change terminology of static and (un)bounded dynamic arrays

[dirk-thomas]

When we wrote the design article about interface definitions we needed to distinguish three different kind of arrays and came up with these terms:

• "static array": arrays with a static size
• "dynamic unbounded arrays": arrays with a dynamic size and no upper boundary
• "dynamic bounded arrays": arrays with a dynamic size and an upper boundary

In preparation of the migration to IDL we should change our terminology to match the specification.

• array equates to "static array" from above
• (unbounded) sequence equates to "dynamic unbounded arrays" from above
• bounded sequence equates to "dynamic bounded arrays" from above

This ticket will serve as a meta ticket for PRs renaming types, functions, variables, files, etc.

[dirk-thomas]

All referenced PRs are ready for review.

• Linux
• Linux-aarch64
• macOS
• Windows

[wjwwood]

Renaming sounds fine to me.

[mjcarroll]

Looks like there were a few places where array -> sequence caused us to violate line lengths.

[jacobperron]

Should ROS message types in rcl_interfaces/test_msgs be renamed for consistency?

[dirk-thomas]

• Fixed the header guard in ros2/rosidl@14a5b3e#diff-e120ebd828429d798c09b4002e0667aeR15.
• Fixed the line length in ros2/system_tests@6157b1f#diff-831fe03f2d54d3547eae0bc2f01db942R1158.

Should ROS message types in rcl_interfaces/test_msgs be renamed for consistency?

That could be done. Since it doesn't affect the API touched by the IDL work I didn't include it in this set of patches.

[mjcarroll]

All look good to me provided that those few linting issues are resolved.

[dirk-thomas]

And another linter fix: ros2/rosidl@2d11c37#diff-e989a753d48150acb8507440277b41b4R15

CI only running the linter of the failing test: