[quicwg/base-drafts] Consider making SETTINGS part of the control stream header (#2783)

[Lucas Pardue <[e-mail]>]

Right now we have a SETTINGS frame that can only be sent once (in each direction), and if it is sent it has to be sent at a very precise point (the start of the stream) or the connection is bust. 

For these reasons, it strikes me that we could simplify things, in line with **today's** HTTP/3 design, by unframing settings and placing them as the control stream header. For illustration this would look something like below, where Identifier and Value can be sent multiple times.

```
0                   1                   2                   3
 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                           0x00 (i)                          ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                          Setting Length (i)             ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                        Identifier (i)                       ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                           Value (i)                         ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
```

**For**
* Simplifies some parts of error handling. 
  - The HTTP_MISSING_SETTINGS snowflake becomes redundant
  - Not possible to send multiple SETTINGS on the control stream to trigger HTTP_UNEXPECTED_FRAME
  - Not possible to send a SETTINGS frame on the wrong stream to trigger HTTP_WRONG_STREAM 
* A peer that is happy with defaults can send 00 (2 bytes) rather than 040 (3 bytes). This is a tinsy tiny benefit
* Forces endpoints to commit to SETTINGS very early (this is of dubious benefit if SETTINGS only loosen values)

**Against**
* We already have implementations that handle SETTINGS just fine
* Keeping the frame allows us change today's design to allow multiple SETTINGS in the future.

The arguments for and against seem to net out evenly. Changing control stream seems like a cleaner design to me but shipping is a feature :tm: .


-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/2783