Arnout Engelen

Build failures might be due to https://github.com/lampepfl/dotty/issues/11022 (though funny that we didn't see that problem on 3.1.2 in Akka)

I see the appeal, on the other hand the current behavior of 'all tweets since the latest toot' is pretty nice and predictable - I think I like it this...

We decided not to include the wildcard support in Akka HTTP. To simulate the Netty wildcard behavior we could pass in `upgrade.requestedProtocols.headOption` at https://github.com/playframework/playframework/blob/ca4b5fa132eee54ea3a8d7e313d198c0d713d8b8/framework/src/play-akka-http-server/src/main/scala/play/core/server/AkkaHttpServer.scala#L294 . Additionally it might make sense...

@HealthyPeace is it refused by Chrome itself or by a websocket library on top of chrome? Is it possible to share a minimal piece of code demonstrating the problem?

> Isn't the problem only triggered by aborting the connection? In that case, there's no good way to report an error back to the client. It's also triggered by closing...

> This should be spared for the next minor version. `master` is the upcoming 2.9, so that would just mean 'not backporting to 2.8.x', right?

Nonetheless there's something to be said for holding back bigger changes to master to reduce conflicts while backporting important smaller changes - WDYT?

> In light of https://doc.akka.io/docs/akka-http/current/security/2021-CVE-2021-42697-stack-overflow-parsing-user-agent.html, is it maybe worth going straight to 10.2.7? I'm not sure if that adds a lot of work/risk though. The vulnerability seems somewhat severe I...

@solarmosaic-kflorence that sounds reasonable to me, could you prepare a PR?

Again in https://travis-ci.com/github/lagom/lagom/jobs/339346526