Moved PointXYZLAB to common point types

[ueqri]

Fixes #4600

• moved PointXYZLAB definition from gicp6d.h to point_types.hpp
• added struct PointXYZLAB declaration in point_types.h
• moved function RGB2Lab & convertRGBAToLAB (refactored) from gicp6d.h to point_types_conversion.h, in line with other conversions.
• overloaded output stream operator for PointXYZLAB in point_types.cpp
• added IntensityFieldAccessor class for PointXYZLAB in intensity.hpp
  • I learned the basic knowledge of color space from this tutorial , and not very sure whether my implementation is exactly correct.

I've compiled this branch on Ubuntu 20.10 GCC and passed all unit tests 😆. I'd be grateful if you could have a review on it. Thank you!

[kunaltyagi]

You should add a git pre-commit hook to run make format or ninja format 😆 for the CI

[ueqri]

You should add a git pre-commit hook to run make format or ninja format for the CI

Thank you for reminding me of the format check 😃 in CI ! I've applied the patch to yesterday's commit, and added pre-commit hook with .dev/format.sh to ensure the behavior is the same as CI.

[ueqri]

I think I've fixed the compiling error of LUT (built by Clang Darwin and MSVC) in fa0782c, by adding the header of std::array which needs to be included explicitly especially for those compilers.

And the CI shows the branch passed {Clang macOS, GCC Ubuntu, MSVC Windows x86} build, but met error in MSVC Windows x64 build due to the lack of space (detailed log) .

Could you give me some advice about solving this error, or just let it go? BTW, I found a relevant issue #3387 and it seems to be a tricky problem, but I'm willing to work on it.:smiley: Is there any plan to tackle this issue and enhance the quality of CI? I'm really glad to do it!

[larshg]

I just restarted the win x64 build.

The space error is quite spurious and its not really space problem. We tried moving the build from the temporary D drive to the C drivce, which has about 78G free space at start.

But what causes the error, is probably a task for the DevOps people at MS. But if you have some good ideas, please go ahead.

[ueqri]

Thank you for rerunning the CI and the helpful explanations:blush:! I would try to find out the causes and solutions if possible.

[mvieth]

Thank you for working on this!