Unpin cppcheck 1.90

[nuclearsandwich]

Reverts #472

Related to ros2/ros2#942

[claireyywang]

Update: I still need to add macro dependency for a bunch of repos reported failure in cppcheck in the original issue. ros2/rclpy#577 proves adding additional dependency solves the immediate problem, but will require a more general solution later on to improve performance.

[claireyywang]

PRs opened and in review
ros2/rosidl#505
ros2/rosidl_typesupport#78
ros2/rosidl_typesupport_fastrtps#49
ros2/rosidl_typesupport_connext#57

Should be good to test out this PR once they are merged

[ivanpauno]

@claireyywang friendly ping.
are the linked PRs ready for review?

[claireyywang]

@ivanpauno yup!

[ivanpauno]

I've approved them all.
They need fresher CI and a double check of those windows/macOS failures (if they are still happening) before merging.

If your PRs aren't getting a review, please pick a random reviewer or ask for one in a waffle meeting.
Thanks!

[claireyywang]

@nuclearsandwich Im trying to run CIs against this branch but both OSX https://ci.ros2.org/job/ci_osx/9902/console and Windows https://ci.ros2.org/job/ci_windows/12058/console are failing. It seems to have something to do with the rolling release. Can you please update this branch and resolve the conflicts?

[nuclearsandwich]

macOS
Windows

[nuclearsandwich]

@claireyywang it looks like your rosidl_typesupport_fastrtps branch has conflicts that need attention. You should be able to re-trigger the two builds above to retry.

[clalancette]

With recent changes, here is another try on CI with cppcheck unpinned:

Windows

[clalancette]

At this point, I think we've determined that cppcheck > 1.9 isn't really salvageable. So we've actually disabled it in ament_lint instead. With that in mind, I'm going to close this out as I don't think there is a path forward here. Feel free to reopen if you disagree.