I'd swap the order of those 2 checks as if there's no joint trajectory passed there is no way to get the warning about multidof trajectory.

It may be of personal taste but I'd also revamp this file a bit in order to remove all those continue-s. A goto is a goto even if they give it a new name :D

I forgot about this request...

You are right about the order of the checks. I swapped them.

It may be of personal taste but I'd also revamp this file a bit in order to remove all those continue-s.

There is only one continue in this function and I very much prefer it over a big if { ...} block around the whole loop content. I would like to keep it as is.

Please have another look.

Another question: shouldn't we just error out when a multidof trajectory arrives? This allows for passing in a bunch of stuff that then gets ignored.

In theory we could error out and tell the user to either submit a patch for working validation or disable validation explicitly. But that would entail that everyone planning with a multi-dof joint has to disable start state validation by hand. This use-case "just worked" before we added validation, so it would break peoples workflow.

I believe the warning is enough in this case to tell the user "hey submit a patch or don't complain if something breaks".

Also, this topic is probably beyond the scope of this PR.

👍