If I wish to fetch data from a remote server, then JSONP is the tool of choice I believe. But I am confused by an example I have seen:
<script src="http://ajax.googleapis.com/ajax/libs/jquery/1.8.0/jquery.min.js"></script>
<script type="text/javascript">
$(document).ready(function() {
$.ajax({
dataType: 'jsonp',
data: 'p3=c',
jsonp: 'callback',
url: 'http://someserver.com/app?p1=a&p2=b',
success: function (data) {
console.log("data="+data);
$.each(data, function (i, r) {
console.log("i="+i);
console.log("r="+r);
});
},
});
});
</script>
I can see that in the request, a callback parameter has been added with value in the format jQuery1234567890. When I look at the app that processes that request, it extracts the callback parameter from the request and wraps the json data to be returned with that and relevant brackets, so it ends up returning something like this:
jQuery1234567890([{"x":"100","y":"101"},{"x":"200","y":"201"}])
So my first questions are:
(1) Is the app correct to have done what it has?
(2) What has jQuery / JSONP actually done for us?
I was assuming that jQuery would see the dataType of "jsonp", insert a script tag into the DOM, the browser would then download and execute the script. If that's right, has jQuery created the function jQuery1234567890, the implementation of which is to pass the parameter on to the success function?
(3) Is my understanding correct (I don't think it is)?
Thank you,
Paul
(1) Is the app correct to have done what it has?
Yes, that's a correct JSONP format
(2) What has jQuery / JSONP actually done for us?
Notified the server application that JSONP is desired by placing a &callback=jQuery1234567890 in the request
I was assuming that jQuery would see the dataType of "jsonp", insert a script tag into the DOM, the browser would then download and execute the script. If that's right, has jQuery created the function jQuery1234567890, the implementation of which is to pass the parameter on to the success function?
(3) Is my understanding correct (I don't think it is)?
Yes, your understanding is correct. It has created a script with a jQuery1234567890 function which is invoked when the requested scripted is loaded. And as you stated the parameter receives the data and passes it on to the $.ajax internals, which invokes the success callback
From the ajax docs for the jsonp option:
Override the callback function name in a jsonp
request. This value will be used instead of 'callback' in the
'callback=?' part of the query string in the url. So
{jsonp:'onJSONPLoad'} would result in 'onJSONPLoad=?' passed to the
server.
So using jsonp: 'callback' overrides callback with callback, essentially doing nothing.
The other stuff you're seeing is generated by jQuery so that you don't have to do it yourself. You get to simply treat this request like any other ajax request in jquery and not worry about the implementation of the jsonp.
Related
I need to synchronously call a controller (url) in my Javascript. I could do this with an Ajax function like so:
$.ajax({
type: 'POST',
url: BASE + '/publication/storeWorldCat',
data: data,
success: success,
dataType: dataType,
async:false
});
I don't understand why I should use an Ajax function with a async:false parameter as Ajax is asynchronous by definition. $post() is asynchronous too.
Is there a better way?
Thanks for your help!
Looking at the documentation of jQuery.ajax (quoting) :
By default, all requests are sent
asynchronously (i.e. this is set to
true by default). If you need
synchronous requests, set this option
to false.
Granted, AJAX stands for Asynchronous Javascript and XML so any request sent in Javascript that is not sent asynchronously or does not use XML as the data format should technically not be called an AJAX request. With the popularity of JSON nowadays, it looks like the $.ajax function is REALLY badly named since it allows any data format.
But many other acronyms are chosen to be easy to pronounce at the cost of not being totally correct, and I think we can live with it. So to answer your questions:
Why has the jQuery team decided to name this function ajax? Because this acronym was already used by everyone, regardless of the data format used and the "synchronousness". Why is everyone using this acronym wrongly? Probably because it did not make sense to make a new acronym for each data format, each time with an optional 'A' for whether it is asynchronous.
Why [you] should use an Ajax function with a async:false parameter as Ajax is asynchronous by definition?
Well, you don't have to:
$.sendRequest = $.ajax
And now you can use a function with a better name in your code. You can even change the default behaviour:
$.sendRequest = function(options) {
if (typeof options.async === undefined) {
options.async = false;
}
$.ajax.apply($, arguments);
}
I'm building an SPA using Sammy and Knockout powered by a REST Web Service available on a different URL.
I'm noticing some odd behavior when returning JSONP versus JSON when using $.when().done()...
.done() never fires, but .fail() will, even though the status code I receive is 200, and JSONP Linter tells me that my JSONP is valid:
(function($) {
$(function() {
$.when($.getJSON('endpoint1?callback=?', null),
$,getJSON('endpoint2?callback=?', null))
.done(function(resp1, resp2) {
console.log(resp1); // this is never called
});
})
.fail(function(obj) {
console.log(obj); // this is called, but why?
});
});
})(jQuery);
A sample response returned is:
callback({
"external-links": [
{
"nav_link_text": "Stack Overflow",
"url": "http://stackoverflow.com"
}
]
});
If I return JSON instead of JSONP, .done() works as expected. What am I doing wrong or need to change?
The problem is specified here:
ReferenceError: callback is not defined
Your JSONP response has callback hard-coded. That's incorrect. JSONP needs to set the function name dynamically.
When jQuery sends a JSONP request, it creates a function called jQuery123456 (or something like that) and sends that name in the request. It calls endpoint1?callback=jQuery123456. The job of JSONP is to make a call to that function. Your JSONP needs to return:
jQuery123456({
your: 'data'
})
You need to use the value of the callback parameter.
If for some reason, creating the JSONP "dynamically" isn't an option, you can force jQuery to name the callback function it creates. You need to use $.ajax for this:
$.ajax({
url: 'endpoint1',
dataType: 'jsonp',
jsonp: false, // Don't add the "?callback=?" param,
// you're not using it anyway
jsonpCallback: 'callback' // Force jQuery to use "callback"
// as the function name
});
Note: jQuery probably won't like having the same callback value for multiple requests.
I am experiencing an issue in jQuery when I do multiple jsonp requests, all with the same jsonpCallback function. It seems that only for the one of those the callback function is triggered. Are JSONP requests somehow overwriting each other?
Below an example of doing 2 jsonp request to github, and even though both firebug shows that both of them return, the callback function getName is only called for one of them:
function getName(response){
alert(response.data.name);
}
function userinfo(username){
$.ajax({
url: "https://api.github.com/users/" + username,
jsonpCallback: 'getName',
dataType: "jsonp"
});
}
users = ["torvalds", "twitter", "jquery"]
for(var i = 0; i < users.length; i++){
userinfo(users[i]);
}
Your request fired only once because of how jsonp works.
Jsonp means adding a script tag to the page from an outside domain to get around Cross-Site Scripting protections built into modern browsers (and now IE6 and 7 as of April 2011). In order to have that script interact with the rest of the script on the page, the script being loaded in needs to call a function on the page. That function has to exist in the global namespace, meaning there can only be one function by that name. In other words, without JQuery a single jsonp request would look like this:
<script>
function loadJson(json) {
// Read the json
}
</script>
<script src="//outsidedomain.com/something.js"></script>
Where something.js would look like this:
loadJson({name:'Joe'})
something.js in this case has a hard-coded callback to load the JSON it carries, and the page has a hard-coded loadJson function waiting for scripts like this one to load and call it.
Now suppose you want to be able to load json from multiple sources and tell when each finishes, or even load JSON from the same source multiple times, and be able to tell when each call finishes - even if one call is delayed so long it completes after a later call. This hard-coded approach isn't going to work anymore, for 2 reasons:
Every load of something.js calls the same loadJson() callback - you have no way of knowing which request goes with which reply.
Caching - once you load something.js once, the browser isn't going to ask the server for it again - it's going to just bring it back in from the cache, ruining your plan.
You can resolve both of these by telling the server to wrap the JSON differently each time, and the simple way is to pass that information in a querystring parameter like ?callback=loadJson12345. It's as though your page looked like this:
<script>
function loadJson1(json) {
// Read the json
}
function loadJson2(json) {
// Read the json
}
</script>
<script src="//outsidedomain.com/something.js?callback=loadJson1"></script>
<script src="//outsidedomain.com/somethingelse.js?callback=loadJson2"></script>
With JQuery, this is all abstracted for you to look like a normal call to $.ajax, meaning you're expecting the success function to fire. In order to ensure the right success function fires for each jsonp load, JQuery creates a long random callback function name in the global namespace like JQuery1233432432432432, passes that as the callback parameter in the querystring, then waits for the script to load. If everything works properly the script that loads calls the callback function JQuery requested, which in turn fires the success handler from the $.ajax call.
Note that "works properly" requires that the server-side reads the ?callback querystring parameter and includes that in the response, like ?callback=joe -> joe({.... If it's a static file or the server doesn't play this way, you likely need to treat the file as cacheable - see below.
Caching
If you wanted your json to cache, you can get JQuery to do something closer to my first example by setting cache: true and setting the jsonpCallback property to a string that is hardcoded into the cacheable json file. For example this static json:
loadJoe({name:'Joe'})
Could be loaded and cached in JQuery like so:
$.ajax({
url: '//outsidedomain.com/loadjoe.js',
dataType: 'jsonp',
cache: true,
jsonpCallback: 'loadJoe',
success: function(json) { ... }
});
Use the success callback instead..
function userinfo(username){
$.ajax({
url: "https://api.github.com/users/" + username,
success: getName,
dataType: "jsonp"
});
}
$(function() {
function userinfo(username){
var XHR = $.ajax({
url: "https://api.github.com/users/" + username,
dataType: "jsonp"
}).done(function(data) {
console.log(data.data.name);
});
}
users = ["torvalds", "twitter", "jquery"];
for(var i = 0; i < users.length; i++){
userinfo(users[i]);
}
});
Not sure but the response I get from that call to the github API does not include gravatar_id.
This worked for me:
function getGravatar(response){
var link = response.data.avatar_url;
$('#list').append('<div><img src="' + link + '"></div>');
}
I want to get live currency rates from an external source, so I found this great webservice:
Currency Convertor
This service is working like a charm, the only downside is that it does not provide JSONP results, only XML. Therefore we have a cross browser problem while trying to consume this webservice using jQuery $.ajax().
So I found Yahoo Query Language which returns results as JSONP and also mangae to consume other webservices and return me the results. This is also working, here is an example URL:
http://query.yahooapis.com/v1/public/yql?q=select%20*%20from%20xml%20where%20url%3D'http%3A%2F%2Fwww.webservicex.net%2FCurrencyConvertor.asmx%2FConversionRate%3FFromCurrency%3DNOK%26ToCurrency%3DEUR'&format=json&diagnostics=true&callback=cbfunc
This URL return JSONP result and is working like a charm, but the problem appears when I use this in my code:
$.ajax({
type: "GET",
url: urlToWebservice,
contentType: "application/json; charset=utf-8",
dataType: "jsonp",
success: function(data) {
$("#status").html("OK: " + data.text);
},
error: function(xhr, textStatus, errorThrown) {
$("#status").html("Unavailable: " + textStatus);
}
});
When I try to run this code nothing happens, and I can see this error message in my Firebug javascript debugger:
cbfunc is not defined
cbfunc is the name of the container which surrounds the JSON response, but why does it say not defined?
EDIT:
This is my new code, but I still get the cbfunc is not defined
$.ajax({
url: "http://query.yahooapis.com/v1/public/yql?q=select%20*%20from%20xml%20where%20url%3D'http%3A%2F%2Fwww.webservicex.net%2FCurrencyConvertor.asmx%2FConversionRate%3FFromCurrency%3DNOK%26ToCurrency%3DEUR'&format=json&callback=cbfunc",
dataType: 'jsonp',
jsonp: 'callback',
jsonpCallback: 'cbfunc'
});
function cbfunc(data) {
alert("OK");
}
And the "OK" message is never fired...
If available, use the jsonpCallback parameter in the call to $.ajax like:
jsonpCallback: "cbfunc",
Its description, from the jQuery API docs reads:
Specify the callback function name for a jsonp request. This value will be used instead of the random name automatically generated by jQuery.
The docs later go on to say:
It is preferable to let jQuery generate a unique name as it'll make it easier to manage the requests and provide callbacks and error handling. You may want to specify the callback when you want to enable better browser caching of GET requests.
However it is not advised to make use of this "preferable" behaviour when making use of YQL. Precisely why that approach is not ideal might make this answer far too verbose, so here is a link (from the YQL blog) detailing the problems with jQuery's preferred approach, making use of jsonpCallback and so on: Avoiding rate limits and getting banned in YQL and Pipes: Caching is your friend
You should let jQuery handle the callback by changing urlToWebservice to end in callback=?
The reason it's not working is because by specifying callback=cbfunc in the querystring generates a URL of the type:
http://query.yahooapis.com/...&callback=cbfunc&callback=jsonp1277417828303
Stripped out all uninteresting parts, but the URL contains two callback parameters. One of them is managed by jQuery, and the other one not. YQL only looks at the first callback parameter and returns a response wrapped around that.
cbfunc({"query":{...}});
However, there is no function named cbfunc in your script, so that's why you are getting the undefined error. jQuery created an implicit function named jsonp1277417828303 in the above example, and the response from YQL should instead have been:
jsonp1277417828303({"query":{...}});
for jQuery to act upon it, and return the response to your success callback which it never got to do.
So, as #SLaks suggested, remove the &callback=cbfuncfrom your URL, or replace it with &callback=? to let jQuery handle things.
See a working example.
You definitely should give jQuery-JSONP a try: http://code.google.com/p/jquery-jsonp/
Simplifies everything :)
Does anyone know what is the difference between $("#id").load and $.ajax?
Let me clarify things for you a little bit :
$.ajax() is the basic and low-level ajax function jQuery provides which means you can do what ever you want to like you can work with XmlHttpRequest object. But once upon a time jQuery Developers thought that actually besides $.ajax(), they could provide more specific methods to developers so they wouldn't need to pass more parameters to make $.ajax() method work the way they want. For example they said instead of passing json as a parameter to $.ajax() to indicate return data type, they provided $.getJSON() so we would all know that the return type we expected was json, or instead of indicating send method as post or get, you could use $.post() or $.get() respectively.
So load() is the same thing, it can help you inject html data into your html. with load() method you know that an html portion is being expected.
Isn't that cool ?
I think I've been fallen in love.
For more information, you can visit jquery.com, they are even providing their new library and api tutorial page.
Edit :
$.ajax({
type: "POST",
url: "some.php",
data: "name=John&location=Boston",
success: function(msg){
alert( "Data Saved: " + msg );
}
});
is the same as below :
$.post("some.php", { name: "John", time: "2pm" },
function(data){
alert("Data Loaded: " + data);
});
Now as you can see it is the simplified version of $.ajax(), to make post call, you need to pass some information of send method type which is post as shown at the first example but instead of doing this you can use $.post() because you know what you are doing is post so this version is more simplified and easy to work on.
But don't forget something. Except for load(), all other ajax methods return XHR (XmlHttpRequest instance) so you can treat them as if you were working with XmlHttpRequest, actually you are working with it tho :) and but load() returns jQuery which means :
$("#objectID").load("test.php", { 'choices[]': ["Jon", "Susan"] } );
in the example above, you can easly inject the return html into #objectID element. Isn't it cool ? If it wasn't returning jQuery, you should have been working with callback function where you probably get the result out of like data object and inject it manually into the html element you want. So it would be hassle but with $.load() method, it is really simplified in jQuery.
$("#feeds").load("feeds.php", {limit: 25}, function(){
alert("The last 25 entries in the feed have been loaded");
});
You can even post parameters, so according to those parameters you can do some work at server-side and send html portion back to the client and your cute jQuery code takes it and inject it into #feeds html element in the example right above.
load() initiates an Ajax request to retrieve HTML that, when returned, is set to the given selector.
All the jQuery Ajax functions are simply wrappers for $.ajax() so:
$("#id").load(...);
is probably equivalent to:
$.ajax({
url: "...",
dataType: "html",
success: function(data) {
$("#id").html(data);
}
});
A more concise summary and the most important difference is that $.ajax allows you to set content-type and datatype.
These two are important for making JSON requests, or XML requests. ASP.NET is more fussy with a missing content-type field (atleast when you use [WebMethod]) and will simply return the HTML of the page instead of JSON.
$.load() is intended to simply return straight HTML. $.ajax also gives you
caching
error handling
filtering of data
password
plus others.
From the documentation ...
$(selector).load(..)
Load HTML from a remote file and inject it into the DOM.
$.ajax(...)
Load a remote page using an HTTP request. This is jQuery's low-level AJAX implementation.
load is specifically for fetching (via GET unless parameters are provided, then POST is used) an HTML page and directly inserting it into the selected nodes (those selected by the $(selector) portion of $(selector).load(...).
$.ajax(...) is a more general method that allows you to make GET and POST requests, and does nothing specific with the response.
I encourage you to read the documentation.
Here's the source code for the load function: http://github.com/jquery/jquery/blob/master/src/ajax.js#L15
As you can see, it's a $ajax with some options handling. In other words, a convenience method.
The above answer may not be valid anymore in light of the use of deferred and promise objects. I believe that with .ajax you can use .when but you cannot do so with .load. In short, I believe that .ajax is more powerful than .load. For example:
some_promise = $.ajax({....});
.when(some_promise).done(function(){.... });
You get more granular control over the html loading. There is also .fail and .always for failure and "no matter what" cases. You don't get this in load. Hope I am correct on this.