Re: [quicwg/base-drafts] Why are error codes not a varint? (#1940)

[Kazuho Oku <[e-mail]>]

I think there are pros and cons in using a varint.

The pro is that we'd have a chance to save an octet as have been pointed out.

The cons that I see are:
* It would not give us any benefit when the application protocol defines error codes in groups (e.g., upper 8 bits indicate the category of the error and lower 8 bits indicate the details). To generalize, I am not sure if we can anticipate that the number space will be used from 0 and onwards unlike other variables that we send.
* It is a bit strange to state that an error code is a *62*-bit variable rather than a type that uses a multiple of 8-bits.

I do not have a strong opinion but considering the cons sticking with `uint16_t` is not too strange to me.

-- 
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/1940#issuecomment-434592735