Improvements:

• substantial improvement of the cloud merge operation (thanks to Thomas Watson)

• symbolic link files (or shortcut or alias) should now be properly handled

• command line:

  • increase the timestamp resolution of the registration matrix filename and best fit plane information filename so as to avoid overwriting them if generated too quickly
  • new sub-option -NO_LABEL after -O
    • prevents any label from being loaded/created automatically (in case text columns are present in the input file)
    • for ASCCI files only

• Korean translation updated (thanks to Yun-Ho Chung)

• The Animation plugin now uses ffmpeg 6.1

• The 'Normals computation' dialog should remember whether normals 'orientation' should be resolved or not

• PCD files can now be loaded or saved with local characters. PCD files will also be saved as compressed files by default.

Bug fixes:

• The LAS dialog could be be wrongly initialized with a point format of 0 in some cases (with FWF data).
  • In command line this could result in missing waveforms when saving. Thanks to Paul Leroy for the fix ;)
• The Rasterize tool was not letting the user use '0' as the max edge length parameter for the Delaunay-based raster interpolation mode (forcing the user to set a high value to keep all triangles)
• The scale value in the bottom right corner of the 3D view (orthographic mode) was wrong if the screen height was larger than the width
• After the -FEATURE command line command was run, the automatically exported filename of clouds was containing some duplicated contents
• The Poisson Reconstruction progress dialog 'cancel' button was ineffective (the process cannot be canceled). It is now hidden.
• Per-vertex colors were never saved in a Maya (MA) file
• Loading a corrupted STL filter could result in a corrupted mesh or a crash
• The Compass plugin had some minor but numerous memory leaks
• The torus primitive mesh topology was broken
• LAS files: the synthetic flag could be mistakenly set at save time if some non-zero classification values were present
• The envelope/contour extraction routine of the Cross Section tool could fail or crash in some cases
• Canupo: some scalar fields were not properly removed/cleaned in some cases