Why does PHP store uploaded files in a temporary location and what is the benefit?

Question

Okay I\'m totally new in this field and going through some tutorials and I found that while uploading files in PHP it stores them in a temporary location.

$file         


        

Answer 1

Good question. The short answer is that PHP must process the entire HTTP request - filling out $_POST with data, and $_FILES as needed - before giving control to your script. Since your script doesn't gain control until after the processing, there's no way to tell PHP where to put that file data.

But why does PHP do it this way? Well, let's look at an HTTP POST with file data:

POST /upload?upload_progress_id=12344 HTTP/1.1
Host: localhost:3000
Content-Length: 1325
Origin: http://localhost:3000
... other headers ...
Content-Type: multipart/form-data; boundary=----WebKitFormBoundaryePkpFF7tjBAqx29L

------WebKitFormBoundaryePkpFF7tjBAqx29L
Content-Disposition: form-data; name="MAX_FILE_SIZE"

100000
------WebKitFormBoundaryePkpFF7tjBAqx29L
Content-Disposition: form-data; name="uploadedfile"; filename="hello.o"
Content-Type: application/x-object

... contents of file goes here ...
------WebKitFormBoundaryePkpFF7tjBAqx29L--

Notice that the contents of the request are a multi-part encoded document, with form fields interspersed among file data. In this particular example, the form field occurs before the file data. However, it's possible - indeed likely - that form data occurs after file data.

So, in order to guarantee that PHP can give you all the $_POST data, PHP must process the entire request. So it might as well complete the $_FILES super-global while it's there.

Now, PHP could keep this file data in memory, but this might really be a bad idea. Think about what would happen if PHP needed to store a 100 MiB file a user uploaded. Suddenly, you've got a 100 MiB increase in the RSS of your Apache process, which is really not too good - Apache might be ulimited to not have that much space, or Apache might get swapped: to your users anguish. So, PHP does the next best thing: put this received file in a temporary file.

You might ask why PHP can't be told what file to put the incoming file data first, so you didn't have to move it. Well, that's a bootstrapping problem: PHP hasn't handed control over to the script yet, so the script can't tell PHP where to put the file. Thus, PHP does the best it can: put the file data into a temporary file.

Now, you can keep this file data in a RAM disk, for speed if you want. This is a good approach if you don't mind the infrastructure cost (eg, maintaining the RAM disk setup). But note this isn't like PHP holding it in RAM itself: in that scenario, the PHP container process (usually Apache or some other web server) must have the heap to hold the file (which it might not). In this scenario, the RAM disk is managed by the kernel.