Re: [quicwg/base-drafts] Reorder HTTP frame fields to put Type first (#1351)

[Mike Bishop <[e-mail]>]

The existing order is because it started with what was in HTTP/2, but we've already established consensus that similarity to HTTP/2 isn't a sufficient reason to keep any non-functional aspect.

I'm inclined to think that reading the length first, then knowing how much data to pass to your type-specific parser would be the most efficient.  The varint Length means that frames can be very large (DATA frames, especially), so you don't want to wait until you have the entire frame to start processing it, but knowing when to stop passing data to the parser and dispatch a new frame seems worthwhile nonetheless.

I don't have hands on an implementation of this, though, so if there's consensus that one is easier to parse than the other, I'm fine either way.

-- 
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/1351#issuecomment-388444699