Todd L. Montgomery
Todd L. Montgomery
@benjchristensen I agree. HTTP/2 takes control of the connection flow control so that implementations can schedule frames effectively. It's a frame/fragment scheduling issue. Or at least, this is how I...
Let us revive this. Especially in relation to resumability (#102).
I think the slides are accurate. And with HTTP/2, we don't need anything else. However, with TCP alone, there could be need for more. But I don't think we should...
Personal preference. For me, I like to see all the packet formats in a single place with minimal field explanations and leave processing and operation to their own sections. This...
I am doing my review in stages. Did one last week. Let it sit over the weekend. Doing another tomorrow. Then a final one later this week. Multiple reviews are...
I'd support the idea of rethinking resumption over H3.
There is still a warning here. https://github.com/real-logic/simple-binary-encoding/blob/727281a26157a244947ec8c8afa92176c3f98a9b/sbe-tool/src/main/java/uk/co/real_logic/sbe/xml/CompositeType.java#L207
Sure. Feel free to send us a PR that makes the most sense.
We could write a message header to the start.... but then that is the job of the FrameCodec....
One option might be to have the message header overlay the FrameCodec so it can do double duty. i.e. start of the FrameCodec is the MessageHeader.