Adobe release the AMF3 spec. Grab the documentation here.
I previously reverse engineered the AMF3 integer. And named it in the blog post "29-bit Integer". Someone commented on that* and I was never sure which name fit better: 29-bit or 32-bit Integer**. Until today! In the official documentation it is also called 29-bit Integer.
What I did not get back then, was that it can also be used unsigned with a uint. I also wondered about what happens when valid integer is encoded that is out of the reduced range. The documentation says that it is encoded as a double. ... I can"t remember that it did that when I reverse engineered it. But maybe I got it wrong or it was updated later...
Now after 3 semesters of university the task seems much easier due to some knowledge about the basic two's complement idea. However, I learned this concept before and I think I learned it "better" than in university this way.
(* sorry, comment was lost due to the migration to the new site.)
(** It can take up 8,16,24,32 bit to save but its range is 229 or -228..228.)