Adam Ross

Yes, the unit test coverage I wrote for the helper function consistently reproduces. I've pushed what I have to a draft PR to illustrate: https://github.com/cloudevents/sdk-go/pull/799

Making this more specific, suppose I want to unmarshal the data payload of a CloudEvent delivered as application/json to a protobuf. ### Type Definition ```golang type ProtobufMessageData struct { //...

I'm not sure I follow this idea, I believe you are suggesting we don't want supplied types to have an awareness of the CloudEvent SDK, and so instead we modify...

I've thought this over, and I think your solution has cleaner separation of concerns but a lot more work to implement and optimize. I've also seen priority registry patterns like...

@n3wscott do you have further thoughts on the direction? (If it helps I hope to contribute implementation.)

Hi @n3wscott , this ended up at larger complexity than I initially thought and I am considering the timing of when pursuing this makes sense. Thanks for helping identify a...

This issue is about addressing technical debt and previously decided improvements that already had their primary development completed. I am against closing.

I suggest following the existing pattern and putting this inside the CloudFunctionsDeployTrait or creating a new trait to encapsulate utilities for async testing. When the first example was merged, we...

Do we have any data on the benefits? E.g., A/B performance?

I've reproduced the problem with the sample, it does not work because some change in the HTTP proxy server in the runtime was changed, and is now prefixing stderr messages...