Generally, they don't. Most of the time, when you download a file, it's saved with new metadata, onto your system at time you downloaded it, with the permissions your user possesses. (Documents, image files, pdfs, anything that you download manually.)
There are two approaches. One, as you suggest, is to request that timestamps and other information are included in the headers. This is unreliable, because the server may not recognize or allow such information to be passed during a download, for various reasons.
The second (more common) approach is to have the requesting program note timestamps and other details itself while locating a desired file, and apply those details to the file after the download completes. This tends to be a preferred method, because it avoids situations where a server may not be setup correctly (applying an incorrect date/permissions to everything it sends), or a server that's purposely setup to give downloaders headaches (generally video sites. This is a problem not just for people "borrowing" videos, but archivers and historians).
As for inodes, they're specific to the system you're storing the program on, so inode data is never sent, unless you're transferring a complete disk backup over the internet.
Remember that sending extra information and recombining it at the endpoint is trivial for computers, and is pretty much how all data arrives at your system. The only major issue is whether the server and client care to acknowledge such data.
Last edited by scott_R; 12-04-2017 at 02:45 AM.
Reason: Inodes addition
|