geometry2
geometry2 copied to clipboard
ros2
New approach for toMsg()/fromMsg()
This is a rewrite of #368. Depends (more or less) on #422, #423, #424, #425 and #425, so it should be rebased and merged once these PRs are merged. First actual commit of this PR is 68a1193.
@gleichdick OK, I think all of the other ones have been merged now. When you get a chance, a rebase of this one would be appreciated (or I can do it if you don't have time). Once we have that done, we can take a look at the current state of this.
So here we go... One thing came to my mind: As the messages can be used with a custom allocator, does it make sense to add support for it in the conversion methods?
One thing came to my mind: As the messages can be used with a custom allocator, does it make sense to add support for it in the conversion methods?
Not at the moment, no. It turns out that we need to revamp how the custom allocators are done since they don't quite work right at present. Until we know what that solution looks like, let's just stay away from the custom allocators.
The fixups for the header includes and the doxygen comments are now squashed.
Well, I started working on this PR almost one year ago. Indeed, it is pretty big and hard to review, but I think this redesign of geometry2 is necessary. @clalancette @ahcorde I'm intending to close this PR after the Humble Hawksbill API freeze (April 4, 2022).
@gleichdick I'm assigning myself to this PR since it's gotten so stale and I'd like to get this merged... I'm sorry this took so long to pick up again
Do you think this is this still necessary, and if so, would you be able to rebase it? I'm iffy on downstream impacts
I'm sorry, I stopped using ROS more than a year ago. I don't have any clue on how big the impact is on downstream projects, but things will break for sure. And the ros2 branch has moved quite a bit. I don't see myself spending more time on this.
But in my mind the reasons why I wrote this PR are still valid
I'm sorry, I stopped using ROS more than a year ago. I don't have any clue on how big the impact is on downstream projects, but things will break for sure. And the ros2 branch has moved quite a bit. I don't see myself spending more time on this.
But in my mind the reasons why I wrote this PR are still valid
Alright, I understand; thanks so much for your contributions so far!
@methylDragon I'm going to reopen this, as I do think the idea here is valid. It's just that this is a large change, so needs a lot of thought.
I agree that this is something that would be great to bring in and modernize this part of the API. But it's also something that we'll need to think closely about and make sure that there's a good design document or possibly REP with a migration path clearly defined as well as appropriate deprecation process with so many people potentially using this.
