Markus Vieth

Basically what I described in my comment above: Write more tests to show that the current behaviour is wrong, then fix the wrong behaviour so that the new tests pass....

@sourav25998 I believe this issue is more complex than it initially appeared. So it might not qualify as a "good first issue" and you may want to look for another...

I guess this is something we can think about. It would be possible to again have an empty cloud after saving and loading from PCD file. Important is that the...

> @mvieth May I work on this ticket? @bottlenome That would be great! > I'm thinking of making changes to writeASCII, writeBinary, writeBinaryCompressed and PCDReader. Sounds reasonable. Here is an...

Do you give any (command line) arguments to `narf_keypoint_extrraction.exe`? Which PCL version do you use and how did you install it?

I think adding tests for the new field types to `test/io/test_io.cpp` would be a good first step.

Thank you for suggesting this idea. I will investigate it further, but here are my first thoughts: - Can you try the printf with more digits, e.g. `%.20g`? Default is...

I am trying to weigh the possible benefits of switching from cosine to sine versus the complexity of the necessary code changes. If we switch, `setTransformationRotationEpsilon` has to work the...

Either way seems good to me, feel free to create a pull request with whichever solution you think is best. Are you aware of a file in PCL where this...

Do you have the same problem if you use `IterativeClosestPoint` instead of `IterativeClosestPointWithNormals`? Have you tried scaling the cloud up by 10 (decimeter) or 100 (centimeter)?