What cookies are sent in an HTTP request?
Your browser stores a bunch of “cookies”.
Compared to other client-side web storage
the unique property of cookies is that,
when the browser makes an HTTP request,
cookies are embedded in the request sent to the server.
The cookies are included in an HTTP header called
in the developer tools for this page,
you’ll see an HTTP request for
a picture of me.
In the HTTP headers for that request, you can see:
Cookie: _ga=GA1.2.12345.67890; _gid=GA1.2.111111.222222
But why were these cookies sent, and not the many others stored by my browser? The answer is not simple!
Cookies have many attributes;
among them are the domain and path attributes,
which together are called the cookie’s “scope”.
Both Google Analytics cookies above have the domain
.jameshfisher.com and the path
You can see these attributes in the Application tab of Google Chrome developer tools.
These attributes are set when the cookie is set,
document.cookie = 'test_cookie=foo'.
When setting a cookie,
the default domain is the domain of the current page,
and the default path is the path of the current page,
Note the domain attributes of these two cookies are different:
The leading dot means “this is a domain suffix”;
a cookie with domain
.jameshfisher.com will match
A cookie with domain
jameshfisher.com will only match the domain
and will not match subdomains.
Confusingly, whether the domain attribute is a suffix is determined by
whether you explicitly set the domain attribute when setting the cookie.
If you explicitly set a domain attribute when setting the cookie,
it will be a domain suffix (shown with leading dot).
If you don’t explicitly set a domain attribute,
it will be a fully-qualified domain,
jameshfisher.com in this example.
A cookie is only sent in an HTTP request if the hostname that the request is being sent to matches the domain, or domain suffix, in the cookie. The browser will never send a cookie to a hostname that doesn’t match the cookie’s domain attribute.
An HTTP origin
An HTTP request made by a browser has two relevant origins: the origin that makes the request, and the origin that the request is sent to.
Whether cookies are sent depends on how the request was made. There are many APIs to invoke an HTTP request from a browser:
- The user follows a link, i.e. clicks on an
<a>element. Cookies are sent.
- The user submits a form, e.g. clicks on a
<button type="submit">element. Cookies are sent.
- The page loads some media, e.g. an image via an
<img>element. Cookies are sent.
- The page loads a script from a
<script>element. Cookies are sent, unless the element has
- The attributes of the cookie (domain and path)
- The domain and path of the URL the request is being made to
- The domain and path of the page making the request (???)
To test this behavior,
I’ve set up some domains in
$ cat /etc/hosts ... 127.0.0.1 s1.com 127.0.0.1 s2.com 127.0.0.1 sub.s2.com
And I’m running the “web servers” for
nc, like so:
$ while true; do cat response.http | nc -l 8000; done
This lets me control the full HTTP response by typing it out. Here’s a starter page:
$ cat index.http HTTP/1.1 200 OK Content-Type: text/html; charset=UTF-8 <!doctype html> <html><body><h1>A webpage served by netcat</h1></body></html>
nc in an infinite loop to serve this file for every response:
When the browser makes an HTTP request, it will include each cookie whose scope matches the target URL.
More by Jim
- Smear phishing: a new Android vulnerability
- A probabilistic pub quiz for nerds
- Time is running out to catch COVID-19
- The inception bar: a new phishing method
- The hacker hype cycle
- Project C-43: the lost origins of asymmetric crypto
- How Hacker News stays interesting
- My parents are Flat-Earthers
- The dots do matter: how to scam a Gmail user
- The sorry state of OpenSSL usability
- I hate telephones
- The Three Ts of Time, Thought and Typing: measuring cost on the web
- Granddad died today
- Your syntax highlighter is wrong