[rcl_interfaces] Homogenize iface definition

[mikaelarguedas]

This is just cleanup

• first commit make msg/srv definition consistent 27920fa
• second commit moves the repo's README.md to the right folder 89da210
• thiird commit creates a generic README for the entire repo 5cd70c2

[dirk-thomas]

I wouldn't consider those "internal". Most of them define public interfaces, no?

[mikaelarguedas]

"used for the implementation and testing" then ?

[dirk-thomas]

"defining the interface across client libraries"?

[mikaelarguedas]

Hmm not sure because as you said it is user facing and currently most of them are not used to communicate across client libraries given that they're used mostly/(only?) in rclcpp.

@ros2/team maybe some native speakers have some insights ?

[wjwwood]

They're definitely public and I would say their primary use is implementation of rcl concepts rather than testing, though they could be used for that too I guess.

"This repository contains a set of packages which themselves contain interface files (i.e. .msg and .srv files) used to implement client library concepts like parameters."

[mikaelarguedas]

my goal here is to give a description of the content of the repository and not only describe the rcl_interfaces package, that's why I talked about testing (given that this repo also contains test_msgs).

I'm fine with "This repository contains a set of packages which themselves contain interface files (i.e. .msg and .srv files) used to implement client library concepts" I just wanted to clarify the scope

[wjwwood]

Oh, I was confused by the diff and I didn't know there were test messages here. In that case I agree, something like:

"This repository contains a set of packages which themselves contain interface files (i.e. .msg and .srv files) which are used both to implement client library concepts and in testing."

Or something like that.

[tfoote]

I've updated it with approximately the verbiage of above.

[mikaelarguedas]

Thanks @tfoote
@ros2/team reay for review