tools/nxstyle: Whitelist ROS 2 message type identifiers.

[arjav1528]

Summary

The ROS 2 C client API exposes message types through rosidl token-pasting macros. Each message type ends up with a generated symbol of the form <pkg>__msg__<Type> (e.g. std_msgs__msg__Int32), and the bare type token (Int32, Float64, ...) appears in user code as an argument to ROSIDL_GET_MSG_TYPE_SUPPORT(). Neither form can be renamed without breaking the public API.

Add the ROS 2 standard message package prefixes and the rosidl scalar type names to g_white_prefix so nxstyle does not flag the resulting mixed-case identifiers in apps that consume micro-ROS via apps/system/microros.

Whitelisted prefixes:

• Packages: std_msgs, geometry_msgs, sensor_msgs, nav_msgs, builtin_interfaces, rcl_interfaces.
• Scalars: Int8, Int16, Int32, Int64, UInt8, UInt16, UInt32, UInt64, Float32, Float64.

Follows the existing precedent for external ecosystems in the same list (cJSON, NativeSymbol, Ifx, lua_, TEEC_, ...).

Impact

• tools/nxstyle only. No build, runtime, or hardware effect.
• Pure relaxation of the style checker for a fixed, enumerated setof identifiers. No existing in-tree code becomes invalid; only previously-flagged ROS 2 token-pasted identifiers stop being flagged.

Dependencies

Companion to apache/nuttx-apps#3512 (system/microros: Add UDP and serial custom transports with publisher example). That PR's CI nxstyle step fails on apps/examples/microros_pub/microros_pub_main.c until this lands.

Testing

Local build of the patched tools/nxstyle and run against every C/H file touched by apache/nuttx-apps#3512:

$ gcc -O2 -o tools/nxstyle tools/nxstyle.c
$ for f in <8 touched .c/.h paths>; do ./tools/nxstyle "$f"; done
(no output -> all clean)

Before this patch (same nxstyle build minus the whitelist additions):

apps/examples/microros_pub/microros_pub_main.c:45:7: error: Mixed case identifier found
apps/examples/microros_pub/microros_pub_main.c:84:51: error: Mixed case identifier found

After: 0 errors, 0 warnings on the same 8 files.

No other in-tree files reference any of the added prefixes, so the
whitelist relaxation is scoped to the new ROS 2 surface.

[linguini1]

Should we allow modifying nxstyle for third party libraries?

[acassis]

Should we allow modifying nxstyle for third party libraries?

Normally we don't it, because it makes the update to a new library version more difficult.