Andrew Godwin

Yes, and I appreciate you helping me work through this!

I don't think I'm _that_ concerned about backwards compat in this case - the behaviour here is already undefined depending on the server, and I think making that into an...

An application could, in theory, still try to `send()` after a client disconnection has happened, so I think it can happen with HTTP as well?

To be fair, I wrote that previous message five years ago; in the time since, I have now come to believe that explicit errors in this case _could be_ good,...

It would never get to the end user as, by its nature, they are disconnected at that point - it would just clutter up server error tracking (which it would...

Well I am proposing that we instead say you can return a future that says if it's sent or not - servers can immediately set_result with true/false if they want...

Ah yes, that is an interesting take on it - of course, it does bring us back to the problem of "which exception", since we can't have servers catch and...

Yes - I think this can be considered closed out now.

Agreed - this has sat around for long enough because I simply don't think we can do a good enough job of it that will be less bad for users...

Ah, I did wonder, given our previous discussions on the subject. You did still suggest mapping the other case over, just with a lot of warnings, which is fine in...