mozilla
/
snowl
зеркало из https://github.com/mozilla/snowl.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
snowl/modules/twitter.js

558 строки
20 KiB
JavaScript
Исходник Ответственный История

/* ***** BEGIN LICENSE BLOCK *****
* Version: MPL 1.1/GPL 2.0/LGPL 2.1
*
* The contents of this file are subject to the Mozilla Public License Version
* 1.1 (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
* http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
* for the specific language governing rights and limitations under the
* License.
*
* The Original Code is Snowl.
*
* The Initial Developer of the Original Code is Mozilla.
* Portions created by the Initial Developer are Copyright (C) 2008
* the Initial Developer. All Rights Reserved.
*
* Contributor(s):
* Myk Melez <myk@mozilla.org>
*
* Alternatively, the contents of this file may be used under the terms of
* either the GNU General Public License Version 2 or later (the "GPL"), or
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
* in which case the provisions of the GPL or the LGPL are applicable instead
* of those above. If you wish to allow use of your version of this file only
* under the terms of either the GPL or the LGPL, and not to allow others to
* use your version of this file under the terms of the MPL, indicate your
* decision by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL or the LGPL. If you do not delete
* the provisions above, a recipient may use your version of this file under
* the terms of any one of the MPL, the GPL or the LGPL.
*
* ***** END LICENSE BLOCK ***** */
let EXPORTED_SYMBOLS = ["SnowlTwitter"];
const Cc = Components.classes;
const Ci = Components.interfaces;
const Cr = Components.results;
const Cu = Components.utils;
// modules that come with Firefox
Cu.import("resource://gre/modules/XPCOMUtils.jsm");
Cu.import("resource://gre/modules/ISO8601DateUtils.jsm");
// modules that are generic
Cu.import("resource://snowl/modules/log4moz.js");
Cu.import("resource://snowl/modules/Mixins.js");
Cu.import("resource://snowl/modules/Observers.js");
Cu.import("resource://snowl/modules/request.js");
Cu.import("resource://snowl/modules/URI.js");
// modules that are Snowl-specific
Cu.import("resource://snowl/modules/constants.js");
Cu.import("resource://snowl/modules/datastore.js");
Cu.import("resource://snowl/modules/source.js");
Cu.import("resource://snowl/modules/target.js");
Cu.import("resource://snowl/modules/identity.js");
Cu.import("resource://snowl/modules/message.js");
Cu.import("resource://snowl/modules/utils.js");
Cu.import("resource://snowl/modules/service.js");
// FIXME: make strands.js into a module.
let loader = Cc["@mozilla.org/moz/jssubscript-loader;1"].getService(Ci.mozIJSSubScriptLoader);
loader.loadSubScript("chrome://snowl/content/strands.js");
const TYPE = "SnowlTwitter";
const NAME = "Twitter";
const MACHINE_URI = URI.get("https://twitter.com");
// XXX Should this be simply http://twitter.com ?
const HUMAN_URI = URI.get("http://twitter.com/home");
/**
* The HTTP authentication realm under which to save credentials via the login
* manager. We save them under a realm whose name we define instead of the one
* that Twitter provides (currently "Twitter API") because we set our own
* Authorization header, and that happens before we get a response from Twitter,
* so we can't depend on the value Twitter sets when it responds, because
* we don't know it yet.
*
* Using our own realm also has the beneficial side effect that users browsing
* their saved credentials in preferences will see our realm next to
* the credentials that were saved by Snowl, which seems a better explanation
* of where the credentials come from than the one Twitter provides.
*
* The reason we set our own Authorization header is that Necko assumes users
* will only be logged into a single account for a given authentication realm
* in any given session, so it caches credentials and reuses them for all
* requests to the same realm. Setting the Authorization header ourselves
* ensures that we determine the credentials being used for our requests,
* which is necessary to support multiple Twitter accounts.
*
* We could have theoretically worked around the problem by putting the username
* into the URL (i.e. https://username@twitter.com...) and falling back on our
* notification callback to get the saved credentials, by which point we'd know
* the authentication realm.
*
* But experimentation showed that only worked for serialized requests;
* concurrent requests (like when we refresh two Twitter accounts at the same
* time asynchronously) cause Necko to again use the same credentials for both
* refreshes, even though we've specified different usernames in the URLs
* for those refreshes.
*
* And it had the side-effect that Necko stopped saving credentials at all
* after the requests completed, so a user with a single account who didn't save
* their credentials was prompted to enter them every time we refreshed.
*
* FIXME: file a bug on this bad behavior of Necko during concurrent requests.
*
* We could have also worked around the problem by also injecting the password
* into the request URLs (i.e. https://username:password@twitter.com...),
* but then we'd be putting passwords into URLs, which is considered harmful
* because URLs leak into visible places (like the Error Console).
*/
const AUTH_REALM = "Snowl";
// This module is based on the API documented at http://apiwiki.twitter.com/.
// FIXME: make the constructor accept credentials instead of passing them
// to the subscribe function.
function SnowlTwitter(aID, aName, aMachineURI, aHumanURI, aUsername, aLastRefreshed, aImportance, aPlaceID) {
// Use the given machine URI, if available. We use this in unit tests
// to point the account to a test server rather than the actual Twitter
// servers.
let machineURI = aMachineURI || MACHINE_URI;
// FIXME: figure out a better solution than hanging the first mixed in init()
// method on this object's prototype but calling the second one directly
// because it didn't actually get mixed in because it already existed!
this.init(aID, aName, machineURI, HUMAN_URI, aUsername, aLastRefreshed, aImportance, aPlaceID);
SnowlTarget.prototype.init.call(this);
}
SnowlTwitter.prototype = {
// The constructor property is defined automatically, but we destroy it
// when we redefine the prototype, so we redefine it here in case we ever
// need to check it to find out what kind of object an instance is.
constructor: SnowlTwitter,
get _log() {
let logger = Log4Moz.repository.getLogger("Snowl.Twitter." + this.username);
this.__defineGetter__("_log", function() logger);
return this._log;
},
//**************************************************************************//
// Abstract Class Composition Declarations
_classes: [SnowlSource, SnowlTarget],
implements: function(cls) {
return (this._classes.indexOf(cls) != -1);
},
//**************************************************************************//
// SnowlSource
refreshInterval: 1000 * 60 * 3, // 3 minutes
// refresh is defined elsewhere.
//**************************************************************************//
// SnowlTarget
maxMessageLength: 140,
// send is defined elsewhere.
//**************************************************************************//
// Notification Callbacks for Authentication
// FIXME: factor this out with the equivalent code in feed.js.
// If we prompt the user to authenticate, and the user asks us to remember
// their password, we store the nsIAuthInformation in this property until
// the request succeeds, at which point we store it with the login manager.
_authInfo: null,
get _loginManager() {
let loginManager = Cc["@mozilla.org/login-manager;1"].
getService(Ci.nsILoginManager);
this.__defineGetter__("_loginManager", function() loginManager);
return this._loginManager;
},
/**
* The saved credentials for this Twitter account, if any.
* FIXME: we memoize this and never refresh it, which won't do once we have
* long-lived account objects, so don't memoize this at all (attach it to
* its request and kill it once the request is done) or invalidate it when
* the set of credentials changes.
*/
get _savedLogin() {
// XXX Should we be using channel.URI.prePath instead of
// this.machineURI.prePath in case the old URI redirects us to a new one
// at a different hostname?
return this._loginManager.
findLogins({}, this.machineURI.prePath, null, AUTH_REALM).
filter(function(login) login.username == this.username, this)
[0];
},
// nsISupports
QueryInterface: XPCOMUtils.generateQI([Ci.nsIAuthPrompt2]),
// nsIInterfaceRequestor
getInterface: function(iid) {
return this.QueryInterface(iid);
},
// nsIAuthPrompt2
promptAuth: function(channel, level, authInfo) {
this._log.debug("promptAuth: this.name = " + this.name + "; this.username = " + this.username);
this._log.debug("promptAuth: this.name = " + this.name + "; authInfo.realm = " + authInfo.realm);
let args = Cc["@mozilla.org/supports-array;1"].createInstance(Ci.nsISupportsArray);
args.AppendElement({ wrappedJSObject: this });
args.AppendElement(authInfo);
// |result| is how the dialog passes information back to us. It sets two
// properties on the object:
// |proceed|, which we return from this function, and which determines
// whether or not authentication can proceed using the value(s) entered
// by the user;
// |remember|, which determines whether or not we save the user's login
// with the login manager once the request succeeds.
let result = {};
args.AppendElement({ wrappedJSObject: result });
let ww = Cc["@mozilla.org/embedcomp/window-watcher;1"].getService(Ci.nsIWindowWatcher);
ww.openWindow(null,
"chrome://snowl/content/login.xul",
null,
"chrome,centerscreen,dialog,modal",
args);
if (result.remember)
this._authInfo = authInfo;
else
this._authInfo = null;
return result.proceed;
},
asyncPromptAuth: function() {
throw Cr.NS_ERROR_NOT_IMPLEMENTED;
},
//**************************************************************************//
// Refreshment
get _stmtGetMaxExternalID() {
let statement = SnowlDatastore.createStatement(
"SELECT MAX(externalID) AS maxID FROM messages WHERE sourceID = :sourceID"
);
this.__defineGetter__("_stmtGetMaxExternalID", function() statement);
return this._stmtGetMaxExternalID;
},
/**
* Get the maximum external ID of the messages received from this source.
* Newer messages always have larger integer IDs, so we can query for only
* new messages by specifying since_id=[max ID] in the refresh request.
*
* @returns {Number}
* the maximum external ID, if any
*/
_getMaxExternalID: function() {
let maxID = null;
try {
this._stmtGetMaxExternalID.params.sourceID = this.id;
if (this._stmtGetMaxExternalID.step())
maxID = this._stmtGetMaxExternalID.row["maxID"];
}
finally {
this._stmtGetMaxExternalID.reset();
}
return maxID;
},
/**
* Refresh the feed, retrieving the latest information in it.
*
* @param time {Date} [optional]
* when the refresh occurs; determines the received time of new
* messages; we let the caller specify this so a caller refreshing
* multiple feeds can give their messages the same received time
*/
refresh: function(time) {
if (typeof time == "undefined" || time == null)
time = new Date();
// this._log.info("start refresh " + this.username + " at " + time);
Observers.notify("snowl:subscribe:get:start", this);
// URL parameters that modify the return value of the request.
let params = [];
// Retrieve up to 200 messages, the maximum we're allowed to retrieve.
params.push("count=200");
// Retrieve only messages newer than the most recent one already retrieved.
let (maxID = this._getMaxExternalID()) {
if (maxID)
params.push("since_id=" + maxID);
}
let url = this.machineURI.spec.replace("^(https?://)", "$1" + this.username + "@") +
"/statuses/friends_timeline.json?" + params.join("&");
this._log.debug("refresh: this.name = " + this.name + "; url = " + url);
let requestHeaders = {};
// If the login manager has saved credentials for this account, provide them
// to the server. Otherwise, no worries, Necko will automatically call our
// notification callback, which will prompt the user to enter their credentials.
if (this._savedLogin) {
this._log.info("setting Authorization header with username " + this.username);
let credentials = btoa(this.username + ":" + this._savedLogin.password);
requestHeaders.Authorization = "Basic " + credentials;
}
let request = new Request({
url: url,
notificationCallbacks: this,
requestHeaders: requestHeaders
});
// FIXME: remove subscribe from this notification's name.
Observers.notify("snowl:subscribe:connect:end", this, request.status);
this.lastStatus = request.status + " (" + request.statusText + ")";
if (request.status < 200 || request.status > 299 || request.responseText.length == 0) {
this.onRefreshError();
return;
}
// _authInfo only gets set if we prompted the user to authenticate
// and the user checked the "remember password" box. Since we're here,
// it means the request succeeded, so we save the login.
if (this._authInfo) {
this._saveLogin(this._authInfo);
this._authInfo = null;
}
let items = JSON.parse(request.responseText);
this.messages = this._processItems(items, time);
this.lastRefreshed = time;
Observers.notify("snowl:subscribe:get:end", this);
},
//**************************************************************************//
// Processing
/**
* Process an array of items (from the server) into an array of messages.
*
* @param items {Array} the items to process
* @param received {Date} when the items were received
*/
_processItems: function(items, received) {
this._log.trace("processing items");
let messages = [];
for each (let item in items) {
try {
let message = this._processItem(item, received);
messages.push(message);
}
catch(ex) {
this._log.error("couldn't process item " + item.id + ": " + ex);
}
}
return messages;
},
_processItem: function(item, received) {
this._log.trace("processing item " + item.id);
let message = new SnowlMessage();
message.source = this;
message.externalID = item.id;
message.timestamp = new Date(item.created_at);
message.received = received || new Date();
message.author = new SnowlIdentity(null, this.id, item.user.id);
message.author.person = new SnowlPerson(null, item.user.screen_name, null, item.user.url, item.user.profile_image_url);
message.content =
new SnowlMessagePart({
partType: PART_TYPE_CONTENT,
content: item.text,
mediaType: "text/plain"
});
return message;
},
// XXX Perhaps factor this out with the identical function in feed.js,
// although this function supports multiple accounts with the same server
// and doesn't allow the user to change their username, so maybe that's
// not possible (or perhaps we can reconcile those differences).
_saveLogin: function(authInfo) {
// Create a new login with the auth information we obtained from the user.
let LoginInfo = new Components.Constructor("@mozilla.org/login-manager/loginInfo;1",
Ci.nsILoginInfo,
"init");
// XXX Should we be using channel.URI.prePath in case the old URI
// redirects us to a new one at a different hostname?
let newLogin = new LoginInfo(this.machineURI.prePath,
null,
AUTH_REALM,
authInfo.username,
authInfo.password,
"",
"");
// If there are credentials with the same username, we replace them.
// Otherwise, we add the new credentials.
if (this._savedLogin)
this._loginManager.modifyLogin(this._savedLogin, newLogin);
else
this._loginManager.addLogin(newLogin);
},
//**************************************************************************//
// Sending
_successCallback: null,
_errorCallback: null,
send: function(content, successCallback, errorCallback) {
Observers.notify("snowl:send:start", this);
let data = "status=" + encodeURIComponent(content) + "&source=snowl";
// + "&in_reply_to_status_id=" + encodeURIComponent(inReplyToID);
this._successCallback = successCallback;
this._errorCallback = errorCallback;
let request = Cc["@mozilla.org/xmlextras/xmlhttprequest;1"].createInstance();
// FIXME: make a TwitterRequest (or plain Request) object that caches
// references to the callbacks and the SnowlTwitter instance so we don't
// have to cache them in the object itself, which could cause problems
// if we were to persist the instance and send multiple messages through
// it simultaneously.
request.QueryInterface(Ci.nsIDOMEventTarget);
let (t = this) {
request.addEventListener("load", function(e) { t.onSendLoad(e) }, false);
request.addEventListener("error", function(e) { t.onSendError(e) }, false);
}
request.QueryInterface(Ci.nsIXMLHttpRequest);
request.open("POST", this.machineURI.spec.replace("^(https?://)", "$1" + this.username + "@") + "/statuses/update.json", true);
// If the login manager has saved credentials for this account, provide them
// to the server. Otherwise, no worries, Necko will automatically call our
// notification callback, which will prompt the user to enter their credentials.
if (this._savedLogin) {
let credentials = btoa(this.username + ":" + this._savedLogin.password);
request.setRequestHeader("Authorization", "Basic " + credentials);
}
request.channel.notificationCallbacks = this;
request.setRequestHeader("Content-Type", "application/x-www-form-urlencoded");
request.send(data);
},
onSendLoad: function(event) {
let request = event.target;
// FIXME: the next three chunks of code are the same for multiple
// load handlers; find some way to factor them out.
// If the request failed, let the error handler handle it.
// XXX Do we need this? Don't such failures call the error handler directly?
if (request.status < 200 || request.status > 299) {
this.onSendError(event);
return;
}
// If the response is empty, assume failure.
// XXX What's the right way to handle this?
if (request.responseText.length == 0) {
this.onSendError(event);
return;
}
this._log.info("onSendLoad: " + request.responseText);
// _authInfo only gets set if we prompted the user to authenticate
// and the user checked the "remember password" box. Since we're here,
// it means the request succeeded, so we save the login.
if (this._authInfo)
this._saveLogin(this._authInfo);
this._processSend(request.responseText);
if (this._successCallback)
this._successCallback();
this._resetSend();
},
onSendError: function(event) {
let request = event.target;
// Sometimes an attempt to retrieve status text throws NS_ERROR_NOT_AVAILABLE
let statusText = "";
try { statusText = request.statusText } catch(ex) {}
this._log.error("onSendError: " + request.status + " (" + statusText + ")");
if (this._errorCallback)
this._errorCallback();
this._resetSend();
},
_processSend: function(responseText) {
let JSON = Cc["@mozilla.org/dom/json;1"].createInstance(Ci.nsIJSON);
let item = JSON.decode(responseText);
let message = this._processItem(item);
message.persist();
},
_resetSend: function() {
this._successCallback = null;
this._errorCallback = null;
this._authInfo = null;
}
};
Mixins.mix(SnowlSource).into(SnowlTwitter);
Mixins.mix(SnowlSource.prototype).into(SnowlTwitter.prototype);
Mixins.mix(SnowlTarget).into(SnowlTwitter);
Mixins.mix(SnowlTarget.prototype).into(SnowlTwitter.prototype);
SnowlService.addAccountType(SnowlTwitter);
Ссылка в новой задаче Показать git blame Копировать постоянную ссылку