Mike Purvis

I just hit this as well, I agree with colcon quietly smoothing over `unittest`'s inconsistency here.

Some previous discussion: https://github.com/colcon/colcon-core/issues/208 https://discourse.openrobotics.org/t/call-for-testing-standards-based-python-packaging-with-colcon/32008 https://discourse.openrobotics.org/t/status-of-colcon-building-standards-based-python-packages/40578

> @mikepurvis since you contributed this pull method, could you try this branch in your environment to ensure there is no regression? Thanks for the ping! I'm actually not at...

Hey, like most ROS 1 packages, this is basically unmaintained. If you’d like to take over maintainership for the purposes of preparing a new release, I’m good with that. At...

Do we know if there's an active request, or roadmap to PyPI exposing this information? Is there PEP to add build system info to the standard metadata? If not, would...

The design intent is that colcon can discover and build packages with non-package.xml metadata. Indeed, this capability is one of colcon's tentpole features over `catkin_tools`, which is completely dependent on...

I'm interested in this as well; it's a nuisance to implement it manually, and it would be particularly nice for perf if the filtering could happen on the rust side.

Yes, the rosbag Python API generates message definitions on the fly based on what's stored in the bag file itself. This can be a source of headaches as the `type`...

FWIW, in the Nix builds at Clearpath we correctly propagated/handled build_export_depend and I don't recall having to patch much stuff, at least up to ros-desktop-full on the ROS 1 side.

Can rosindex or some other tool give some visibility into how much build_export_depend is even being used? I would have guessed that in the overwhelming majority of cases, developers haven't...