Make some changes for newer versions of uncrustify

[ahcorde]

Follow up of this PR #1578

I can see some failures here https://build.ros2.org/job/Rpr__rosbag2__ubuntu_noble_amd64/14/

[MichaelOrlov]

@ahcorde A general comment instead of guarding a lot of code segments with the

// *INDENT-OFF*
...
// *INDENT-ON*

statements it will be preferable to reformat everything with the new version of the Uncrustify.

cc: @clalancette It seems a nightmare with the new version of Uncrustify escalades more severely and more quickly than we assumed. The code formatting is going to diverge more significantly from previous ROS 2 distros and it will be an extra burden during backporting.
It seems the idea for switching to the clang-format for core packages is not so bad. However, it seems nobody has time for this effort currently.

[fujitatomoya]

i know this is still draft, but AFAIS those fix the uncrustify warnings. @MichaelOrlov 's comment needs to be discussed... 👓

[clalancette]

So we still need to land ament/ament_lint#475, which smooths over at least some of these differences. I need to get back to that one, hopefully tomorrow.

[MichaelOrlov]

@clalancette Thanks for the update. In this case I would hold off from moving forward with this PR development.

[clalancette]

OK, so I ended up merging in ament/ament_lint#475 today. With that in place, as far as I can tell there are no uncrustify errors here. @ahcorde Can you confirm that, and if that is the case for you, close this PR? Thanks.

[ahcorde]

@clalancette I runned again the CI but I can still see some linter issues https://build.ros2.org/job/Rpr__rosbag2__ubuntu_noble_amd64/44/

[clalancette]

@clalancette I runned again the CI but I can still see some linter issues https://build.ros2.org/job/Rpr__rosbag2__ubuntu_noble_amd64/44/

So the Rpr job is going to show the problem until we do a release of ament_lint into Rolling; I hope to do that early next week.

Can you test locally?

[MichaelOrlov]

@ahcorde According to the other job failures, it seems we have discrepancy only in two files for now. In the rosbag2_transport/logging.hpp and rosbag2_transport/test_record.cpp.

 >>>
2024-04-09T17:12:44.2346707Z build/rosbag2_transport/test_results/rosbag2_transport/uncrustify.xunit.xml: 57 tests, 0 errors, 2 failures, 0 skipped
2024-04-09T17:12:44.2347526Z - rosbag2_transport.uncrustify src/rosbag2_transport/logging.hpp
2024-04-09T17:12:44.2347947Z   <<< failure message
2024-04-09T17:12:44.2348194Z     Diff with 30 lines
2024-04-09T17:12:44.2348436Z   >>>
2024-04-09T17:12:44.2348815Z - rosbag2_transport.uncrustify test/rosbag2_transport/test_record.cpp
2024-04-09T17:12:44.2349248Z   <<< failure message
2024-04-09T17:12:44.2349567Z     Diff with 16 lines
2024-04-09T17:12:44.2349799Z   >>>
2024-04-09T17:12:44.2349917Z 

However, need to make a fresh run on Rolling for sure.

[clalancette]

@ahcorde According to the other job failures, it seems we have discrepancy only in two files for now. In the rosbag2_transport/logging.hpp and rosbag2_transport/test_record.cpp.

Hm, that shouldn't be anymore. I believe that at the moment, we should be uncrustify clean. I'm running a job right now on Noble to check.

[MichaelOrlov]

• I've tried to reproduce on the latest Rolling build locally and my uncrustify is happy.
  The errors reported before

2024-04-09T17:12:44.2346707Z build/rosbag2_transport/test_results/rosbag2_transport/uncrustify.xunit.xml: 57 tests, 0 errors, 2 failures, 0 skipped
2024-04-09T17:12:44.2347526Z - rosbag2_transport.uncrustify src/rosbag2_transport/logging.hpp
2024-04-09T17:12:44.2347947Z   <<< failure message
2024-04-09T17:12:44.2348194Z     Diff with 30 lines
2024-04-09T17:12:44.2348436Z   >>>
2024-04-09T17:12:44.2348815Z - rosbag2_transport.uncrustify test/rosbag2_transport/test_record.cpp
2024-04-09T17:12:44.2349248Z   <<< failure message
2024-04-09T17:12:44.2349567Z     Diff with 16 lines
2024-04-09T17:12:44.2349799Z   >>>
2024-04-09T17:12:44.2349917Z 

was on the old branches before #1578

• Closing this PR as will not be merged since all uncrustify issues has already been addressed.