Re: HPACK, Draft 09, Integer Representation

On Fri, Aug 29, 2014 at 04:51:16PM +0000, RUELLAN Herve wrote:
> I made some more modifications to 6.1, to describe how variable
> length integer works, but I think this could be further improved.
> I'll try to add some drawings to this.

The prose seems to be missing the substraction of (2^N-1) if value
is at least that large.

That is, the second byte would contain the low 7 bits of the raw
value ("the value"), not low 7 bits of value - (2^N-1).

This would give xF 81 02 for 257@4b, whereas pseudocode gives
xF F2 01.

I presume the pseudocode is correct and 2^N-1 should be substracted.


Received on Friday, 29 August 2014 19:00:06 UTC