We Keep Coding

How to track http requests/responses using JavaScript? [duplicate]

I'm trying to intercept all AJAX calls in order to check if that AJAX response contains specific error code that I send as JSON from my PHP script (codes: ACCESS_DENIED, SYSTEM_ERROR, NOT_FOUND).
I know one can do something like this:
$('.log').ajaxSuccess(function(e, xhr, settings) {
});
But - does this work only if "ajaxSuccess" event bubble up to .log div? Am I correct? Can I achieve what I want by binding "ajaxSuccess" event to document?
$(document).ajaxSuccess(function(e, xhr, settings) {
});
I can do this in either jQuery or raw JavaScript.

If you're using jQuery, $.ajaxSuccess is a good option, but here's a generic option that will intercept XHR calls from all frameworks (I've tested it with ExtJS and jQuery - it should work even if multiple frameworks are loaded concurrently). It's been tested to work with IE8, Chrome and Firefox.
(function(XHR) {
"use strict";
var open = XHR.prototype.open;
var send = XHR.prototype.send;
XHR.prototype.open = function(method, url, async, user, pass) {
this._url = url;
open.call(this, method, url, async, user, pass);
};
XHR.prototype.send = function(data) {
var self = this;
var oldOnReadyStateChange;
var url = this._url;
function onReadyStateChange() {
if(self.readyState == 4 /* complete */) {
/* This is where you can put code that you want to execute post-complete*/
/* URL is kept in this._url */
}
if(oldOnReadyStateChange) {
oldOnReadyStateChange();
}
}
/* Set xhr.noIntercept to true to disable the interceptor for a particular call */
if(!this.noIntercept) {
if(this.addEventListener) {
this.addEventListener("readystatechange", onReadyStateChange, false);
} else {
oldOnReadyStateChange = this.onreadystatechange;
this.onreadystatechange = onReadyStateChange;
}
}
send.call(this, data);
}
})(XMLHttpRequest);
I've posted a more specific example on github which intercepts AJAX calls and posts the AJAX call durations back to the server for statistical analysis.

From http://api.jquery.com/ajaxSuccess/ :
Whenever an Ajax request completes successfully, jQuery triggers the ajaxSuccess event. Any and all handlers that have been registered with the .ajaxSuccess() method are executed at this time.
So the selector doesn't define the position where you are "catching" the event (because, honestly, ajax event by its nature doesn't start from a DOM element), but rather defines a scope to which the handling will be defaulted (i.e. this will poitn to that/those element(s)).
In summary - it should be exactly what you wish for

The best way, which I found https://lowrey.me/intercept-2/
const intercept = (urlmatch, callback) => {
let send = XMLHttpRequest.prototype.send;
XMLHttpRequest.prototype.send = function() {
this.addEventListener('readystatechange', function() {
if (this.responseURL.includes(urlmatch) && this.readyState === 4) {
callback(this);
}
}, false);
send.apply(this, arguments);
};
};

Try using Mockjax.js http://code.appendto.com/plugins/jquery-mockjax
It lets you hijack AJAX calls to the server and mock the location.

JavaScript static function in callback

Hi I've been trying to clarify this but there's something I'm still confused about. I know that you can't return values from asynchronous functions so I've referenced this answer's top answer Returning value from asynchronous JavaScript method?
What I'm trying to do is use the flickrAPI to get the biggest size image. The flickrAPI allows one to search images, so I use this to get the photo_id, then I use this photo_id to procses another request to the API's getSize method to get the URL for the biggest size photo.
The code looks a little messy as it is, because I have a method called flickrRequest which sends an XMLHttp request and gets back a JSON string. I know that I can achieve what I want by writing the functions as follows:
function flickRQforimage() {
...got ID
function flickrRQforSize() {
...got maxsizeURL
create image based on maxsizeURL here
}
}
but I was wondering if it was possible to do something like this
function flickRQforimage() {
...got ID
function flickrRQforSize() {
...got maxsizeURL
}
create image based on maxsizeURL here
}
or even create image based on maxsizeURL here
In general my question is whether it is possible to have a callback function that references another statically defined function (I think?). The specifics of the my function is that it takes a callback and the ID and URL processing happens in those callbacks:
flickrRQ(options, cb)
I am wondering whether/what would happen if that unnamed function is instead something else, say flickrRQ(options, processPhoto(data)), and then I define the function in a separate method. This just makes sense for me because I want to keep functionality for the URL processing separate in an attempt to make my code cleaner and more readable.
I tried the following below and it didn't work. Nothing prints. I even have a console.log in the processPhoto method. In fact anything inside of the flickrRQforSize method seems to not evaluate
flickrRQforSize(options, function(data) {
processPhoto(data)
}
even though in the flickrRQforSize definition, a callback function is taken as an argument. I'm suspecting there must be something about functions/async calls that I don't understand.
I hope this is clear -- if not, I can post my actual code.
Here's my code:
var flickrRequest = function(options, xhrRQ, cb) {
var url, xhr, item, first;
url = "https://api.flickr.com/services/rest/";
first = true;
for (item in options) {
if (options.hasOwnProperty(item)) {
url += (first ? "?" : "&") + item + "=" + options[item];
//parses to search equest;
first = false;
}
}
//XMLHttpRQ to flickr
if(xhrRQ == 1 ) {
xhr = new XMLHttpRequest();
xhr.onload = function() { cb(this.response); };
xhr.open('get', url, true);
xhr.send();
};
}
var processPhotoSize = function(photoJSON) {
var parsedJSON = JSON.parse(data);
var last = parsedJSON.sizes.size.length;
console.log(parsedJSON.sizes.size[last-1].source);
return parsedJSON.sizes.size[last-1].source;
}
...
flickrRequest(options, 1, function(data) {
...
flickrRequest(sizesOptions, 0, function(data) {
parsedJSON = JSON.parse(data);
console.log(parsedJSON);
processPhotoSize(data);
});
}

Javascript : return XMLHttpRequest out of scope

Javascript : return XMLHttpRequest out of scope
I need to return the data from my AJAX call
series: [{
data: ( )
in order to update one of the keys data in dictionary series but my function retrieve does not seem to return the data that I am getting.
var myPopulation = {
series: [{
data: (
function() {
function retrieve() {
var httpRequest = new XMLHttpRequest();
httpRequest.onreadystatechange = function() {
if (httpRequest.readyState === 4) {
if (httpRequest.status === 200) {
var obj = JSON.parse(httpRequest.responseText)
console.log(obj.V1, obj.V2, obj.V3, obj.V4);
var data = [],
time = (new Date()).getTime(),
i;
for (i = -60; i <= 0; i++) {
console.log(obj.V2)
data.push({
x: time + i * 60 * 1000,
y: obj.V2
});
}
myPopulation.series[1].data = data
// ???
console.log(data)
}
}
};
httpRequest.open('GET', "/myCall");
httpRequest.send();
}
retrieve();
}()
)
}],
What should I do to return the value out of the function and update data?

Well, since you are using jQuery tag, I think my answer could be valid and I prefer doing this way for what you need and I understood (it is well explained so please read code comments and check browser console, this can be found at the end of the answer).
Remember that you won't be able to return a XMLHttpRequest because ajax calls are async but you can force an ajax call to be sync in order to get your data on a return statement from any function or do other things as you expected. However, forcing is not a good approach to do because UI will freeze for the user until getting the response back from server and you really don't know how much time that will take (specially if you are expecting a big amount of data to be returned - I know that's not entirely a metric but other factors may apply).
Hope this helps and please take your time and read the following post and user comments: Reasons not to use native XMLHttpRequest - why is $.ajax mandatory?
Live Demo: http://jsfiddle.net/4mbjjfx8/
HTML:
<div id="wrapper">
<div id="loader"></div>
<div id="content"></div>
</div>
jQuery
$(function() {
var series = [], // Your series array
loader = $('#loader'), // UI loader sample
request = {}; // Request params
/**
* Set request method, url and data if needed
* In this case I am sending an object with a text property
* that will be returned from jsfiddle "echo" service
*/
request.method = 'GET';
request.url = '/echo/jsonp/';
request.data = {
text: 'Second message returned from "echo" service'
};
// Send ajax call
retrieveData(request, series, handleData);
// Set loading message to UI
loader.html('Loading...');
// Just do some logging to know how process goes
console.log('Populating series for the first time');
/**
* Populate series for the first time, at this point process
* will go on and after the response from server was finally
* done, process will go to the callback (since ajax calls
* are async).
*/
populate(series);
// Just do some logging to know how process goes
console.log('End populating series for the first time');
});
function populate(series) {
var dummy = {
text: 'First message populated over process'
};
// Set dummy object to series array
series.push(dummy);
};
/**
* Used to make ajax call and return data from server
*/
function retrieveData(cfg, series, callback) {
$.ajax({
type: cfg.method,
url: cfg.url,
data: cfg.data
}).done(function(data, status, xhr) {
// Pass args to callback function if defined
if (callback) callback(series, data);
}).fail(function(xhr, status) {
/**
* Pass args to callback function if defined
* At this point, request wasn't success so
* force data arg at callback to be 'null'
*/
if (callback) callback(series, null);
});
};
/**
* Used to handle data returned from server
* Note: Your series array can be modified here since you
* passed it into the callback
*/
function handleData(series, data) {
var loader = $('#loader');
// Just do some logging to know how process goes
console.log('Populating series from server');
// Check if data is defined and not an empty object
if(data && !($.isEmptyObject(data))) {
// Add it to series array
series.push(data);
}
// Set UI loader empty
loader.html('');
// Retrieve series
showData(series);
};
function showData(series) {
var contentDiv = $('#content');
// Loop process and append to UI
for(var i = 0; i < series.length; ++i) {
contentDiv.append(series[i].text + '<br>');
}
};

You should put retrieve function outside. You can invoke retrieve function. And, It will call ajax. When ajax is success, it will update data of population. Like this.
var myPopulation = {
series: [{
data: undefined
}]
};
function retrieve() {
var httpRequest = new XMLHttpRequest();
httpRequest.onreadystatechange = function() {
if (httpRequest.readyState === 4) {
if (httpRequest.status === 200) {
var obj = JSON.parse(httpRequest.responseText)
console.log(obj.V1, obj.V2, obj.V3, obj.V4);
var data = [],
time = (new Date()).getTime(),
i;
for (i = -60; i <= 0; i++) {
console.log(obj.V2)
data.push({
x: time + i * 60 * 1000,
y: obj.V2
});
}
myPopulation.series[0].data = data
// ???
console.log(data)
}
}
};
httpRequest.open('GET', "/myCall");
httpRequest.send();
}
retrieve();

Assuming you're simply trying to set the value of myPopulation.series[0].data when the array is first defined...
myPopulation.series[1].data = data
...should be...
myPopulation.series[0].data = data;
Also, some parts of you code are missing closing semicolons, closing brackets and/or curly brackets. Please make sure you end all statements with a semicolon and you have an equal number of opening and closing (curly) brackets.
I've tested your code with the above changes. The HTTP request I made returned a simple "Test successful" string, so I've replaced the code which handles the structure of the response text to simply var data = httpRequest.responeText;. This worked fine. Of course, this assumes the code which handles the structure of the returned httpRequest.responeText in your case is correct, as I have no way of knowing what the responseText in your case looks like. If you receive any errors regarding this part of your code, we'll need to see what the responseText looks like before we can help you.

I'm not judging whether you are doing the right thing. Im merely presenting you a working version of your code.
Errors in your code:
You mean to set result of the "function" to data but your function is not returning anything in the first place.
XMLHttpRequest is async so even if you return you will not have the data set, simply because the outer function exited after making the http request setting a callback to trigger when it is completed.
Note: The fix is by making XMLHttpRequest synchronous.
https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest/Synchronous_and_Asynchronous_Requests
Here is the corrected version of your code
var myPopulation = {
series: [{
data: (
function() {
function retrieve() {
var httpRequest = new XMLHttpRequest();
var result = []; //[1] just renamed data to result to avoid confusion
httpRequest.onreadystatechange = function() {
if (httpRequest.readyState === 4) {
if (httpRequest.status === 200) {
var obj = JSON.parse(httpRequest.responseText)
console.log(obj.V1, obj.V2, obj.V3, obj.V4);
var time = (new Date()).getTime(),
i;
for (i = -60; i <= 0; i++) {
console.log(obj.V2)
result.push({
x: time + i * 60 * 1000,
y: obj.V2
});
}
//myPopulation.series[1].data = data //[2] commented this as it is not necessary
// ???
console.log(result)
}
}
};
httpRequest.open('GET', "/myCall", false); //[3] Added 3rd argument 'false' to make the call synchronous
httpRequest.send();
return result //[4] to convey the result outside
}
return retrieve(); //[5] added return to set it to the data
}()
)
}],
The above code is not tested however. Here is a tested solution http://jsfiddle.net/98f9amo8/1/
The jsfiddle content is different for obvious reasons.

Working with async code means you have to change the way you code because the code is not executed top-down any more.
So, in your case, you would do something like:
var myPopulation = {series: []};
$.get(..., my_function_that_will_format_the_data_after_they_have_been_received);
...
my_function_that_will_format_the_data_after_they_have_been_received() {
// Do stuff here
var formattedData = ...
myPopulation.series.push(formattedData);
// ONLY NOW, myPopulation is ... populated with data.
// So, whatever you use this for, need to be called here
doMagicWith(myPopulation);
}
...
/// Here, myPopulation is empty. doMagicWith(myPopulation) will fail here.

I do not know the context of how you are doing this, seeing no jQuery tells me you wish to avoid it.
So no matter what happens the call is going to take time, and you need to wait for it for whatever you may need to do with it. Loaders can help tell a user that its processing but there are other ways to do that as well. The common factor is no matter what the data is not going to be there when you need it unless you do some sort of callback.
So here is an idea, create your on onload event more or less. There are many things to keep an eye on so jQuery's is probably the most complete, but going to keep it simple here.
window.isLoaded = false;
window.ajaxLoaded = false;
window.onload = function(){
if(window.ajaxLoaded){
onReadyFunction();
}else{
window.isLoaded = true;
}
}
//skipping most of your code, the key part is the retrieve function.
//So its the only part I am going to include in this part.
function retrieve() {
var httpRequest = new XMLHttpRequest();
httpRequest.onreadystatechange = function() {
if (httpRequest.readyState === 4) {
if (httpRequest.status === 200) {
var obj = JSON.parse(httpRequest.responseText)
console.log(obj.V1, obj.V2, obj.V3, obj.V4);
var data = [],
time = (new Date()).getTime(),
i;
for (i = -60; i <= 0; i++) {
console.log(obj.V2)
data.push({
x: time + i * 60 * 1000,
y: obj.V2
});
}
myPopulation.series[1].data = data
// ???
console.log(data)
}
}
//here is my only addition
if(window.isLoaded){
onReadyFunction();
}else{
window.ajaxLoaded = true;
}
};
httpRequest.open('GET', "/myCall");
httpRequest.send();
}
So all I am doing is adding another part to the typical DOM load. Waiting for the data you need to be available before it initialized the rest of the JS. Doing this you can keep the least downtime for your app (although it depends on where you are trying to get this data though). All you need is to define the onReadyFunction like so.
function onReadyFunction(){
//all the rest of your JS here
}
This can be expanded and organized very easy, just a simple example to get started.

Monkey patch XMLHTTPRequest.onreadystatechange

How would go about monkey patching the XMLHTTPRequest's onreadystatechange function. I'm trying to add a function that would be called when every ajax request made from a page come back.
I know this sounds like a terrible idea, but the use case is quite peculiar. I want to use a certain SDK with a console (jqconsole) but show status and results from ajax calls within the console without modifying the external SDK.
I've looked at this post which had great info, but nothing on monkey patching the callback which seem to exceed my JavaScript skills.
P.S Can't use jQuery since it only supports ajax calls made from jQuery not from XMLHTTPRequests directly which is the case here.

To monkey-patch XMLHttpRequests, you need to know how an AJAX request is generally constructed:
Constructor invocation
Preparation the request (setRequestHeader(), open())
Sending the request (.send).
General-purpose patch
(function(xhr) {
function banana(xhrInstance) { // Example
console.log('Monkey RS: ' + xhrInstance.readyState);
}
// Capture request before any network activity occurs:
var send = xhr.send;
xhr.send = function(data) {
var rsc = this.onreadystatechange;
if (rsc) {
// "onreadystatechange" exists. Monkey-patch it
this.onreadystatechange = function() {
banana(this);
return rsc.apply(this, arguments);
};
}
return send.apply(this, arguments);
};
})(XMLHttpRequest.prototype);
The previous assumed that onreadystatechange was assigned to the onreadystatechange handler. For simplicity, I didn't include the code for other events, such as onload. Also, I did not account for events added using addEventListener.
The previous patch runs for all requests. But what if you want to limit the patch to a specific request only? A request with a certain URL or async flag and a specific request body?
Conditional monkey-patch
Example: Intercepting all POST requests whose request body contains "TEST"
(function(xhr) {
function banana(xhrInstance) { // Example
console.log('Monkey RS: ' + xhrInstance.readyState);
}
//
var open = xhr.open;
xhr.open = function(method, url, async) {
// Test if method is POST
if (/^POST$/i.test(method)) {
var send = this.send;
this.send = function(data) {
// Test if request body contains "TEST"
if (typeof data === 'string' && data.indexOf('TEST') >= 0) {
var rsc = this.onreadystatechange;
if (rsc) {
// Apply monkey-patch
this.onreadystatechange = function() {
banana(this);
return rsc.apply(this, arguments);
};
}
}
return send.apply(this, arguments);
};
}
return open.apply(this, arguments);
};
})(XMLHttpRequest.prototype);
The main techniques used is the transparent rewrite using...
var original = xhr.method;
xhr.method = function(){
/*...*/;
return original.apply(this, arguments);
};
My examples are very basic, and can be extended to meet your exact wishes. That's up to you, however.

Assuming you can ignore IE...
//Totally untested code, typed at the SO <textarea>... but the concept *should* work, let me know if it doesn't.
var OldXMLRequest = XMLHttpRequest;
// Create a new instance
function XMLHttpRequest() {
var ajax = new OldXMLRequest();
// save old function
var f = ajax.onreadystatechange;
ajax.onreadystatechange = function() {
console.log("Whatever!");
f(); // Call the old function
}
return ajax;
}

you can learn from Ajax-hook written by chinese!
it is a advanced js to enable Monkey patch XMLHTTPRequest

Bug in basic Ajax library

Been trying to create a basic Ajax library using JavaScript based upon a tutorial in the book "Build your own AJAX Web Applications" by Matthew Eernise (see here) as I want to get more in-depth knowledge of AJAX XML-RPC and REST. Based on the book I have created a JavaScript constructor function to get AJAX or an XMLHttpRequest going, but somehow I seem to suffer from an out-of-scope issue and the Ajax class is not defined in the following script:
function Ajax() {
// properties
this.req = null;
this.url = null;
this.method = 'GET';
this.asynch = true;
this.status = null;
this.statusText = '';
this.postData = null;
this.readyState = null;
this.responseText = null;
this.responseXML = null;
this.handleResp = null;
this.responseFormat = 'text',
// 'text', 'html', 'xml' or 'object'
this.mimeType = null;
} // End Constructor
//Create XMLHttpRequest method with XMLHttpRequest object
this.init = function() {
if (!this.req) {
try {
//Try to create objects for Firefox, Safari, IE7, etc.
this.req = newXMLHttpRequest();
}
catch(e) {
try {
//Try to create object for later versions of IE.
this.req = new ActiveXObject('MSXML2.XMLHTTP');
}
catch(e) {
try {
//Try to create for early versions of IE.
this.req = new ActiveXObject('Microsoft.XMLHTTP');
}
catch(e) {
//Could not create XMLHttpRequest object.
return false;
}
}
}
}
return this.req;
};
//Sending a Request method
this.doReq = function() {
if (!this.init()) {
alert('Could not create XMLHttpRequest object.');
return;
}
//Setting up a request
//open methods with method, url and asycn yes or no
this.req.open(this.method, this.url, this.asynch);
//Make sure mimetype is OK
if (this.mimeType) {
try {
req.overrideMimeType(this.mimeType);
}
catch(e) {
//couldn't override MIME type ... IE6 or Opera?
}
}
var self = this; // fix loss-of-scope in inner function
this.req.onreadystatechange = function() {
var resp = null;
if (self.req.readyState == 4) {
//do stuff to handle response
switch (self.reponseFormat) {
case 'text':
resp = self.req.responseText;
break;
case 'xml':
resp = self.req.responseXML;
break;
case 'object':
resp = req;
break;
}
if (self.req.status >= 200 && self.req.status <= 299) {
self.handleResp(resp);
}
else {
self.handleErr(resp);
}
}
};
this.req.send(this.postData);
};
this.handleErr = function() {
var errorWin;
try {
errorWin = window.open('', 'errorWin');
errorWin.document.body.innerHTML = this.responseText;
}
catch(e) {
alert('An error occured, but this error message cannot be '
+ 'displayed. This is probably because of your browser\'s '
+ 'pop-up blocker. \n'
+ 'Please allow pop-ups from this website if you want to '
+ 'see the full error messages. \n'
+ '\n'
+ 'Status Code: ' + this.req.status + '\n'
+ 'Status description: ' + this.req.statusText);
}
};
this.abort = function() {
if (this.req) {
this.req.onreadystatechange = function() {};
this.req.abort();
this.req = null;
}
};
this.doGet = function (url, hand, format) {
this.url = url;
this.handleResp = hand;
this.responseFormat = format || 'text' ;
this.doReq();
};
The error I get on the page that loads this script with
var hand = function (str) {
alert(str);
}
var Ajax = new Ajax(); // new instance as can ben done with PHP5 constructor classes
ajax.doGet ('/fakeserverpage.php', hand);
and starts up a new instance of Ajax get the error ajax is not defined even though I did add var self = this; // fix loss-of-scope in inner function
to fix the scope issue. What am I missing?
Update 1
Thanks to a tip here Gave new instance a different name so they don't clash:
var hand = function (str) {
alert(str);
}
var ajax = new Ajax(); // new instance as can ben done with PHP5 constructor classes
ajax.doGet ('/fakeserverpage.php', hand);
Now I am a little further. Now I get a new error: Uncaught TypeError: Object #<Ajax> has no method 'doGet'
Update 2
I tried using Ajax.prototype.init instead of this.init as recommended by a co dev here, but I still have the same error..
Update 3
Thanks to #Soufiana Hassou I improved the code by adding Ajax.prototype to all methods. Did not know it was necessary for all to work with the constructor, but it is. Code is here http://pastebin.com/g86k0z8d . I now get this pop-up saying Could not create XMLHttpRequest object. This error message is built into the method so it is working, but it cannot create the object somehow. This means there must be an error in my request for an XMLHttpRequest as I covered all cases and tested this in Firefox 11 for Mac using code on my local MacPorts MAMP. Either that or there is something else I do not know about..
Update 4
Fixed a typo. Then I got a 404 loading the fake server page. Corrected path ajax.doGet ('/ajax/fakeserverpage.php', hand); so now OK. Only I need to get the PHP to generate the code so I get an OK. The header response is OK, but I do not see the AJAX alert yet. Then I checked the console and found this error:
self.req is undefined
http://localhost/ajax/ajax.js
Line 78
See latest code: http://pastebin.com/g86k0z8d . I added some more Ajax.prototype where I thought they were still needed. Now I get:
this.req is null
http://localhost/ajax/ajax.js
Line 100
Update 5
Made some more changes removing some selfs used initially for the out-of-scope issue using var self = this. Code is still the same pastebin, but I have updated it. Now I have:
Ajax.prototype.handleResp is not a function
http://localhost/ajax/ajax.js
Line 92
Update 6
I cleaned up some of the mistakes I made in the req.onreadystatechange = function() function and now I does run. I turned of Firefox pop-up blocker for localhost and on reload it opened another tab and showed the text undefined. So almost there. No errors, just no pop-up with OK. Chrome showed a pop-up with the undefined in the body. Updated code here: http://pastebin.com/g86k0z8d as usual

You are using the same name for your instance and the class itself.
Also, you are declaring Ajax and using ajax, Javascript is case-sensitive.

First, you have var Ajax = new Ajax(); You should have var ajax = new Ajax(); instead.
Secondly, using this outside of the constructor isn't referring to the Ajax object. Try using its prototype instead:
function Ajax() {
// Set properties here
}
Ajax.prototype.init = function() {
// ...
}
See this article on Javascript classes for more information.