Ugprade to 3.1.6

[JamieMair]

No description provided.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 0.78125% with 508 lines in your changes missing coverage. Please review.

Project coverage is 25.98%. Comparing base (0cab161) to head (d4f6f60).

Files	Patch %	Lines
src/LibMuJoCo/functions.jl	0.00%	323 Missing ⚠️
src/visualiser_wrappers.jl	0.00%	99 Missing ⚠️
src/function_constraints.jl	0.00%	50 Missing ⚠️
src/LibMuJoCo/structs.jl	0.00%	29 Missing ⚠️
src/named_access.jl	42.85%	4 Missing ⚠️
src/MuJoCo.jl	0.00%	2 Missing ⚠️
ext/VisualiserExt/defaulthandlers.jl	0.00%	1 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #109      +/-   ##
==========================================
- Coverage   27.47%   25.98%   -1.49%     
==========================================
  Files          22       22              
  Lines        8568    10097    +1529     
==========================================
+ Hits         2354     2624     +270     
- Misses       6214     7473    +1259     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[nic-barbara]

Looking great overall but some issues with finding file path for humanoid demo.

I really like the new API to show docs for Model and Data, that will be extremely useful and the Python bindings should look into doing the same :)

[nic-barbara]

@JamieMair it all works fine on my Mac (except for loading the humanoid but that's just a file path issue). However, I have an Intel Mac, not one of the new M1/2/3 chips. The build error is: MuJoCo cannot be run under Rosetta 2 on an Apple Silicon machine.

I'm pretty sure MuJoCo works fine on modern Macs, but they might be doing something funky in the CMakeLists.txt Any ideas?

[JamieMair]

I've noticed that the builds on GitHub actions also fail for a previous PR? Perhaps there is an issue with the runners themselves?

[nic-barbara]

Possibly. I'll find someone with a more recent mac in my lab and get them to test a build.

[nic-barbara]

Can confirm that all tests pass on my colleague's mac, M1 (maybe M2) chip with Mac OS 14.5. Not sure what's failing in the CI tests.

[JamieMair]

Can confirm that all tests pass on my colleague's mac, M1 (maybe M2) chip with Mac OS 14.5. Not sure what's failing in the CI tests.

It turns out that Rosetta 2 is a translation layer to let ARM MacOS computers run x86 code, which is specifically disallowed in MuJoCo. The CI was set to "macos-latest" which is now macos-14 which is an arm based system, but the arch was set to x86_64, which causes Julia to load the MuJoCo_jll version that is built for x86_64, not the ARM version which is the problem. I have set the build to use macos-13 for x86_64 and macos-14 for ARM builds, but currently Julia actions breaks with aarch64, so I'm removed the CI for ARM for now.