Stefan Post

@pjfanning I see it failed on binary compatibility checks this time... Will fix this as well...

@pjfanning You were right about the tests. They are failing. I know how to fix them, but it is not going to be as clean as I would like it...

@pjfanning PR should be good now. Instead of introducing new QoS flags for the Subscribe message, I thought it was best to make it work with the existing QoS flags...

@pjfanning Test coverage could be nice indeed. For this to make sense we would need to test against a mqtt broker that is known to be strict on following the...

@pjfanning As for mqtt 5: For my work we rely on mqtt 3. Part of my job is to make sure everything runs stable and if there is a problem,...

@pjfanning I think I would like to test with AWS IoT just once more, to be absolutely sure. I will post the results when I am done.

@pjfanning I have tested with AWS IoT. The results are below. Subscribing with QoSAtLeastOnceDelivery *without* the encoding fix gives "Upstream finished": ``` [info] {"level":"DEBUG","origin":"org.apache.pekko.stream.connectors.mqtt.streaming.scaladsl.ActorMqttClientSession","sc":{},"title":"[client-commandFlow] Element: Right(Subscribe(PacketId(1),Vector((22574/lwt,ControlPacketFlags(2)), (13182/lwt,ControlPacketFlags(2)), (67782/lwt,ControlPacketFlags(2)), (13185/lwt,ControlPacketFlags(2)), (114855/lwt,ControlPacketFlags(2)),...

@Hazza88 Ah that would explain why I can't reproduce. To implement warp, AutoRaise specifically intercepts the cmd+backtick combination. Any other means of switching between windows of the same app isn't...

@Hazza88 Thanks for sharing :-). I think the used key combination in theory can be made configurable but this would add a considerable amount of code to AutoRaise to make...

I think this happens because the image element might not support the Accessibility API. As such, AutoRaise uses a fallback method to determine the element underneath the mouse and activates...