf(':') === -1 ? this.hostname : '[' + this.hostname + ']');\n\n if (this.port) {\n host += ':' + this.port;\n }\n }\n\n if (this.query && util.isObject(this.query) && Object.keys(this.query).length) {\n query = querystring.stringify(this.query);\n }\n\n var search = this.search || query && '?' + query || '';\n if (protocol && protocol.substr(-1) !== ':') protocol += ':'; // only the slashedProtocols get the //. Not mailto:, xmpp:, etc.\n // unless they had them to begin with.\n\n if (this.slashes || (!protocol || slashedProtocol[protocol]) && host !== false) {\n host = '//' + (host || '');\n if (pathname && pathname.charAt(0) !== '/') pathname = '/' + pathname;\n } else if (!host) {\n host = '';\n }\n\n if (hash && hash.charAt(0) !== '#') hash = '#' + hash;\n if (search && search.charAt(0) !== '?') search = '?' + search;\n pathname = pathname.replace(/[?#]/g, function (match) {\n return encodeURIComponent(match);\n });\n search = search.replace('#', '%23');\n return protocol + host + pathname + search + hash;\n};\n\nfunction urlResolve(source, relative) {\n return urlParse(source, false, true).resolve(relative);\n}\n\nUrl.prototype.resolve = function (relative) {\n return this.resolveObject(urlParse(relative, false, true)).format();\n};\n\nfunction urlResolveObject(source, relative) {\n if (!source) return relative;\n return urlParse(source, false, true).resolveObject(relative);\n}\n\nUrl.prototype.resolveObject = function (relative) {\n if (util.isString(relative)) {\n var rel = new Url();\n rel.parse(relative, false, true);\n relative = rel;\n }\n\n var result = new Url();\n var tkeys = Object.keys(this);\n\n for (var tk = 0; tk < tkeys.length; tk++) {\n var tkey = tkeys[tk];\n result[tkey] = this[tkey];\n } // hash is always overridden, no matter what.\n // even href=\"\" will remove it.\n\n\n result.hash = relative.hash; // if the relative url is empty, then there's nothing left to do here.\n\n if (relative.href === '') {\n result.href = result.format();\n return result;\n } // hrefs like //foo/bar always cut to the protocol.\n\n\n if (relative.slashes && !relative.protocol) {\n // take everything except the protocol from relative\n var rkeys = Object.keys(relative);\n\n for (var rk = 0; rk < rkeys.length; rk++) {\n var rkey = rkeys[rk];\n if (rkey !== 'protocol') result[rkey] = relative[rkey];\n } //urlParse appends trailing / to urls like http://www.example.com\n\n\n if (slashedProtocol[result.protocol] && result.hostname && !result.pathname) {\n result.path = result.pathname = '/';\n }\n\n result.href = result.format();\n return result;\n }\n\n if (relative.protocol && relative.protocol !== result.protocol) {\n // if it's a known url protocol, then changing\n // the protocol does weird things\n // first, if it's not file:, then we MUST have a host,\n // and if there was a path\n // to begin with, then we MUST have a path.\n // if it is file:, then the host is dropped,\n // because that's known to be hostless.\n // anything else is assumed to be absolute.\n if (!slashedProtocol[relative.protocol]) {\n var keys = Object.keys(relative);\n\n for (var v = 0; v < keys.length; v++) {\n var k = keys[v];\n result[k] = relative[k];\n }\n\n result.href = result.format();\n return result;\n }\n\n result.protocol = relative.protocol;\n\n if (!relative.host && !hostlessProtocol[relative.protocol]) {\n var relPath = (relative.pathname || '').split('/');\n\n while (relPath.length && !(relative.host = relPath.shift())) {\n ;\n }\n\n if (!relative.host) relative.host = '';\n if (!relative.hostname) relative.hostname = '';\n if (relPath[0] !== '') relPath.unshift('');\n if (relPath.length < 2) relPath.unshift('');\n result.pathname = relPath.join('/');\n } else {\n result.pathname = relative.pathname;\n }\n\n result.search = relative.search;\n result.query = relative.query;\n result.host = relative.host || '';\n result.auth = relative.auth;\n result.hostname = relative.hostname || relative.host;\n result.port = relative.port; // to support http.request\n\n if (result.pathname || result.search) {\n var p = result.pathname || '';\n var s = result.search || '';\n result.path = p + s;\n }\n\n result.slashes = result.slashes || relative.slashes;\n result.href = result.format();\n return result;\n }\n\n var isSourceAbs = result.pathname && result.pathname.charAt(0) === '/',\n isRelAbs = relative.host || relative.pathname && relative.pathname.charAt(0) === '/',\n mustEndAbs = isRelAbs || isSourceAbs || result.host && relative.pathname,\n removeAllDots = mustEndAbs,\n srcPath = result.pathname && result.pathname.split('/') || [],\n relPath = relative.pathname && relative.pathname.split('/') || [],\n psychotic = result.protocol && !slashedProtocol[result.protocol]; // if the url is a non-slashed url, then relative\n // links like ../.. should be able\n // to crawl up to the hostname, as well. This is strange.\n // result.protocol has already been set by now.\n // Later on, put the first path part into the host field.\n\n if (psychotic) {\n result.hostname = '';\n result.port = null;\n\n if (result.host) {\n if (srcPath[0] === '') srcPath[0] = result.host;else srcPath.unshift(result.host);\n }\n\n result.host = '';\n\n if (relative.protocol) {\n relative.hostname = null;\n relative.port = null;\n\n if (relative.host) {\n if (relPath[0] === '') relPath[0] = relative.host;else relPath.unshift(relative.host);\n }\n\n relative.host = null;\n }\n\n mustEndAbs = mustEndAbs && (relPath[0] === '' || srcPath[0] === '');\n }\n\n if (isRelAbs) {\n // it's absolute.\n result.host = relative.host || relative.host === '' ? relative.host : result.host;\n result.hostname = relative.hostname || relative.hostname === '' ? relative.hostname : result.hostname;\n result.search = relative.search;\n result.query = relative.query;\n srcPath = relPath; // fall through to the dot-handling below.\n } else if (relPath.length) {\n // it's relative\n // throw away the existing file, and take the new path instead.\n if (!srcPath) srcPath = [];\n srcPath.pop();\n srcPath = srcPath.concat(relPath);\n result.search = relative.search;\n result.query = relative.query;\n } else if (!util.isNullOrUndefined(relative.search)) {\n // just pull out the search.\n // like href='?foo'.\n // Put this after the other two cases because it simplifies the booleans\n if (psychotic) {\n result.hostname = result.host = srcPath.shift(); //occationaly the auth can get stuck only in host\n //this especially happens in cases like\n //url.resolveObject('mailto:local1@domain1', 'local2@domain2')\n\n var authInHost = result.host && result.host.indexOf('@') > 0 ? result.host.split('@') : false;\n\n if (authInHost) {\n result.auth = authInHost.shift();\n result.host = result.hostname = authInHost.shift();\n }\n }\n\n result.search = relative.search;\n result.query = relative.query; //to support http.request\n\n if (!util.isNull(result.pathname) || !util.isNull(result.search)) {\n result.path = (result.pathname ? result.pathname : '') + (result.search ? result.search : '');\n }\n\n result.href = result.format();\n return result;\n }\n\n if (!srcPath.length) {\n // no path at all. easy.\n // we've already handled the other stuff above.\n result.pathname = null; //to support http.request\n\n if (result.search) {\n result.path = '/' + result.search;\n } else {\n result.path = null;\n }\n\n result.href = result.format();\n return result;\n } // if a url ENDs in . or .., then it must get a trailing slash.\n // however, if it ends in anything else non-slashy,\n // then it must NOT get a trailing slash.\n\n\n var last = srcPath.slice(-1)[0];\n var hasTrailingSlash = (result.host || relative.host || srcPath.length > 1) && (last === '.' || last === '..') || last === ''; // strip single dots, resolve double dots to parent dir\n // if the path tries to go above the root, `up` ends up > 0\n\n var up = 0;\n\n for (var i = srcPath.length; i >= 0; i--) {\n last = srcPath[i];\n\n if (last === '.') {\n srcPath.splice(i, 1);\n } else if (last === '..') {\n srcPath.splice(i, 1);\n up++;\n } else if (up) {\n srcPath.splice(i, 1);\n up--;\n }\n } // if the path is allowed to go above the root, restore leading ..s\n\n\n if (!mustEndAbs && !removeAllDots) {\n for (; up--; up) {\n srcPath.unshift('..');\n }\n }\n\n if (mustEndAbs && srcPath[0] !== '' && (!srcPath[0] || srcPath[0].charAt(0) !== '/')) {\n srcPath.unshift('');\n }\n\n if (hasTrailingSlash && srcPath.join('/').substr(-1) !== '/') {\n srcPath.push('');\n }\n\n var isAbsolute = srcPath[0] === '' || srcPath[0] && srcPath[0].charAt(0) === '/'; // put the host back\n\n if (psychotic) {\n result.hostname = result.host = isAbsolute ? '' : srcPath.length ? srcPath.shift() : ''; //occationaly the auth can get stuck only in host\n //this especially happens in cases like\n //url.resolveObject('mailto:local1@domain1', 'local2@domain2')\n\n var authInHost = result.host && result.host.indexOf('@') > 0 ? result.host.split('@') : false;\n\n if (authInHost) {\n result.auth = authInHost.shift();\n result.host = result.hostname = authInHost.shift();\n }\n }\n\n mustEndAbs = mustEndAbs || result.host && srcPath.length;\n\n if (mustEndAbs && !isAbsolute) {\n srcPath.unshift('');\n }\n\n if (!srcPath.length) {\n result.pathname = null;\n result.path = null;\n } else {\n result.pathname = srcPath.join('/');\n } //to support request.http\n\n\n if (!util.isNull(result.pathname) || !util.isNull(result.search)) {\n result.path = (result.pathname ? result.pathname : '') + (result.search ? result.search : '');\n }\n\n result.auth = relative.auth || result.auth;\n result.slashes = result.slashes || relative.slashes;\n result.href = result.format();\n return result;\n};\n\nUrl.prototype.parseHost = function () {\n var host = this.host;\n var port = portPattern.exec(host);\n\n if (port) {\n port = port[0];\n\n if (port !== ':') {\n this.port = port.substr(1);\n }\n\n host = host.substr(0, host.length - port.length);\n }\n\n if (host) this.hostname = host;\n};","map":null,"metadata":{},"sourceType":"script"} |