Why my REST API clients need JSONP request? - c#

I created REST base webAPI in MVC 4 and hosted on server when I call this from HTML on other domain and on my local pc I need to call it as JSONP request like put callback=? in url so it can be jsonp. My question is that why this is so? if its due to cross domain then how google and facbook and other companies host their api we also call it from our own domain but we dont keep callback=? in their url.
so why my API need callback=? in url if i call it from other domain or on my local pc with simple jquery html.

Its because of the Same Origin Policy imposed by the browsers.
See
http://www.w3.org/Security/wiki/Same_Origin_Policy
http://en.wikipedia.org/wiki/Same_origin_policy
.
Also note that CORS might be a better option than JSONP in the future
http://en.wikipedia.org/wiki/Cross-origin_resource_sharing
EDIT: ------------
If you have gone through above links you would see that JSONP allows users to work-around this Same Origin Policy security measure imposed by the browsers.
Trick is browsers allow tags to refer files in other domains than the origin.
Basically what happens is with JSONP, you send a callback function name to the server appended to the query string. Then the server will pad or prefix it's otherwise JSON request with a call to this function, hence the P in the name to denote response is padded or prefixed.
For example you can create a script tag like
then the target server, should send a response such that
mymethod({normal: 'json response'})
when this repsone is evaluated on the client side (as for any other javascript file) it will effectively call your method with the JSON response from that server.
However, this can only do GET requests.
If you want to make POST (PUT/DELETE) requests you need to use CORS in which server needs to set a specific header beforehand.
Access-Control-Allow-Origin: www.ext.site.com
Hope this helps.

Because of the same-origin policy limitations. The same-origin policy prevents a script loaded from one domain from getting or manipulating properties of a document from another domain. That is, the domain of the requested URL must be the same as the domain of the current Web page. This basically means that the browser isolates content from different origins to guard them against manipulation.

Related

Browser debugger says my response.body is null but fiddler and the developer tools show it as populated [duplicate]

I can hit this endpoint, http://catfacts-api.appspot.com/api/facts?number=99 via Postman and it returns JSON
Additionally I am using create-react-app and would like to avoid setting up any server config.
In my client code I am trying to use fetch to do the same thing, but I get the error:
No 'Access-Control-Allow-Origin' header is present on the requested
resource. Origin 'http://localhost:3000' is therefore not allowed
access. If an opaque response serves your needs, set the request's
mode to 'no-cors' to fetch the resource with CORS disabled.
So I am trying to pass in an object, to my Fetch which will disable CORS, like so:
fetch('http://catfacts-api.appspot.com/api/facts?number=99', { mode: 'no-cors'})
.then(blob => blob.json())
.then(data => {
console.table(data);
return data;
})
.catch(e => {
console.log(e);
return e;
});
Interestingly enough the error I get is actually a syntax error with this function. I am not sure my actual fetch is broken, because when I remove the { mode: 'no-cors' } object, and supply it with a different URL it works just fine.
I have also tried to pass in the object { mode: 'opaque'} , but this returns the original error from above.
I belive all I need to do is disable CORS.. What am I missing?
mode: 'no-cors' won’t magically make things work. In fact it makes things worse, because one effect it has is to tell browsers, “Block my frontend JavaScript code from seeing contents of the response body and headers under all circumstances.” Of course you never want that.
What happens with cross-origin requests from frontend JavaScript is that browsers by default block frontend code from accessing resources cross-origin. If Access-Control-Allow-Origin is in a response, then browsers relax that blocking and allow your code to access the response.
But if a site sends no Access-Control-Allow-Origin in its responses, your frontend code can’t directly access responses from that site. In particular, you can’t fix it by specifying mode: 'no-cors' (in fact that’ll ensure your frontend code can’t access the response contents).
However, one thing that will work: if you send your request through a CORS proxy.
You can also easily deploy your own proxy to Heroku in just 2-3 minutes, with 5 commands:
git clone https://github.com/Rob--W/cors-anywhere.git
cd cors-anywhere/
npm install
heroku create
git push heroku master
After running those commands, you’ll end up with your own CORS Anywhere server running at, for example, https://cryptic-headland-94862.herokuapp.com/.
Prefix your request URL with your proxy URL; for example:
https://cryptic-headland-94862.herokuapp.com/https://example.com
Adding the proxy URL as a prefix causes the request to get made through your proxy, which:
Forwards the request to https://example.com.
Receives the response from https://example.com.
Adds the Access-Control-Allow-Origin header to the response.
Passes that response, with that added header, back to the requesting frontend code.
The browser then allows the frontend code to access the response, because that response with the Access-Control-Allow-Origin response header is what the browser sees.
This works even if the request is one that triggers browsers to do a CORS preflight OPTIONS request, because in that case, the proxy also sends back the Access-Control-Allow-Headers and Access-Control-Allow-Methods headers needed to make the preflight successful.
I can hit this endpoint, http://catfacts-api.appspot.com/api/facts?number=99 via Postman
https://developer.mozilla.org/en-US/docs/Web/HTTP/Access_control_CORS explains why it is that even though you can access the response with Postman, browsers won’t let you access the response cross-origin from frontend JavaScript code running in a web app unless the response includes an Access-Control-Allow-Origin response header.
http://catfacts-api.appspot.com/api/facts?number=99 has no Access-Control-Allow-Origin response header, so there’s no way your frontend code can access the response cross-origin.
Your browser can get the response fine and you can see it in Postman and even in browser devtools—but that doesn’t mean browsers expose it to your code. They won’t, because it has no Access-Control-Allow-Origin response header. So you must instead use a proxy to get it.
The proxy makes the request to that site, gets the response, adds the Access-Control-Allow-Origin response header and any other CORS headers needed, then passes that back to your requesting code. And that response with the Access-Control-Allow-Origin header added is what the browser sees, so the browser lets your frontend code actually access the response.
So I am trying to pass in an object, to my Fetch which will disable CORS
You don’t want to do that. To be clear, when you say you want to “disable CORS” it seems you actually mean you want to disable the same-origin policy. CORS itself is actually a way to do that — CORS is a way to loosen the same-origin policy, not a way to restrict it.
But anyway, it’s true you can—in your local environment—do suff like give a browser runtime flags to disable security and run insecurely, or you can install a browser extension locally to get around the same-origin policy, but all that does is change the situation just for you locally.
No matter what you change locally, anybody else trying to use your app is still going to run into the same-origin policy, and there’s no way you can disable that for other users of your app.
You most likely never want to use mode: 'no-cors' in practice except in a few limited cases, and even then only if you know exactly what you’re doing and what the effects are. That’s because what setting mode: 'no-cors' actually says to the browser is, “Block my frontend JavaScript code from looking into the contents of the response body and headers under all circumstances.” In most cases that’s obviously really not what you want.
As far as the cases when you would want to consider using mode: 'no-cors', see the answer at What limitations apply to opaque responses? for the details. The gist of it is:
In the limited case when you’re using JavaScript to put content from another origin into a <script>, <link rel=stylesheet>, <img>, <video>, <audio>, <object>, <embed>, or <iframe> element (which works because embedding of resources cross-origin is allowed for those)—but for some reason you don’t want to/can’t do that just by having the markup of the document use the resource URL as the href or src attribute for the element.
When the only thing you want to do with a resource is to cache it. As alluded to in What limitations apply to opaque responses?, in practice the scenario that’s for is when you’re using Service Workers, in which case the API that’s relevant is the Cache Storage API.
But even in those limited cases, there are some important gotchas to be aware of; see the answer at What limitations apply to opaque responses? for the details.
I have also tried to pass in the object { mode: 'opaque'}
There is no 'opaque' request mode — opaque is instead just a property of the response, and browsers set that opaque property on responses from requests sent with no-cors mode.
But incidentally the word opaque is a pretty explicit signal about the nature of the response you end up with: “opaque” means you can’t see into any of its details; it blocks you from seeing.
If you are trying to address this issue temporarily on your localhost, you can use this chrome extension : Allow CORS Access-Control-Allow-Origin
https://chrome.google.com/webstore/detail/allow-cors-access-control/lhobafahddgcelffkeicbaginigeejlf
If you are using Express as back-end you just have to install cors and import and use it in app.use(cors());.
If it is not resolved then try switching ports.
It will surely resolve after switching ports
So if you're like me and developing a website on localhost where you're trying to fetch data from Laravel API and use it in your Vue front-end, and you see this problem, here is how I solved it:
In your Laravel project, run command php artisan make:middleware Cors. This will create app/Http/Middleware/Cors.php for you.
Add the following code inside the handles function in Cors.php:
return $next($request)
->header('Access-Control-Allow-Origin', '*')
->header('Access-Control-Allow-Methods', 'GET, POST, PUT, DELETE, OPTIONS');
In app/Http/kernel.php, add the following entry in $routeMiddleware array:
‘cors’ => \App\Http\Middleware\Cors::class
(There would be other entries in the array like auth, guest etc. Also make sure you're doing this in app/Http/kernel.php because there is another kernel.php too in Laravel)
Add this middleware on route registration for all the routes where you want to allow access, like this:
Route::group(['middleware' => 'cors'], function () {
Route::get('getData', 'v1\MyController#getData');
Route::get('getData2', 'v1\MyController#getData2');
});
In Vue front-end, make sure you call this API in mounted() function and not in data(). Also make sure you use http:// or https:// with the URL in your fetch() call.
Full credits to Pete Houston's blog article.
You can also set up a reverse proxy which adds the CORS headers using a self-hosted CORS Anywhere or Just CORS if you want a managed solution.
https://justcors.com/<id>/<your-requested-resource>
http://cors-anywhere.com/<your-requested-resource>
Very easy solution (2 min to config) is to use local-ssl-proxy package from npm
The usage is straight pretty forward:
1. Install the package:
npm install -g local-ssl-proxy
2. While running your local-server mask it with the local-ssl-proxy --source 9001 --target 9000
P.S: Replace --target 9000 with the -- "number of your port" and --source 9001 with --source "number of your port +1"
Solution for me was to just do it server side
I used the C# WebClient library to get the data (in my case it was image data) and send it back to the client. There's probably something very similar in your chosen server-side language.
//Server side, api controller
[Route("api/ItemImage/GetItemImageFromURL")]
public IActionResult GetItemImageFromURL([FromQuery] string url)
{
ItemImage image = new ItemImage();
using(WebClient client = new WebClient()){
image.Bytes = client.DownloadData(url);
return Ok(image);
}
}
You can tweak it to whatever your own use case is. The main point is client.DownloadData() worked without any CORS errors. Typically CORS issues are only between websites, hence it being okay to make 'cross-site' requests from your server.
Then the React fetch call is as simple as:
//React component
fetch(`api/ItemImage/GetItemImageFromURL?url=${imageURL}`, {
method: 'GET',
})
.then(resp => resp.json() as Promise<ItemImage>)
.then(imgResponse => {
// Do more stuff....
)}
I had a similar problem with my browser debugger saying my response.body was null but fiddler and the developer tools show it as populated that turned out to be basically the same scenario as this. I was using a local Angular application hitting a Web Api service running on IISExpress. I fixed it by following the steps outlined here to find the correct applicationhost.config file to add a Access-Control-Allow-Origin header like so:
<customHeaders>
<clear />
<add name="X-Powered-By" value="ASP.NET" />
<add name="Access-Control-Allow-Origin" value="*" />
<add name="Access-Control-Allow-Headers" value="Content-Type" />
</customHeaders>
If all the above solutions don't work, probably it's because of the file permissions as sometimes even if you have fixed the non-cors problem using Heroku or another way, it throws 403 forbidden error. Set the directory/file permissions like this:
Permissions and ownership errors
A 403 Forbidden error can also be caused by incorrect ownership or permissions on your web content files and folders.
Permissions
Rule of thumb for correct permissions:
Folders: 755
Static Content: 644
Dynamic Content: 700

Request.Headers["Referer"] does not exist

I have a solution with two ASP.NET Core MVC projects. One project (Client) is making a request to the other (Server) using HttpClient. When the action in Server receives the request, I want to get the URL of the thing that sent it. Every article I have read purports Request.Headers["Referer"] as the solution, but in my case Headers does not contain a "referer" key (or "referrer").
When receiving the request in Server, how should I find the URL of the Client that sent it?
That is how you you get the referring url for a request. But the referer isn't the thing that sent the request. The referer gets set in the headers by the browser when a person clicks on a link from one website to go to another website. When that request is made by the browser to the new website the request will typically have the Referer header which will contain the url of the prior website.
The receiving server can't get the url of the "client" making the request, remember a typical web browser client isn't at any url. All the receiving server can get is the IP address of the client typically.
Since you have control of the client software, if you wanted you could have the client put whatever info you want in the header of the request before it's sent to the server and the server could then get that info out of the header.
If you're using HttpClient, then it is up to the site making the request to add that header. It isn't added automatically in this case. So: change the code - or request that the code is changed - so as to add the header and value that you expect. If you are proxying through a request, you might get the value from the current request's Referer header, and add that.
Even in the general case of a browser making the request as part of a normal page cycle, you can't rely on it: the Referer header is often deliberately not sent; depending on the browser version, configuration, whether you're going between different domains, whether it is HTTPS or not, and rel markers on a <a href=... such as "noreferrer".

Check query string (or maybe not a query string) to see if value equals [duplicate]

For example if I type in the URL:
http://www.foo.com/page.php?parameter=kickme#MOREURL
Then on the server there is no part: #MOREURL
Is possible to send or get these part to the server without jQuery AJAX?.
No, it is available to the browser only, so you have to deal it with Javascript. The server can not read it.
Explanation:
Basically the hash component of the page URL (the part following the # sign) is processed by the browser only - the browser never passes it to the server. This sadly is part of the HTML standard and is the same whether or not you are using IE or any other browser (and for that matter PHP or any other server side technology).
Here's what Wikipedia says about it:
The fragment identifier functions differently than the rest of the URI: namely, its processing is exclusively client-side with no participation from the server. When an agent (such as a Web browser) requests a resource from a Web server, the agent sends the URI to the server, but does not send the fragment. Instead, the agent waits for the server to send the resource, and then the agent processes the resource according to the fragment value. In the most common case, the agent scrolls a Web page down to the anchor element which has an attribute string equal to the fragment value. Other client behaviors are possible
https://www.rfc-editor.org/rfc/rfc2396#section-4
When a URI reference is used to
perform a retrieval action on the
identified resource, the optional
fragment identifier, separated from
the URI by a crosshatch ("#")
character, consists of additional
reference information to be
interpreted by the user agent after
the retrieval action has been
successfully completed. As such, it is
not part of a URI, but is often used
in conjunction with a URI.
I would like to extend the answer on the reason WHY the fragment is not sent to the server. Because it is intentional and desired behavior. Let's look at URL string in whole.
/path/to/element?query=string&for=server#?optional=fragment&for=browser
<----- URI ----> <---- QUERY STRING ---> <----- FRAGMENT STRING ------>
URI uniquely specifies resource fetched from a server
QUERY defines operations to be performed by the server on the resource
FRAGMENT controls browser (application) behavior. Fragment should be used to
store application state which should be visible to the user so the user can send link to another user to get the same application state.
Fragment is the only part of URL free for you to transparently implement single-page web applications (which can run offline on your mobile phone for example).
Therefore it must not be sent to the server.
The hash component is not passed on to the server but it is extensively used on the client side. Specifically, in single page applications, the text following a hash is used to represent state of the application as different routes. Thus, what happens is: following an initial request to the server which serves the 'home' page along with additional js files which include client-side routing logic such as the router, whenever the user navigates anywhere on the page by clicking an anchor tag, only the part of the URL following the hash component is changed. This prevents a GET request to the server and in response to this 'onhashchange' event, the content of the single page application can be updated depending on the exact route.

MVC HttpGet and HttpPost

Recently I have attended a training in mvc. The trainer said that - As per the security concerns we have to use HttpPost instead of HttpGet. Always use HttpPost.
Can anyone explain - what is the security issue when we use HttpGet?
When transmitting data over secure connection (https) body of the post request is encrypted and practically undreadable, you can only see address where data is going but not the data itself. Get on the other hand has no body and data has to be transmitted in either query string or as a path parameter. While it is true that query string does get encrypted as well, due to request logging on the server and browser it is possible to get hold of that data.
Anyone can insert image on public forum or stackoverflow with link to your web-site. Then happens next:
Browser looks at url in image tag
Browser find cookies corresponding to domain in url
Browser sends request to url with cookies of user
Your server performs action
Browser tries to parse response as image and fails
Browser renders error instead of image
But if you mark your action as Http Post only then this scenario isn't applicable for 90% of sites. But you should also consider that if hacker can create a form on other web-site then he still can make browser to perform request. So you need CSRF. Well, browsers made a lot to prevent cross-site requests, but it's still possible in some scenarios.

IFrame referer question - asp.net c#

One of our application will be run in an iframe, inside salesforce and I'm having troubles with accessing the referer. They'd like us to do some referer checks, to make sure the request is coming from salesforce and we've been given the IP addresses to check against.
My problem is that anytime I try to access the referer through either of the following two methods:
HttpContext.Current.Request.ServerVariables["HTTP_REFERER"]
HttpContext.Current.Request.UrlReferrer
it returns me null.
Any ideas how could I get hold of the referer?
PS: I'm aware that you can spoof the referer, but it's part of the requirement.
If I understand the question correctly you have client sites that refer to your site by embedding IFrames in their webpages the point to your site. You wish to "ensure" that the requests are coming from host page which itself is part of a designated set of sites. The set of designated sites is described by a set of IP addresses. Does that cover it?
Tricky. First off lets assume you've got a referer. You will need to aquire the host name from it (easy enough using the Uri type). Then you need to resolve the IP address for the host name using DNS (again not too difficult with .NET framework).
Of course you need to get a referer and that is the trickiest bit. Browsers do not always place a referer header in the request. This is especially true when the referee address is not in the same domain as the referer, which is the case here. IOW, this is a showstopper.
A better approach to solving this problem (and is not prone to spoofing) is to use some hash based authentication. Doesn't have to be too sophisticated (if the original requirements felt the referer testing was sufficient anyway).
A referrer is only there if the page was requested through a link. When a page is opened say from the address bar in a browser by typing in the address directly (or in your case y setting the src. of the IFRAME), the referrer will be empty.

Categories