Karsten Knese

You might want to rebase once again as it looks like you've got some duplicated commits on your branch. I've just checked and your PR compiles locally on my OSX...

> Isn't this PR required #1081 for this change to make sense? I would try holding off from #1081 until these changes are settled. The changes introduced in the other...

The typetrait will return correctly as well with `rcl_serialized_message_t`. But I agree, before calling the callback, I believe that `get_rcl_serialized_message()` has to be called to get the rcl type.

> That's true, but I think it is worth it to do that change now. Sounds good. I am on it ;-)

@DensoADAS with Foxy being out, I am happy to resume our work on the two outstanding PRs. Are you still planning on working on them?

@DensoADAS friendly ping

@flynneva thanks for picking this up. The basic idea here is as follows: Right now, the package is receiving an externally initiated `rclcpp::Node` in its constructor rather then itself being...

@Rayman I won't really work on this in the near future. It might be easier for you to open a new PR - incorporate my existing commits from this PR...