Prevent crash in TimeStampHandler for 2D and 3D simulated laser data without/identical timestamp fields (fixes #31)

[MikeDegany]

Summary

This PR fixes a crash in the TimeStampHandler when processing 3D or 2D laser data in simulations. In scenarios where the PointCloud2 message does not include a per-point timestamp (common with simulated 3D laser scanners), and where the PointCloud2 messages have identical timestamps (common with simulated 2D laser scanner), the current implementation returns an empty vector and later attempts to dereference iterators from it, resulting in a crash.

Changes Introduced

Early Check for Empty Timestamps:
Added an explicit check for an empty timestamps vector in ProcessTimestamps. If no timestamps are found (i.e., the timestamp field is missing), the function now logs a warning and returns early without attempting normalization.

Strict Equality for 2D Data:
Retained the strict equality check (*max_it != *min_it) for normalization when timestamps are present. This ensures that 2D laser data is processed correctly, while bypassing normalization only when all timestamps are identical.

Impact

3D Laser Data:
The fix ensures that 3D laser data in some simulation scenarios—which often lack per-point timestamps—does not lead to crashes, while scan deskewing is disabled gracefully.
2D Laser Data:
It occures in 2D laser data in some simulation scenarios, where timestamps are the same resulting in a crash exit code -6. The fix bypass the normalization in such cases.

The behavior remains unchanged, with normalization occurring only when a genuine timestamp range exists and changes.

Rationale

By handling the absence of timestamp data explicitly, and checking if timestamps are not identical, this change improves the robustness of the kinematic-ICP package and prevents runtime crashes due to invalid iterator dereferencing when users run kinematic-icp on simulation data and logs appropriate warning in each case. This approach is consistent with the intent to only process meaningful timestamp variations and gracefully handle missing data.

[benemer]

Looks good to me!

[benemer]

@tizianoGuadagnino can you please have a look at this PR?

[edelwiw]

Thank you!