Sergey Alexandrov

Do I understand right that the correspondences are correct and we can narrow down the problem to visualization?

No Bash script this time, sorry. From my side the process always involved two PRs with a tag in between: * 1st PR: * Bump version to x.x.x in CMakeLists.txt...

I would suggest you to try compiling the library from the latest source. Then try to apply the fixes proposed in that [thread](http://dev.pointclouds.org/issues/463).

Generally, I would say keep the behavior and adjust the documentation, otherwise someone's code may be broken. In this case, however, if I understand you correctly, this class has been...

In the old SVN times devs were committing their in-progress work directly to master. Seems like this stick model is such a half-baked feature. What would you propose: remove entirely...

I guess you are now the leading stick model expert. Please update us with your further findings so we can decide what to do with it.

Also, the existing tests should be updated to test the classes instantiated with both `float` and `double` precision.

I don't see a safe way to fix it. The PCD file format has a concept of version, so in theory we can introduce a new revision `.8` where data...

Our ["official" "specification"](http://pointclouds.org/documentation/tutorials/pcd_file_format.php) does not even talk about `binary_compressed` storage format :disappointed:

I agree that PCL should fail gracefully when the file is too large to be saved in binary compressed format. However, I don't think that automatically saving in uncompressed format...