Christian Palmstierna

@danielwertheim Sorry, was a bit swamped lately.. I like the idea of making it less intrusive on the implementors. Would it be possible to take it one step further and...

Ah my bad, I mixed something up in my head :) Letting it be up to the user makes sense, I agree. Although the question then is, does it even...

I was thinking to throw `InvalidOperationException` if the handler is null. Sound good? One question about the publish methods - should it resolve all handlers first and validate them before...

No, it's probably better not to re-use them. Was just thinking that it could make sense to first ensure all handlers can actually be resolved before executing any of them....

I think that makes sense as well. Another question - should the dispatcher receive an pre-baked `Action` or would it make sense to just send an `IEnumerable` instead? I'm leaning...

Good point. Using Confluent.Kafka you can set a queue buffering delay on the producer, so it waits for messages to buffer up before flushing. Having batched spans then could actually...

@adriancole That's good input. In that case my initial idea might work, which was to implement buffering in `ZipkinTracerReporter` instead. Thoughts?

That's an interesting idea. This would make the serializers/senders quite a bit more complex though - since the size isn't known until after serialization, the sender would need to serialize...

@roend83 I hit the same issue as described in this PR discussion. @mconnew explained the underlying problem further very well there - it's apparently not trivial to solve. https://github.com/CoreWCF/CoreWCF/pull/550#issuecomment-1083106378 If...

I have been toying with this idea before; it could definitely be very useful for non-actor-based clients. Perhaps even browser-based clients once grpc-web matures. Unfortunately, this project has been on...