Matt Bonneau

@nivshah Do you have any blocking code (maybe a PDO database call or something?)

@nineff Is this with RFC6455 0.2.6? This is an interesting find. It appears that the so-called "sensible default" maximum size limit of 25% of memory for messages and frames is...

I agree with your statements. A possible solution would be to allow a `MessageBuffer` factory to be injected into the `WsServer` so that users would be able to specify more...

@cboden - Is this still worth implementing? Fast failing on UTF-8 would only be useful when there is a misbehaving client. We still check before passing this off to the...

If protocol error means that the connection is shut down, I would be in favor of invalid URI being an application error. If you expand the scope of the question...

:+1: I would like to see the `disclose_caller`/`disclose_publisher` option provide all this information if the router has it.

@gammazero Thruway is a little out of sync with the spec. @oberstet Would still like to see the `authroles` officially supported see #99

+1

> whereas x_ prefixed attributes must be forwarded (not dropped) by a router. > > eg CALL.options.x_foobar should be blindly forwarded by a router to INVOCATION.details.x_foobar > >does that make...

:+1: for standards. We will update Thruway to match the standard once there is a definitive one. Will the uri and match option be the only things that determine the...