Move setting registrations into QPACK #1358
Conversation
MikeBishop
added
editorial
| SETTINGS_HEADER_TABLE_SIZE (0x1): | ||
| : An integer with a maximum value of 2^30 - 1. The default value is 4,096 | ||
| bytes. | ||
| The following setting is defined in HTTP/QUIC: | ||
|
|
||
| SETTINGS_MAX_HEADER_LIST_SIZE (0x6): |
There was a problem hiding this comment.
Does anything in this document explain what SETTINGS_MAX_HEADER_LIST_SIZE does?
There was a problem hiding this comment.
Good point. RFC7540 doesn't have much, but it has more than HQ does currently.
|
|
||
| SETTINGS_MAX_HEADER_LIST_SIZE (0x6): | ||
| : An integer with a maximum value of 2^30 - 1. The default value is | ||
| unlimited. | ||
|
|
||
| SETTINGS_QPACK_BLOCKED_STREAMS (0x7): | ||
| : An integer with a maximum value of 2^16 - 1. The default value is 100. | ||
| Additional settings MAY be defined by extensions to HTTP/QUIC. |
There was a problem hiding this comment.
Are we calling QPACK an extension to HTTP/QUIC? Or it is for purposes of SETTINGS?
There was a problem hiding this comment.
Kind of. It's required to parse HEADERS and PUSH_PROMISE, but it is its own thing. Kind of a mandatory extension, I suppose?
| @@ -1418,13 +1413,11 @@ The entries in the following table are registered by this document. | |||
| |----------------------------|------|-------------------------| | |||
| | Setting Name | Code | Specification | | |||
| |----------------------------|:----:|-------------------------| | |||
| | HEADER_TABLE_SIZE | 0x1 | {{settings-parameters}} | | |||
There was a problem hiding this comment.
It strikes me as bizarre that 0x1 is omitted from this table / or that QPACK - as an extension or subordinate document - gets to define the lowest numbered setting value.
There was a problem hiding this comment.
I agree. That's because we're trying to keep consistency with the setting codes in HTTP/2, where this is defined in the HTTP draft and the value passed to HPACK as a parameter.
Fixes #1323. Trying to describe how the max table size was used made me realize we never actually say, which is why I also wrote #1357; there's a reference that needs to be updated in whichever PR lands second.