Maximilian Hils

Yeah, supporting STARTTLS etc would be nice. That's probably more important once we support server greetings, and should be easier to implement for sans-io. :) refs #1775

Thanks again for the very nice report even though information is understandably limited at this time. :) Can you verify that this does not occur when disabling HTTP/2 keepalives? Assuming...

Yes, that sounds reasonable if it can be implemented without too much new complexity. 👍

Thanks! Is there a special reason/convention I'm unaware of for adding backspace? Feels like that's something that is easily fatfingered, maybe just adding the delete key is better?

To confirm, your point is that plain HTTP traffic to other hosts should not appear in the UI if `--allow-hosts` is passed? https://docs.mitmproxy.org/dev/howto-ignoredomains/#limitations and https://docs.mitmproxy.org/dev/howto-ignoredomains/#fn:1 in particular should explain why...

Yes, that is correct! Sorry for the slow reply. :)

> @mhils Just to clarify the summary by @Joshfindit, mitmproxy still acts on HTTPS requests other than _allow-hosts_ by proxying the connections as a normal proxy service. Yes, correct. More...

Yeah, I see where you're coming from. The waters get a bit murky because it also applies to plain TCP proxying, it's only plain HTTP without CONNECT that we cannot...

Yes, that sounds good to me! It makes sense to add an event hook for this. Would you be willing to give the implementation a stab? We should also ponder...

> I suspect this would require so much time wrt the actual value my contribution can bring, to understand the events logic and provide an acceptably clean implementation. I prefer...