What happened to Meshlabserver?
A few days ago, one reader comment on the blog (thank you Lance) made me aware of the changes happening with the Meshlabserver program. I misread the comment first as if the MeshLab project was discontinued, and after checking I replied back that a new released was just a few days old. But then it hit me, Lance was spot on, MeshLab project was deprecating the Meshlabserver program, now becoming "unsupported" and users were diverted to a new program called PyMeshLab which, as you all can guess, is now Python-based.
So that was interesting news for me as I have been using meshlabserver for various automation processes dealing with 3D meshes. And the even better news is that now installing it should be quite simple, just type pip3 install pymeshlab.
Linux and OSX install was uneventful and it all worked as expected. I learned that you could even use .mlx files created with Meshlab software. Windows 10 install was equally simple but, once I imported the library in Python I was warned with a longish message. I assumed it was an error and I went ahead and reported the issue on GitHub but I stand corrected: it was only a warning, and while a couple of filters will not work, the rest of the library would work ok, which it did!
I tried to find what was the cause of the issue in Windows 10, but after some tests (you can see the details here), I am not closer to a fix than I was at the beginning. Apparently, Qt5Network.dll is not loaded, but the reasons are unknown, and the usual suspect (wrong path) seems not to be the case here.
Anyway, the good news is that PyMeshLab does what meshlabserver did in a much more convenient way (if you like Python coding). And with the exception of the Windows 10 oddity I experienced, it keeps on being a fantastic multiplatform program.
While I was looking for a similar program, I came across another interesting mesh-oriented software library called libigl, which I am going to keep an eye on, as Python binding is already in beta now. I have used Pymesh (unrelated to PyMeshLab) in the past and I liked it too.
Comments