Alexis DUBURCQ

This issue has been addressed. Closing. https://github.com/duburcqa/jiminy/pull/784 https://github.com/duburcqa/jiminy/pull/786 https://github.com/duburcqa/jiminy/pull/787 https://github.com/duburcqa/jiminy/pull/792

Ideally, it should be possible to instantiate standalone sensors without shared storage, then update and fetch sensor measurement in isolation. The manager should only be helper that provides a bunch...

Just add an extra requirement `cmeel` that install all these libs with the expected versions, so that the user has to opt-in a to get them. This way, building everything...

> Add option to disable extra computations The speed up was very marginal (less than 2%), so I think it is not worth adding an extra option for this. Closing.

It appears to be much slower than expected, at least to load first render meshes. Once fully loaded it is reasonably fast on both integrated and dedicated GPU though. Full...

For the record, the PR https://github.com/duburcqa/jiminy/pull/826 is bringing many optimisation regarding panda3d rendering speed, for both offscreen and onscreen rendering. So it is even more likely today than replacing it...

Note that it is still relevant to create theoretical visual and collision models (at least internally), as they are used to generate the actual visual and collision models later on....

Fixed by https://github.com/duburcqa/jiminy/pull/753

I made some progress thanks to numpy team. I'm just having a segfault in one of the unit tests: ```diff diff --git a/include/eigenpy/numpy-allocator.hpp b/include/eigenpy/numpy-allocator.hpp index 6165394..58f6415 100644 --- a/include/eigenpy/numpy-allocator.hpp +++...

Done ! Opening a PR.