强制浏览器下载大文件

Question

I'm using a POST request with body to dynamically create a specified file on the server side and streaming it to the client as a response. The problem on the client side is that the file is being downloaded into the response object (browser's cache memory I suppose, as a temporary file) instead of the PC's hard drive.

How can I pass the byte stream directly for the browser to handle and direct into the /Downloads folder or wherever the user has set the downloads go?

The file size is really large and there's not enough space on the gadget's hard drive to create it there and give a link for GET request. So I need to create it on the fly and stream to the client as a POST response.

Right now I've written my code in Angular, but the language doesn't really matter. I'd like to understand how the browser handles the situation.
When I execute the HTTP request, I bet the browser does something with it before it's passed to the socket. And I'm sure the browser does something with the received request too before passing it to the piece of code which performed the request.

I mean can you only by setting the response headers make the browser download the file into /Downloads? What happens then to the waiting HTTP request if the packets are "stolen" by the browser? Or when the response headers arrive to the response object, can you somehow notify the browser that the body of the response is a file byte stream we wish to download into /Downloads?

And I have the possibility to modify my server code too.

EDIT More detailed explanation:
My server is sending the 5GB file to the client. Instead of just receiving that file in the temporary memory of the browser I would like to save the file directly on hard drive.
But naturally the browser prevents the direct access to the users computer so I can't write it there. So how do I pass the byte stream for the browser to handle and save into a file?

Answer 1

Ok, I have solved the problem. So if you send a file request yourself to the server from your javascript code the browser will just pass the received response directly to your response object, but you can't store it on the hard drive without some external (browser supported) tools.

What I mean is that if you perform any kind of $http.post() or .get() from your scripts the response will be passed to the $http.post().then(function(response){...}) directly and it's passed only after the whole response is received. Meaning that if you're expecting a file with the size of 5GB it will fail for being too large to be received inside the response object.

Instead what you have to do is to trigger the download in another way. What I did in my code is that I built a hidden

<form method="POST" action="/set.uri.here/" id="download">...</form>

and when the user clicks the button to download the file, I run a function that builds the body of the form and then submits it. Which looks like

var form = document.getElementById("download"); // magic here... create the body for the form and so on form.submit();

This way you can pass the request to be done by your browser so that then the response will be also handled by the browser. This is the crucial difference between doing the request yourself or by the browser. If you do the request yourself then the browser won't download the object but you will receive it yourself to do something with it.

If you wish the browser to download the object, then, make the browser do the request for the object as well!