Edouard Maleix

@trautonen Yes, there is. I am pretty interested in contributing to this one.

The `ValidationPipe` constructor can also receive an `exceptionFactory` to create your exceptions. Then, you could have the options for the `WebSocketGateway` validators as a shared constant. I have been doing...

To make @jonasof solution even more useful i suggest to also resolve the dynamic type so that @ValidateNested can also be used. To continue with his example: ```ts export abstract...

@ansd @kjnilsson I am very interested to know more about the solution based on the RabbitMQ streams that you presented [here](https://github.com/rabbitmq/rabbitmq-server/issues/8096#issuecomment-1535144257). Since I have no experience with streams yet, my...

This could be an opportunity to revise how modules are exported and solve #878 by the same occasion.

I will work on a PR soon. For reference i am copying the patch that i applied for [aedes-otel-instrumentation](https://github.com/moscajs/aedes-otel-instrumentation). ```diff diff --git a/node_modules/aedes/aedes.js b/node_modules/aedes/aedes.js index c02d289..668b162 100644 --- a/node_modules/aedes/aedes.js +++...

> This seems a massive change to do (not opposed). > > I think modifying aedes-packet is incorrect, as the transaction should be decouple in the tracing flow: publishing terminates...

> Changing mqtt-packet is a non starter unfortunately:(. Maybe i wasn’t clear, no change will be requested in the mqtt-packet source code.

> A _better_ approach would be to design a system that requires no monkeypatching at all. This would definitely be faster. How do you imagine this design ? Allowing the...

I agree that it would make less of a performance penalty for MQTT 5. For MQTT 3 if some users truly wish to propagate trace between their systems, i don't...