RE: HeaderDiff specification

[RUELLAN Herve <[e-mail]>]

Hi Wang,

Good catches! Thanks for them.

Herv¨¦.

> -----Original Message-----
> From: Íõ°| [mailto:[e-mail]]
> Sent: mardi 26 mars 2013 11:02
> To: RUELLAN Herve
> Cc: [e-mail]
> Subject: ´ð¸´: HeaderDiff specification
> 
> Hi, Herv¨¦.
> 
>    I'm reading your draft "draft-ruellan-headerdiff-00". According to my
> understanding and your code at https://github.com/hruellan/compression-
> test/tree/HeaderDiff, there seems to be some minor mistakes when you
> explain the Integer representation in Section 4.1.1 of the specification.
> 
>    1. On [Page 8], in the section "The algorithm to represent an integer I", step
> 1 of step 3, "Set I to (I - 2^N - 1) and Q to 1" should be "Set I to (I
> - (2^N - 1)) and Q to 1" or "Set I to (I - 2^N + 1) and Q to 1".
> 
>    2. On [Page 9], in the section "4.1.1.2.  Example 2: Encoding 1337 using a 5-
> bit prefix", "The value to represent on next bytes is I = 1337 - 2^5 = 1305"
> should be "The value to represent on next bytes is I = 1337 - (2^5 -
> 1) = 1306". And the following calculating should use 1306 instead of 1305.
> 
> ---------------------------------------------------------------------
> Wang Qian
> 
> >
> > All,
> >
> > Here is the detailed specification for our header representation
> > format HeaderDiff.
> >
> > We have also published an implementation, which can currently be found
> at:
> > https://github.com/hruellan/compression-test/tree/HeaderDiff
> >
> > (I've sent a pull request to have it included in
> > https://github.com/http2/compression-test).
> >
> > Comments are welcome,
> >
> > Herv¨¦.
>