Re: [quicwg/base-drafts] Forwarding upstream errors, and the implications (#3300)

[Mike Bishop <[e-mail]>]

Taking a look at the text, I still think the right principle here is that errors which might have been tunneled from an origin shouldn't be required to be stream errors.  Looking at the text, here are the things that might fall into this category:

- "Malformed" requests or responses are already stream errors
- Other frame types on CONNECT streams are connection errors, but with a CONNECT stream, the tunneled content is the DATA payload.  This could go either way.
- Frames that don't have the right size/content are connection errors, including final frames on streams that close cleanly.
- Control frames on non-control streams are connection errors.

Does this principle hold for the opposite direction?  Should servers consider garbage on streams as stream errors, because the "client" might actually be a proxy for several clients?  It's probably the simplest to keep things symmetrical.

I've put #3336 as an alternative solution to this.

-- 
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/3300#issuecomment-573871628