new package and interfaces for describing other types

[wjwwood]

Still in development, but related to "evolving message types" REP-2011: ros-infrastructure/rep#358

There are still TODOs and known issues, but I think can iterate on those after merging this initial MVP. Merging will also unblock others working on related features.

[jhurliman]

Is this ready for review, or what components are missing / still in flux?

[wjwwood]

@jhurliman There are some TODO's left for me to update. We've been working on other technical parts of the REP lately so this hasn't made any progress lately. However, we will be returning to it before long.

[wjwwood]

Also, we took some time to reconsider having our own definition versus re-using something existing like the dds-xtypes_typeobject.idl from DDS-XTypes, but we decided to recommend against that and keep our custom format. I have to write up those thoughts into the REP as well.

[emersonknapp]

@clalancette @wjwwood what next steps do you think we need to take to get this landed? perhaps a set of consolidated TODO items will help us get it going. are any further updates needed to REP-2011 before we can move forward with these interfaces, or have we solidified that documented design well enough that it's only implementation details now?

[emersonknapp]

@wjwwood would you mind also rebasing this PR on rolling, when checked out it's an issue because of new service_msgs package being required by dependents

[wjwwood]

Rebased it on rolling.

[wjwwood]

I've moved this out of draft pr mode, and I think it's ready for review.

[methylDragon]

Small nits, but other than that I can't think of any fundamental issues with the message definitions

[allenh1]

Looks good to me as is.

[emersonknapp]

LGTM (after fixing typos)

[allenh1]

@wjwwood on second pass, I have a couple of suggestions.

[clalancette]

@allenh1 @wjwwood @methylDragon @emersonknapp @james-rms I've now done the updates to this PR that we discussed last night. I think this should be ready for final review. Here is CI:

• Linux
• Linux-aarch64
• Windows

[wjwwood]

lgtm

[allenh1]

Thanks for the update. Looks good to me.

[clalancette]

CI was...very yellow. Even though I don't think it is the fault of this PR, here is another try:

• Linux
• Linux-aarch64
• Windows

[clalancette]

The flake8 failures are known (and fixed elsewhere). The one additional failure is on Windows, but that is a known flake for right now. So this all looks good, I'm going to go ahead and merge this in.