I could not find anything in the RFC docs about whether the User-Agent field needed to be formatted the way it commonly is. For example, could I just do:
User-Agent: Mozilla windows NT
and still pass a valid request?
Add a comment
|
2 Answers
The formal definition of the User-Agent header is defined in RFC 2616 Section 14.43 as follows:
User-Agent = "User-Agent" ":" 1*( product | comment )
product is defined in Section 3.8 as follows:
product = token ["/" product-version]
product-version = token
token and comment are defined in Section 2.2 as follows:
token = 1*<any CHAR except CTLs or separators>
comment = "(" *( ctext | quoted-pair | comment ) ")"
Whitespace is a separator, so to answer your question, Mozilla windows NT is NOT a valid product by the above token definition. If you want to follow the spec, you could instead format it more like this: Mozilla (Windows NT).
2 Comments
jeffrey
What if I did Mozilla ((windows NT)), or something derivative. I'm guessing this would be invalid, as parenthesis are key characters here?
Remy Lebeau
Look at the definitions. A
comment is delimited by parenthesis, and a comment can contain another comment, thus parenthesis are nestable. So what do you think ((windows NT)) means? A nested comment.Sure, you can 'fake' a user agent. Check for yourself with this tool for instance: https://chrome.google.com/webstore/detail/user-agent-switcher-for-c/djflhoibgkdhkhhcedjiklpkjnoahfmg
1 Comment
jeffrey
I understand that you can fake a user agent, but this is still keeping the format of the user-agent to the RFC standard. I'm asking whether you can ignore the standard (standard is stated as name of application (comments) )