lukechilds
/
node
mirror of https://github.com/lukechilds/node.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
Browse Source

lint url.js

v0.7.4-release
isaacs 14 years ago
committed by Ryan Dahl
parent
c5c1dc5dda
commit
7c57eb2aec
1 changed files with 146 additions and 115 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 261
      lib/url.js

261
lib/url.js
View File

@ -3,26 +3,35 @@ exports.resolve = urlResolve;
exports.resolveObject = urlResolveObject;
exports.format = urlFormat;
// define these here so at least they only have to be compiled once on the first module load.
// define these here so at least they only have to be
// compiled once on the first module load.
var protocolPattern = /^([a-z0-9]+:)/,
portPattern = /:[0-9]+$/,
nonHostChars = ["/", "?", ";", "#"],
hostlessProtocol = {
"file":true,
"file:":true
},
slashedProtocol = {
"http":true, "https":true, "ftp":true, "gopher":true, "file":true,
"http:":true, "https:":true, "ftp:":true, "gopher:":true, "file:":true
},
path = require("path"), // internal module, guaranteed to be loaded already.
querystring = require('querystring');
function urlParse (url, parseQueryString, slashesDenoteHost) {
if (url && typeof(url) === "object" && url.href) return url;
var out = { href : url },
rest = url;
portPattern = /:[0-9]+$/,
nonHostChars = ['/', '?', ';', '#'],
hostlessProtocol = {
'file': true,
'file:': true
},
slashedProtocol = {
'http': true,
'https': true,
'ftp': true,
'gopher': true,
'file': true,
'http:': true,
'https:': true,
'ftp:': true,
'gopher:': true,
'file:': true
},
path = require('path'), // internal module, guaranteed to be loaded already.
querystring = require('querystring');
function urlParse(url, parseQueryString, slashesDenoteHost) {
if (url && typeof(url) === 'object' && url.href) return url;
var out = { href: url },
rest = url;
var proto = protocolPattern.exec(rest);
if (proto) {
@ -36,27 +45,29 @@ function urlParse (url, parseQueryString, slashesDenoteHost) {
// resolution will treat //foo/bar as host=foo,path=bar because that's
// how the browser resolves relative URLs.
if (slashesDenoteHost || proto || rest.match(/^\/\/[^@\/]+@[^@\/]+/)) {
var slashes = rest.substr(0, 2) === "//";
var slashes = rest.substr(0, 2) === '//';
if (slashes && !(proto && hostlessProtocol[proto])) {
rest = rest.substr(2);
out.slashes = true;
}
}
if (!hostlessProtocol[proto] && (slashes || (proto && !slashedProtocol[proto]))) {
if (!hostlessProtocol[proto] &&
(slashes || (proto && !slashedProtocol[proto]))) {
// there's a hostname.
// the first instance of /, ?, ;, or # ends the host.
// don't enforce full RFC correctness, just be unstupid about it.
var firstNonHost = -1;
for (var i = 0, l = nonHostChars.length; i < l; i ++) {
for (var i = 0, l = nonHostChars.length; i < l; i++) {
var index = rest.indexOf(nonHostChars[i]);
if (index !== -1 && (firstNonHost < 0 || index < firstNonHost)) firstNonHost = index;
if (index !== -1 &&
(firstNonHost < 0 || index < firstNonHost)) firstNonHost = index;
}
if (firstNonHost !== -1) {
out.host = rest.substr(0, firstNonHost);
rest = rest.substr(firstNonHost);
} else {
out.host = rest;
rest = "";
rest = '';
}
// pull out the auth and port.
@ -66,22 +77,23 @@ function urlParse (url, parseQueryString, slashesDenoteHost) {
var key = keys[i];
out[key] = p[key];
}
// we've indicated that there is a hostname, so even if it's empty, it has to be present.
out.hostname = out.hostname || "";
// we've indicated that there is a hostname,
// so even if it's empty, it has to be present.
out.hostname = out.hostname || '';
}
// now rest is set to the post-host stuff.
// chop off from the tail first.
var hash = rest.indexOf("#");
var hash = rest.indexOf('#');
if (hash !== -1) {
// got a fragment string.
out.hash = rest.substr(hash);
rest = rest.slice(0, hash);
}
var qm = rest.indexOf("?");
var qm = rest.indexOf('?');
if (qm !== -1) {
out.search = rest.substr(qm);
out.query = rest.substr(qm+1);
out.query = rest.substr(qm + 1);
if (parseQueryString) {
out.query = querystring.parse(out.query);
}
@ -90,52 +102,57 @@ function urlParse (url, parseQueryString, slashesDenoteHost) {
if (rest) out.pathname = rest;
return out;
};
}
// format a parsed object into a url string
function urlFormat (obj) {
// ensure it's an object, and not a string url. If it's an obj, this is a no-op.
// this way, you can call url_format() on strings to clean up potentially wonky urls.
if (typeof(obj) === "string") obj = urlParse(obj);
var protocol = obj.protocol || "",
host = (obj.host !== undefined) ? obj.host
: obj.hostname !== undefined ? (
(obj.auth ? obj.auth + "@" : "")
+ obj.hostname
+ (obj.port ? ":" + obj.port : "")
)
: false,
pathname = obj.pathname || "",
search = obj.search || (
obj.query && ( "?" + (
typeof(obj.query) === "object"
? querystring.stringify(obj.query)
: String(obj.query)
))
) || "",
hash = obj.hash || "";
if (protocol && protocol.substr(-1) !== ":") protocol += ":";
function urlFormat(obj) {
// ensure it's an object, and not a string url.
// If it's an obj, this is a no-op.
// this way, you can call url_format() on strings
// to clean up potentially wonky urls.
if (typeof(obj) === 'string') obj = urlParse(obj);
var protocol = obj.protocol || '',
host = (obj.host !== undefined) ? obj.host :
obj.hostname !== undefined ? (
(obj.auth ? obj.auth + '@' : '') +
obj.hostname +
(obj.port ? ':' + obj.port : '')
) :
false,
pathname = obj.pathname || '',
search = obj.search || (
obj.query && ('?' + (
typeof(obj.query) === 'object' ?
querystring.stringify(obj.query) :
String(obj.query)
))
) || '',
hash = obj.hash || '';
if (protocol && protocol.substr(-1) !== ':') protocol += ':';
// only the slashedProtocols get the //. Not mailto:, xmpp:, etc.
// unless they had them to begin with.
if (obj.slashes || (!protocol || slashedProtocol[protocol]) && host !== false) {
host = "//" + (host || "");
if (pathname && pathname.charAt(0) !== "/") pathname = "/" + pathname;
} else if (!host) host = "";
if (obj.slashes ||
(!protocol || slashedProtocol[protocol]) && host !== false) {
host = '//' + (host || '');
if (pathname && pathname.charAt(0) !== '/') pathname = '/' + pathname;
} else if (!host) {
host = '';
}
if (hash && hash.charAt(0) !== "#") hash = "#" + hash;
if (search && search.charAt(0) !== "?") search = "?" + search;
if (hash && hash.charAt(0) !== '#') hash = '#' + hash;
if (search && search.charAt(0) !== '?') search = '?' + search;
return protocol + host + pathname + search + hash;
};
}
function urlResolve (source, relative) {
function urlResolve(source, relative) {
return urlFormat(urlResolveObject(source, relative));
};
}
function urlResolveObject (source, relative) {
function urlResolveObject(source, relative) {
if (!source) return relative;
source = urlParse(urlFormat(source), false, true);
@ -144,7 +161,7 @@ function urlResolveObject (source, relative) {
// hash is always overridden, no matter what.
source.hash = relative.hash;
if (relative.href === "") return source;
if (relative.href === '') return source;
// hrefs like //foo/bar always cut to the protocol.
if (relative.slashes && !relative.protocol) {
@ -153,55 +170,62 @@ function urlResolveObject (source, relative) {
}
if (relative.protocol && relative.protocol !== source.protocol) {
// if it's a known url protocol, then changing the protocol does weird things
// first, if it's not file:, then we MUST have a host, and if there was a path
// if it's a known url protocol, then changing
// the protocol does weird things
// first, if it's not file:, then we MUST have a host,
// and if there was a path
// to begin with, then we MUST have a path.
// if it is file:, then the host is dropped, because that's known to be hostless.
// if it is file:, then the host is dropped,
// because that's known to be hostless.
// anything else is assumed to be absolute.
if (!slashedProtocol[relative.protocol]) return relative;
source.protocol = relative.protocol;
if (!relative.host && !hostlessProtocol[relative.protocol]) {
var relPath = (relative.pathname || "").split("/");
var relPath = (relative.pathname || '').split('/');
while (relPath.length && !(relative.host = relPath.shift()));
if (!relative.host) relative.host = "";
if (relPath[0] !== "") relPath.unshift("");
if (relPath.length < 2) relPath.unshift("");
relative.pathname = relPath.join("/");
if (!relative.host) relative.host = '';
if (relPath[0] !== '') relPath.unshift('');
if (relPath.length < 2) relPath.unshift('');
relative.pathname = relPath.join('/');
}
source.pathname = relative.pathname;
source.search = relative.search;
source.query = relative.query;
source.host = relative.host || "";
source.host = relative.host || '';
delete source.auth;
delete source.hostname;
source.port = relative.port;
return source;
}
var isSourceAbs = (source.pathname && source.pathname.charAt(0) === "/"),
isRelAbs = (
relative.host !== undefined
|| relative.pathname && relative.pathname.charAt(0) === "/"
),
mustEndAbs = (isRelAbs || isSourceAbs || (source.host && relative.pathname)),
removeAllDots = mustEndAbs,
srcPath = source.pathname && source.pathname.split("/") || [],
relPath = relative.pathname && relative.pathname.split("/") || [],
psychotic = source.protocol && !slashedProtocol[source.protocol] && source.host !== undefined;
// if the url is a non-slashed url, then relative links like ../.. should be able
var isSourceAbs = (source.pathname && source.pathname.charAt(0) === '/'),
isRelAbs = (
relative.host !== undefined ||
relative.pathname && relative.pathname.charAt(0) === '/'
),
mustEndAbs = (isRelAbs || isSourceAbs ||
(source.host && relative.pathname)),
removeAllDots = mustEndAbs,
srcPath = source.pathname && source.pathname.split('/') || [],
relPath = relative.pathname && relative.pathname.split('/') || [],
psychotic = source.protocol &&
!slashedProtocol[source.protocol] &&
source.host !== undefined;
// if the url is a non-slashed url, then relative
// links like ../.. should be able
// to crawl up to the hostname, as well. This is strange.
// source.protocol has already been set by now.
// Later on, put the first path part into the host field.
if ( psychotic ) {
if (psychotic) {
delete source.hostname;
delete source.auth;
delete source.port;
if (source.host) {
if (srcPath[0] === "") srcPath[0] = source.host;
if (srcPath[0] === '') srcPath[0] = source.host;
else srcPath.unshift(source.host);
}
delete source.host;
@ -211,17 +235,18 @@ function urlResolveObject (source, relative) {
delete relative.auth;
delete relative.port;
if (relative.host) {
if (relPath[0] === "") relPath[0] = relative.host;
if (relPath[0] === '') relPath[0] = relative.host;
else relPath.unshift(relative.host);
}
delete relative.host;
}
mustEndAbs = mustEndAbs && (relPath[0] === "" || srcPath[0] === "");
mustEndAbs = mustEndAbs && (relPath[0] === '' || srcPath[0] === '');
}
if (isRelAbs) {
// it's absolute.
source.host = (relative.host || relative.host === "") ? relative.host : source.host;
source.host = (relative.host || relative.host === '') ?
relative.host : source.host;
source.search = relative.search;
source.query = relative.query;
srcPath = relPath;
@ -234,9 +259,9 @@ function urlResolveObject (source, relative) {
srcPath = srcPath.concat(relPath);
source.search = relative.search;
source.query = relative.query;
} else if ("search" in relative) {
} else if ('search' in relative) {
// just pull out the search.
// like href="?foo".
// like href='?foo'.
// Put this after the other two cases because it simplifies the booleans
if (psychotic) {
source.host = srcPath.shift();
@ -254,58 +279,64 @@ function urlResolveObject (source, relative) {
// resolve dots.
// if a url ENDs in . or .., then it must get a trailing slash.
// however, if it ends in anything else non-slashy, then it must NOT get a trailing slash.
// however, if it ends in anything else non-slashy,
// then it must NOT get a trailing slash.
var last = srcPath.slice(-1)[0];
var hasTrailingSlash = (
(source.host || relative.host) && (last === "." || last === "..")
|| last === ""
);
(source.host || relative.host) && (last === '.' || last === '..') ||
last === '');
// Figure out if this has to end up as an absolute url, or should continue to be relative.
// Figure out if this has to end up as an absolute url,
// or should continue to be relative.
srcPath = path.normalizeArray(srcPath, true);
if (srcPath.length === 1 && srcPath[0] === ".") srcPath = [];
if (srcPath.length === 1 && srcPath[0] === '.') srcPath = [];
if (mustEndAbs || removeAllDots) {
// all dots must go.
var dirs = [];
srcPath.forEach(function (dir, i) {
if (dir === "..") dirs.pop();
else if (dir !== ".") dirs.push(dir);
srcPath.forEach(function(dir, i) {
if (dir === '..') {
dirs.pop();
} else if (dir !== '.') {
dirs.push(dir);
}
});
if (mustEndAbs && dirs[0] !== "" && (!dirs[0] || dirs[0].charAt(0) !== "/")) {
dirs.unshift("");
if (mustEndAbs && dirs[0] !== '' &&
(!dirs[0] || dirs[0].charAt(0) !== '/')) {
dirs.unshift('');
}
srcPath = dirs;
}
if (hasTrailingSlash && (srcPath.join("/").substr(-1) !== "/")) {
srcPath.push("");
if (hasTrailingSlash && (srcPath.join('/').substr(-1) !== '/')) {
srcPath.push('');
}
var isAbsolute = srcPath[0] === "" || (srcPath[0] && srcPath[0].charAt(0) === "/");
var isAbsolute = srcPath[0] === '' ||
(srcPath[0] && srcPath[0].charAt(0) === '/');
// put the host back
if ( psychotic ) {
source.host = isAbsolute ? "" : srcPath.shift();
if (psychotic) {
source.host = isAbsolute ? '' : srcPath.shift();
}
mustEndAbs = mustEndAbs || (source.host && srcPath.length);
if (mustEndAbs && !isAbsolute) {
srcPath.unshift("");
srcPath.unshift('');
}
source.pathname = srcPath.join("/");
source.pathname = srcPath.join('/');
return source;
};
}
function parseHost (host) {
function parseHost(host) {
var out = {};
var at = host.indexOf("@");
var at = host.indexOf('@');
if (at !== -1) {
out.auth = host.substr(0, at);
host = host.substr(at+1); // drop the @
host = host.substr(at + 1); // drop the @
}
var port = portPattern.exec(host);
if (port) {

Write Preview
Loading…
Cancel
Save