I have a link: Hello.
When someone clicks the link I'd like to check via JavaScript if the page the href-attribute points to exists or not. If the page exists the browser redirects to that page ("www.example.com" in this example) but if the page doesn't exist the browser should redirect to another URL.
It depends on whether the page exists on the same domain or not. If you're trying to determine if a page on an external domain exists, it won't work – browser security prevents cross-domain calls (the same-origin policy).
If it is on the same domain however, you can use jQuery like Buh Buh suggested. Although I'd recommend doing a HEAD-request instead of the GET-request the default $.ajax() method does – the $.ajax() method will download the entire page. Doing a HEAD request will only return the headers and indicate whether the page exists (response codes 200 - 299) or not (response codes 400 - 499). Example:
$.ajax({
type: 'HEAD',
url: 'http://yoursite.com/page.html',
success: function() {
// page exists
},
error: function() {
// page does not exist
}
});
See also: http://api.jquery.com/jQuery.ajax/
A pretty good work around is to proxy. If you don't have access to a server side you can use YQL. Visit: http://developer.yahoo.com/yql/console/
From there you can do something like: select * from htmlstring where url="http://google.com". You can use the "REST query" they have on that page as a starting point for your code.
Here's some code that would accept a full URL and use YQL to detect if that page exists:
function isURLReal(fullyQualifiedURL) {
var URL = encodeURIComponent(fullyQualifiedURL),
dfd = $.Deferred(),
checkURLPromise = $.getJSON('http://query.yahooapis.com/v1/public/yql?q=select%20*%20from%20htmlstring%20where%20url%3D%22' + URL + '%22&format=json');
checkURLPromise
.done(function(response) {
// results should be null if the page 404s or the domain doesn't work
if (response.query.results) {
dfd.resolve(true);
} else {
dfd.reject(false);
}
})
.fail(function() {
dfd.reject('failed');
});
return dfd.promise();
}
// usage
isURLReal('http://google.com')
.done(function(result) {
// yes, or request succeded
})
.fail(function(result) {
// no, or request failed
});
Update August 2nd, 2017
It looks like Yahoo deprecated "select * from html", although "select * from htmlstring" does work.
Based on the documentation for XMLHttpRequest:
function returnStatus(req, status) {
//console.log(req);
if(status == 200) {
console.log("The url is available");
// send an event
}
else {
console.log("The url returned status code " + status);
// send a different event
}
}
function fetchStatus(address) {
var client = new XMLHttpRequest();
client.onreadystatechange = function() {
// in case of network errors this might not give reliable results
if(this.readyState == 4)
returnStatus(this, this.status);
}
client.open("HEAD", address);
client.send();
}
fetchStatus("/");
This will however only work for URLs within the same domain as the current URL. Do you want to be able to ping external services? If so, you could create a simple script on the server which does your job for you, and use javascript to call it.
If it is in the same domain, you can make a head request with the xmlhttprequest object [ajax] and check the status code.
If it is in another domain, make an xmlhttprequest to the server and have it make the call to see if it is up.
why not just create a custom 404 handler on the web server? this is probably the more "good-bear" way to do this.
$.ajax({
url: "http://something/whatever.docx",
method: "HEAD",
statusCode: {
404: function () {
alert('not found');
},
200: function() {
alert("foundfile exists");
}
}
});
If you are happy to use jQuery you could do something like this.
When the page loads make an ajax call for each link. Then just replace the href of all the links which fail.
<script type="text/javascript" src="http://ajax.googleapis.com/ajax/libs/jquery/1.4/jquery.min.js"></script>
<script type="text/javascript">
<!--
$.fn.checkPageExists = function(defaultUrl){
$.each(this, function(){
var $link = $(this);
$.ajax({
url: $link.attr("href"),
error: function(){
$link.attr("href", defaultUrl);
}
});
});
};
$(document).ready(function(){
$("a").checkPageExists("default.html");
});
//-->
</script>
You won't be able to use an ajax call to ping the website because of same-origin policy.
The best way to do it is to use an image and if you know the website you are calling has a favicon or some sort of icon to grab, you can just use an html image tag and use the onerror event.
Example:
function pingImgOnWebsite(url) {
var img = document.createElement('img');
img.style.visibility = 'hidden';
img.style.position = 'fixed';
img.src = url;
img.onerror = continueBtn; // What to do on error function
document.body.appendChild(img);
}
Another way to do this is is with PHP.
You could add
<?php
if (file_exists('/index.php'))
{
$url = '/index.php';
} else {
$url = '/notindex.php';
}
?>
And then
<a href="<?php echo $url; ?>Link</a>
Related
I'm making a web app that requires that I check to see if remote servers are online or not. When I run it from the command line, my page load goes up to a full 60s (for 8 entries, it will scale linearly with more).
I decided to go the route of pinging on the user's end. This way, I can load the page and just have them wait for the "server is online" data while browsing my content.
If anyone has the answer to the above question, or if they know a solution to keep my page loads fast, I'd definitely appreciate it.
I have found someone that accomplishes this with a very clever usage of the native Image object.
From their source, this is the main function (it has dependences on other parts of the source but you get the idea).
function Pinger_ping(ip, callback) {
if(!this.inUse) {
this.inUse = true;
this.callback = callback
this.ip = ip;
var _that = this;
this.img = new Image();
this.img.onload = function() {_that.good();};
this.img.onerror = function() {_that.good();};
this.start = new Date().getTime();
this.img.src = "http://" + ip;
this.timer = setTimeout(function() { _that.bad();}, 1500);
}
}
This works on all types of servers that I've tested (web servers, ftp servers, and game servers). It also works with ports. If anyone encounters a use case that fails, please post in the comments and I will update my answer.
Update: Previous link has been removed. If anyone finds or implements the above, please comment and I'll add it into the answer.
Update 2: #trante was nice enough to provide a jsFiddle.
http://jsfiddle.net/GSSCD/203/
Update 3: #Jonathon created a GitHub repo with the implementation.
https://github.com/jdfreder/pingjs
Update 4: It looks as if this implementation is no longer reliable. People are also reporting that Chrome no longer supports it all, throwing a net::ERR_NAME_NOT_RESOLVED error. If someone can verify an alternate solution I will put that as the accepted answer.
Ping is ICMP, but if there is any open TCP port on the remote server it could be achieved like this:
function ping(host, port, pong) {
var started = new Date().getTime();
var http = new XMLHttpRequest();
http.open("GET", "http://" + host + ":" + port, /*async*/true);
http.onreadystatechange = function() {
if (http.readyState == 4) {
var ended = new Date().getTime();
var milliseconds = ended - started;
if (pong != null) {
pong(milliseconds);
}
}
};
try {
http.send(null);
} catch(exception) {
// this is expected
}
}
you can try this:
put ping.html on the server with or without any content, on the javascript do same as below:
<script>
function ping(){
$.ajax({
url: 'ping.html',
success: function(result){
alert('reply');
},
error: function(result){
alert('timeout/error');
}
});
}
</script>
You can't directly "ping" in javascript.
There may be a few other ways:
Ajax
Using a java applet with isReachable
Writing a serverside script which pings and using AJAX to communicate to your serversidescript
You might also be able to ping in flash (actionscript)
You can't do regular ping in browser Javascript, but you can find out if remote server is alive by for example loading an image from the remote server. If loading fails -> server down.
You can even calculate the loading time by using onload-event. Here's an example how to use onload event.
Pitching in with a websocket solution...
function ping(ip, isUp, isDown) {
var ws = new WebSocket("ws://" + ip);
ws.onerror = function(e){
isUp();
ws = null;
};
setTimeout(function() {
if(ws != null) {
ws.close();
ws = null;
isDown();
}
},2000);
}
Update: this solution does not work anymore on major browsers, since the onerror callback is executed even if the host is a non-existent IP address.
To keep your requests fast, cache the server side results of the ping and update the ping file or database every couple of minutes(or however accurate you want it to be). You can use cron to run a shell command with your 8 pings and write the output into a file, the webserver will include this file into your view.
The problem with standard pings is they're ICMP, which a lot of places don't let through for security and traffic reasons. That might explain the failure.
Ruby prior to 1.9 had a TCP-based ping.rb, which will run with Ruby 1.9+. All you have to do is copy it from the 1.8.7 installation to somewhere else. I just confirmed that it would run by pinging my home router.
There are many crazy answers here and especially about CORS -
You could do an http HEAD request (like GET but without payload).
See https://ochronus.com/http-head-request-good-uses/
It does NOT need a preflight check, the confusion is because of an old version of the specification, see
Why does a cross-origin HEAD request need a preflight check?
So you could use the answer above which is using the jQuery library (didn't say it) but with
type: 'HEAD'
--->
<script>
function ping(){
$.ajax({
url: 'ping.html',
type: 'HEAD',
success: function(result){
alert('reply');
},
error: function(result){
alert('timeout/error');
}
});
}
</script>
Off course you can also use vanilla js or dojo or whatever ...
If what you are trying to see is whether the server "exists", you can use the following:
function isValidURL(url) {
var encodedURL = encodeURIComponent(url);
var isValid = false;
$.ajax({
url: "http://query.yahooapis.com/v1/public/yql?q=select%20*%20from%20html%20where%20url%3D%22" + encodedURL + "%22&format=json",
type: "get",
async: false,
dataType: "json",
success: function(data) {
isValid = data.query.results != null;
},
error: function(){
isValid = false;
}
});
return isValid;
}
This will return a true/false indication whether the server exists.
If you want response time, a slight modification will do:
function ping(url) {
var encodedURL = encodeURIComponent(url);
var startDate = new Date();
var endDate = null;
$.ajax({
url: "http://query.yahooapis.com/v1/public/yql?q=select%20*%20from%20html%20where%20url%3D%22" + encodedURL + "%22&format=json",
type: "get",
async: false,
dataType: "json",
success: function(data) {
if (data.query.results != null) {
endDate = new Date();
} else {
endDate = null;
}
},
error: function(){
endDate = null;
}
});
if (endDate == null) {
throw "Not responsive...";
}
return endDate.getTime() - startDate.getTime();
}
The usage is then trivial:
var isValid = isValidURL("http://example.com");
alert(isValid ? "Valid URL!!!" : "Damn...");
Or:
var responseInMillis = ping("example.com");
alert(responseInMillis);
const ping = (url, timeout = 6000) => {
return new Promise((resolve, reject) => {
const urlRule = new RegExp('(https?|ftp|file)://[-A-Za-z0-9+&##/%?=~_|!:,.;]+[-A-Za-z0-9+&##/%=~_|]');
if (!urlRule.test(url)) reject('invalid url');
try {
fetch(url)
.then(() => resolve(true))
.catch(() => resolve(false));
setTimeout(() => {
resolve(false);
}, timeout);
} catch (e) {
reject(e);
}
});
};
use like this:
ping('https://stackoverflow.com/')
.then(res=>console.log(res))
.catch(e=>console.log(e))
I don't know what version of Ruby you're running, but have you tried implementing ping for ruby instead of javascript? http://raa.ruby-lang.org/project/net-ping/
let webSite = 'https://google.com/'
https.get(webSite, function (res) {
// If you get here, you have a response.
// If you want, you can check the status code here to verify that it's `200` or some other `2xx`.
console.log(webSite + ' ' + res.statusCode)
}).on('error', function(e) {
// Here, an error occurred. Check `e` for the error.
console.log(e.code)
});;
if you run this with node it would console log 200 as long as google is not down.
You can run the DOS ping.exe command from javaScript using the folowing:
function ping(ip)
{
var input = "";
var WshShell = new ActiveXObject("WScript.Shell");
var oExec = WshShell.Exec("c:/windows/system32/ping.exe " + ip);
while (!oExec.StdOut.AtEndOfStream)
{
input += oExec.StdOut.ReadLine() + "<br />";
}
return input;
}
Is this what was asked for, or am i missing something?
just replace
file_get_contents
with
$ip = $_SERVER['xxx.xxx.xxx.xxx'];
exec("ping -n 4 $ip 2>&1", $output, $retval);
if ($retval != 0) {
echo "no!";
}
else{
echo "yes!";
}
It might be a lot easier than all that. If you want your page to load then check on the availability or content of some foreign page to trigger other web page activity, you could do it using only javascript and php like this.
yourpage.php
<?php
if (isset($_GET['urlget'])){
if ($_GET['urlget']!=''){
$foreignpage= file_get_contents('http://www.foreignpage.html');
// you could also use curl for more fancy internet queries or if http wrappers aren't active in your php.ini
// parse $foreignpage for data that indicates your page should proceed
echo $foreignpage; // or a portion of it as you parsed
exit(); // this is very important otherwise you'll get the contents of your own page returned back to you on each call
}
}
?>
<html>
mypage html content
...
<script>
var stopmelater= setInterval("getforeignurl('?urlget=doesntmatter')", 2000);
function getforeignurl(url){
var handle= browserspec();
handle.open('GET', url, false);
handle.send();
var returnedPageContents= handle.responseText;
// parse page contents for what your looking and trigger javascript events accordingly.
// use handle.open('GET', url, true) to allow javascript to continue executing. must provide a callback function to accept the page contents with handle.onreadystatechange()
}
function browserspec(){
if (window.XMLHttpRequest){
return new XMLHttpRequest();
}else{
return new ActiveXObject("Microsoft.XMLHTTP");
}
}
</script>
That should do it.
The triggered javascript should include clearInterval(stopmelater)
Let me know if that works for you
Jerry
You could try using PHP in your web page...something like this:
<html><body>
<form method="post" name="pingform" action="<?php echo $_SERVER['PHP_SELF']; ?>">
<h1>Host to ping:</h1>
<input type="text" name="tgt_host" value='<?php echo $_POST['tgt_host']; ?>'><br>
<input type="submit" name="submit" value="Submit" >
</form></body>
</html>
<?php
$tgt_host = $_POST['tgt_host'];
$output = shell_exec('ping -c 10 '. $tgt_host.');
echo "<html><body style=\"background-color:#0080c0\">
<script type=\"text/javascript\" language=\"javascript\">alert(\"Ping Results: " . $output . ".\");</script>
</body></html>";
?>
This is not tested so it may have typos etc...but I am confident it would work. Could be improved too...
UPDATE: I failed to mention that I already read the suggested duplicate answer and it didn't get me any further as I didn't see how it related to my problem. The user ponury-kostek below, did however manage to explain it simply enough without all that clutter, for me to understand. So that's how I don't see it as a duplicate.
I'm trying to implement saving user data into a database when the user logs in with LinkedIn (to keep track of who watched my page). I found a tutorial that used jQuery, and I found a GitHub (here) page for conversion of jQuery to Vanilla JS, but I'm struggling to understand what I need to do to convert this specific statement.
I got the whole thing working using just that one line of jQuery, no problems - but I don't want to force users to load the jQuery lib!
I'll post the jQuery I'm trying to convert, the Vanilla JS solution I have so far, and the conversion "formula" suggested on the GitHub page:
jQuery I'm trying to convert:
$.post('saveUserData.php',
{
oauth_provider: 'linkedin',
userData: JSON.stringify(userData)
},
function(data){ return true; });
My attempt at a Vanilla JS solution
var theUrl = 'saveUserData.php';
var httpRequest = new XMLHttpRequest();
httpRequest.onreadystatechange = function (data) {
};
httpRequest.setRequestHeader('Content-Type', 'application/x-www-form-urlencoded');
httpRequest.open('POST', theUrl);
httpRequest.send({oauth_provider:'linkedin',userData: JSON.stringify(userData)}, function(data){ return true; });
Error thrown:
script.js:10 Uncaught DOMException: Failed to execute 'setRequestHeader' on 'XMLHttpRequest': The object's state must be OPENED.
at saveUserData (http://localhost:8012/linkedCV/script.js:10:14)
at displayProfileData (http://localhost:8012/linkedCV/index.php:43:4)
at B.<anonymous> (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:3350:17)
at B.runHandler (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:172:9)
at B.<anonymous> (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:3355:6)
at B.handleSuccessResults (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:172:9)
at Object.g [as success] (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:3243:4)
at Object.incoming (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:817:38)
at _window_onMessage (https://platform.linkedin.com/js/framework?v=1.0.350-1429&lang=undefined:581:102)
My JS (in the index header):
<script type="text/javascript" src="//platform.linkedin.com/in.js">
api_key: thecorrectAPIkey aka 'Client ID'
authorize: true
onLoad: onLinkedInLoad
scope: r_basicprofile r_emailaddress
</script>
<script type="text/javascript">
// Setup an event listener to make an API call once auth is complete
function onLinkedInLoad() {
IN.Event.on(IN, "auth", getProfileData);
}
// Use the API call wrapper to request the member's profile data
function getProfileData() {
IN.API.Profile("me").fields("id", "first-name", "last-name", "headline", "location", "picture-url", "public-profile-url", "email-address").result(displayProfileData).error(onError);
}
// Handle the successful return from the API call
function displayProfileData(data){
var user = data.values[0];
document.getElementById("picture").innerHTML = '<img src="'+user.pictureUrl+'" />';
document.getElementById("name").innerHTML = user.firstName+' '+user.lastName;
document.getElementById("intro").innerHTML = user.headline;
document.getElementById("email").innerHTML = user.emailAddress;
document.getElementById("location").innerHTML = user.location.name;
document.getElementById("link").innerHTML = 'Visit profile';
document.getElementById('profileData').style.display = 'block';
saveUserData(user);
}
// Handle an error response from the API call
function onError(error) {
console.log(error);
}
// Destroy the session of linkedin
function logout(){
IN.User.logout(removeProfileData);
}
// Remove profile data from page
function removeProfileData(){
document.getElementById('profileData').remove();
}
</script>
GitHub conversion suggestion:
// jQuery
$.post('//example.com', { username: username }, function (data) {
// code
})
// Vanilla
var httpRequest = new XMLHttpRequest()
httpRequest.onreadystatechange = function (data) {
// code
}
httpRequest.setRequestHeader('Content-Type', 'application/x-www-form-urlencoded')
httpRequest.open('POST', url)
httpRequest.send('username=' + encodeURIComponent(username))
Since this works perfectly as long as I use the suggested jQuery (the one I want to convert to Vanilla JS), it all works fine. So I'm going to assume the rest of the code of my page is not needed (the PHP for the DB connection and for saving user data to the DB).
https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest/setRequestHeader
When using setRequestHeader(), you must call it after calling open(),
but before calling send().
var theUrl = 'saveUserData.php';
var httpRequest = new XMLHttpRequest();
httpRequest.onreadystatechange = function (data) {
};
httpRequest.open('POST', theUrl);
httpRequest.setRequestHeader('Content-Type', 'application/x-www-form-urlencoded');
httpRequest.send({oauth_provider:'linkedin',userData: JSON.stringify(userData)}, function(data){ return true; });
<script>
window.addEventListener('load',function(){
var unique_code="3412313ad"// Initialize it with the unique code provided to you.
var param1="1"; // Initialize this with the value that you wish to see.For example 1 for navbar display , 2 for the side floating pop up
//while 3 for a transparent overlay on the whole page.
var domain=window.location.hostname;// current domain.
function jsonp(url, callback) {
var callbackName = 'jsonp_callback_' + Math.round(100000 * Math.random());
window[callbackName] = function(data) {
delete window[callbackName];
document.body.removeChild(script);
callback(data);
};
var script = document.createElement('script');
script.src = url + (url.indexOf('?') >= 0 ? '&' : '?') + 'callback=' + callbackName;
document.body.appendChild(script);
script.onerror=function(){
alert("failed to load snippet!");
}
}
jsonp('http://localhost/server.php?unique_code='+unique_code+'&domain='+domain, function(data) {
alert(data);
if(data.status=='success'){
alert('success');
}else alert(data.reason);
});
});
</script>
This is a code that mimics jsonp of the jquery to get a script from the remote server.
I used the answer given in this question JavaScript XMLHttpRequest using JsonP
Server side code would be
if(isset($_GET['unique_code']) && !empty($_GET['unique_code']) && isset($_GET['domain']) && !empty($_GET['domain'])){
$unique_code=$_GET['unique_code'];
$domain=$_GET['domain'];
$statement=$mysqli->prepare('select * from `snippet_users` where unique_code=? AND domain=?');
$statement->bind_param('ss',$unique_code,$domain);
if(!$statement->execute())
die(json_encode(array('status'=>'error','reason'=>'Server error.')));
$result=$statement->get_result();
if(mysqli_num_rows($result)>0)
die (json_encode(array('status'=>'success')));
else die(json_encode(array('status'=>'error','reason'=>'Unique code/Domain error.')));
}else{
die(json_encode(array('status'=>'error','reason'=>'Unique code/Domain error.')));
}
Everything is working perfectly fine but i see error in the console , somewhat like this :
What would be my solution so that i dont get this error as well as i get my data in the alert box?
You are outputting application/json instead of application/javascript, so your browser thinks it's not valid. The json should be in a function call (callback parameter). The callback parameter should be validated on the server side however to prevent xss injection:
Is it necessary to validate or escape the jsonp callback string
If we do window.location = "http://MyApi.com/Pdf";, browser does a GET of the URL http://MyApi.com/Pdf. But if we want to set authentication header of the request before doing GET of the URL because the server is a REST server and it doesn't support cookies. How to do this?
In all of the cases, I'm using $.ajax to call service but this time I need to show the response in a new window. Response is a PDF file content.
Thanks in advance.
In more recent browsers, you might be able to use blobs:
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<button onclick="tryit();">PDF</button>
<script>
function tryit() {
var win = window.open('_blank');
downloadFile('/pdf', function(blob) {
var url = URL.createObjectURL(blob);
win.location = url;
});
}
function downloadFile(url, success) {
var xhr = new XMLHttpRequest();
xhr.open('GET', url, true);
xhr.setRequestHeader("Authorization", "Basic " + btoa("username:password"));
xhr.responseType = "blob";
xhr.onreadystatechange = function() {
if (xhr.readyState == 4) {
if (success) success(xhr.response);
}
};
xhr.send(null);
}
</script>
</body>
</html>
In IE, ask the user:
window.navigator.msSaveOrOpenBlob(blob, 'readme.pdf');
P.S.
You can test the backend in Node:
router.get('/pdf', function(req, res) {
if(req.headers.authorization !== 'Basic dXNlcm5hbWU6cGFzc3dvcmQ=') return res.status(403).send('Not allowed');
res.sendFile(path.join(__dirname, 'public', 'render.pdf'));
});
router.get('/', function(req, res) {
res.render('index');
});
I think this is what you are looking for... Or correct me if i am wrong.
https://developer.mozilla.org/en/docs/Setting_HTTP_request_headers
If you don't care about hiding or obfuscating the user credentials then just use plain GET authentification:
use http://username:password#MyApi.com/ instead of http://MyApi.com/
Does it have to be a GET?
The reason I am asking is that you could just have a POST form (to a target="_BLANK") that posts whatever but shows an embedded file in a new window. Of course this wouldn't solve the issue with your custom headers, but then since you can also POST using jquery.ajax - which does allow you to set your own headers - you'd have the best of both worlds.
Here's a jQuery plugin that creates such a form dynamically in order to download whichever file. You could use this as a reference...
Hope this helps
You may consider setting the header in beforeunload or onunload event handler
You should configure $.ajax using beforeSend. Below an example, but of course I don't know if the exact setup will work for you without any code to look at.
$.ajax( {
url : '/model/user.json',
dataType : 'json',
'beforeSend' : function(xhr) {
var bytes = Crypto.charenc.Binary.stringToBytes(username + ":" + password);
var base64 = Crypto.util.bytesToBase64(bytes);
xhr.setRequestHeader("Authorization", "Basic " + base64);
},
error : function(xhr, ajaxOptions, thrownError) {
reset();
onError('Invalid username or password. Please try again.');
$('#loginform #user_login').focus();
},
success : function(model) {
cookies();
...
}
});
For this to work you need crypto-js.
I want to reload an image on a page if it has been updated on the server. In other questions it has been suggested to do something like
newImage.src = "http://localhost/image.jpg?" + new Date().getTime();
to force the image to be re-loaded, but that means that it will get downloaded again even if it really hasn't changed.
Is there any Javascript code that will cause a new request for the same image to be generated with a proper If-Modified-Since header so the image will only be downloaded if it has actually changed?
UPDATE: I'm still confused: if I just request the typical URL, I'll get the locally cached copy. (unless I make the server mark it as not cacheable, but I don't want to do that because the whole idea is to not re-download it unless it really changes.) if I change the URL, I'll always re-download, because the point of the new URL is to break the cache. So how do I get the in-between behavior I want, i.e. download the file only if it doesn't match the locally cached copy?
Javascript can't listen for an event on the server. Instead, you could employ some form of long-polling, or sequential calls to the server to see if the image has been changed.
You should have a look at the xhr.setRequestHeader() method. It's a method of any XMLHttpRequest object, and can be used to set headers on your Ajax queries. In jQuery, you can easily add a beforeSend property to your ajax object and set up some headers there.
That being said, caching with Ajax can be tricky. You might want to have a look at this thread on Google Groups, as there's a few issues involved with trying to override a browser's caching mechanisms. You'll need to ensure that your server is returning the proper cache control headers in order to be able to get something like this to work.
One way of doing this is to user server-sent events to have the server push a notification whenever the image has been changed. For this you need a server-side script that will periodically check for the image having been notified. The server-side script below ensures that the server sends an event at least once every (approximately) 60 seconds to prevent timeouts and the client-side HTML handles navigation away from and to the page:
sse.py
#!/usr/bin/env python3
import time
import os.path
print("Content-Type: text/event-stream\n\n", end="")
IMG_PATH = 'image.jpg'
modified_time = os.path.getmtime(IMG_PATH)
seconds_since_last_send = 0
while True:
time.sleep(1)
new_modified_time = os.path.getmtime(IMG_PATH)
if new_modified_time != modified_time:
modified_time = new_modified_time
print('data: changed\n\n', end="", flush=True)
seconds_since_last_send = 0
else:
seconds_since_last_send += 1
if seconds_since_last_send == 60:
print('data: keep-alive\n\n', end="", flush=True)
seconds_since_last_send = 0
And then your HTML would include some JavaScript code:
sse.html
<html>
<head>
<meta charset="UTF-8">
<title>Server-sent events demo</title>
</head>
<body>
<img id="img" src="image.jpg">
<script>
const img = document.getElementById('img');
let evtSource = null;
function setup_sse()
{
console.log('Creating new EventSource.');
evtSource = new EventSource('sse.py');
evtSource.onopen = function() {
console.log('Connection to server opened.');
};
// if we navigate away from this page:
window.onbeforeunload = function() {
console.log('Closing connection.');
evtSource.close();
evtSource = null;
};
evtSource.onmessage = function(e) {
if (e.data == 'changed')
img.src = 'image.jpg?version=' + new Date().getTime();
};
evtSource.onerror = function(err) {
console.error("EventSource failed:", err);
};
}
window.onload = function() {
// if we navigate back to this page:
window.onfocus = function() {
if (!evtSource)
setup_sse();
};
setup_sse(); // first time
};
</script>
</body>
</html>
Here am loading an image, tree.png, as binary data dynamically with AJAX and saving the Last-Modified header. Periodically (every 5 second in the code below). I issue another download request sending backup a If-Modified-Since header using the saved last-modified header. I check to see if data has been returned and re-create the image with the data if present:
<!doctype html>
<html>
<head>
<title>Test</title>
<script>
window.onload = function() {
let image = document.getElementById('img');
var lastModified = ''; // 'Sat, 11 Jun 2022 19:15:43 GMT'
function _arrayBufferToBase64(buffer) {
var binary = '';
var bytes = new Uint8Array(buffer);
var len = bytes.byteLength;
for (var i = 0; i < len; i++) {
binary += String.fromCharCode(bytes[i]);
}
return window.btoa( binary );
}
function loadImage()
{
var request = new XMLHttpRequest();
request.open("GET", "tree.png", true);
if (lastModified !== '')
request.setRequestHeader("If-Modified-Since", lastModified);
request.responseType = 'arraybuffer';
request.onload = function(/* oEvent */) {
lastModified = request.getResponseHeader('Last-Modified');
var response = request.response;
if (typeof response !== 'undefined' && response.byteLength !== 0) {
var encoded = _arrayBufferToBase64(response);
image.src = 'data:image/png;base64,' + encoded;
}
window.setTimeout(loadImage, 5000);
};
request.send();
}
loadImage();
};
</script>
</head>
<body>
<img id="img">
</body>
</html>
You can write a server side method which just returns last modified date of the image resource,
Then you just use polling to check for the modified date and then reload if modified date is greater than previous modified date.
pseudo code (ASP.NET)
//server side ajax method
[WebMethod]
public static string GetModifiedDate(string resource)
{
string path = HttpContext.Current.Server.MapPath("~" + resource);
FileInfo f = new FileInfo(path);
return f.LastWriteTimeUtc.ToString("yyyy-dd-MMTHH:mm:ss", CultureInfo.InvariantCulture);//2020-05-12T23:50:21
}
var pollingInterval = 5000;
function getPathFromUrl(url) {
return url.split(/[?#]/)[0];
}
function CheckIfChanged() {
$(".img").each(function (i, e) {
var $e = $(e);
var jqxhr = $.ajax({
type: "POST",
contentType: "application/json; charset=utf-8",
url: "/Default.aspx/GetModifiedDate",
data: "{'resource':'" + getPathFromUrl($e.attr("src")) + "'}"
}).done(function (data, textStatus, jqXHR) {
var dt = jqXHR.responseJSON.d;
var dtCurrent = $e.attr("data-lastwrite");
if (dtCurrent) {
var curDate = new Date(dtCurrent);
var dtLastWrite = new Date(dt);
//refresh if modified date is higher than current date
if (dtLastWrite > curDate) {
$e.attr("src", getPathFromUrl($e.attr("src")) + "?d=" + new Date());//fool browser with date querystring to reload image
}
}
$e.attr("data-lastwrite", dt);
});
}).promise().done(function () {
window.setTimeout(CheckIfChanged, pollingInterval);
});
}
$(document).ready(function () {
window.setTimeout(CheckIfChanged, pollingInterval);
});
<script src="https://cdnjs.cloudflare.com/ajax/libs/jquery/3.3.1/jquery.min.js"></script>
<img class="img" src="/img/rick.png" alt="rick" />
If you are going to check whether files has changed on the server you have to make http request from the server for the file time, because there is no other way for your check the file time once page get loaded to the browser.
So that time check script will like
filetimecheck.php
<?php
echo filemtime(string $filename);
?>
Then you can check the file time using your Javascript. BTW I have put jQuery $.get for check the file time.
dusplayimage.php
<img id="badge" src="image.jpg"> />
<script>
var image_time = <?php echo filemtime(string $filename); ?>;
var timerdelay = 5000;
function imageloadFunction(){
$.get("filetimecheck.php", function(data, status){
console.log("Data: " + data + "\nStatus: " + status);
if(image_time < parseInt(data)) {
document.getElementById('yourimage').src = "image.jpg?random="+new Date().getTime();
}
});
setTimeout(imageloadFunction, timerdelay);
}
imageloadFunction();
</script>
You will be using extra call to the server to check the file time which you can't avoid however you can use the time delay to fine-tune the polling time.
Yes, you can customize this behavior. Even with virtually no change to your client code.
So, you will need a ServiceWorker (caniuse 96.59%).
ServiceWorker can proxy your http requests. Also, ServiceWorker has already built-in storage for the cache. If you have not worked with ServiceWorker, then you need to study it in detail.
The idea is the following:
When requesting a picture (in fact, any file), check the cache.
If there is no such picture in the cache, send a request and fill the cache storage with the date of the request and the file.
If the cache contains the required file, then send only the date and path of the file to the special API to the server.
The API returns either the file and modification date at once (if the file was updated), or the response that the file has not changed {"changed": false}.
Then, based on the response, the worker either writes a new file to the cache and resolves the request with the new file, or resolves the request with the old file from the cache.
Here is an example code (not working, but for understanding)
s-worker.js
self.addEventListener('fetch', (event) => {
if (event.request.method !== 'GET') return;
event.respondWith(
(async function () {
const cache = await caches.open('dynamic-v1');
const cachedResponse = await cache.match(event.request);
if (cachedResponse) {
// check if a file on the server has changed
const isChanged = await fetch('...');
if (isChanged) {
// give file, and in the background write to the cache
} else {
// return data
}
return cachedResponse;
} else {
// request data, send from the worker and write to the cache in the background
}
})()
);
});
In any case, look for "ways to cache statics using ServiceWorker" and change the examples for yourself.
WARNING this solution is like taking a hammer to crush a fly
You can use sockets.io to pull information to browser.
In this case you need to monitor image file changes on the server side, and then if change occur emit an event to indicate the file change.
On client (browser) side listen to the event and then then refresh image each time you get the event.
set your image source in a data-src property,
and use javascript to periodicaly set it to the src attribute of that image with a anchor (#) the anchor tag in the url isn't send to the server.
Your webserver (apache / nginx) should respond with a HTTP 304 if the image wasn't changed, or a 200 OK with the new image in the body, if it was
setInterval(function(){
l= document.getElementById('logo');
l.src = l.dataset.src+'#'+ new Date().getTime();
},1000);
<img id="logo" alt="awesome-logo" data-src="https://upload.wikimedia.org/wikipedia/commons/1/11/Test-Logo.svg" />
EDIT
Crhome ignores http cache-control headers, for subsequent image reloads.
but the fetch api woks as expected
fetch('https://upload.wikimedia.org/wikipedia/commons/1/11/Test-Logo.svg', { cache: "no-cache" }).then(console.log);
the no-cache instructs the browser to always revalidate with the server, and if the server responds with 304, use the local cached version.