Moved messages to trintrin, plus some cleanup

[S-Dafarra]

In this PR:

• removed deprecated HapticGlove device. Now it is in https://github.com/ami-iit/yarp-device-haptic-gloves/tree/8edbffbedebc907e585852e0a62e7059da8a4f24/SenseGlove
• removed the messages and the bindings. They have been moved to https://github.com/ami-iit/trintrin/ that now is a dependency
• updated the CODEOWNER file
• updated the README
• did some cleanup of Paexo and XSens devices. The latter I could not try compiling since I don't have the SDK available
• fixed iDynTree deprecations

[S-Dafarra]

@davidegorbani @Gianlucamilani I am not sure if these changes will affect your workflow. Let me know if there are issues

[traversaro]

@davidegorbani @Gianlucamilani I am not sure if these changes will affect your workflow. Let me know if there are issues

For sure we need to add trintrin in the superbuild before merging this. If this PR works, can yo do a first tag of trintrin to simplify superbuild integration?

[S-Dafarra]

For sure we need to add trintrin in the superbuild before merging this. If this PR works, can yo do a first tag of trintrin to simplify superbuild integration?

Sure! https://github.com/ami-iit/trintrin/releases/tag/v0.0.1

[davidegorbani]

@davidegorbani @Gianlucamilani I am not sure if these changes will affect your workflow. Let me know if there are issues

It shouldn't, as we switched to biomechanical-analysis-framework, and we don't currently use this repo.

[traversaro]

Also, I guess we need to update also internal Python repos using import hde and import wearables to import trintrin, perhaps we want to do that so that we can add here a minimal migration guide?

[S-Dafarra]

Also, I guess we need to merge also internal Python repos using import hde and import wearables to import trintrin, perhaps we want to do that so that we can add here a minimal migration guide?

I did not understand the suggestion 😊

[traversaro]

Also, I guess we need to merge also internal Python repos using import hde and import wearables to import trintrin, perhaps we want to do that so that we can add here a minimal migration guide?

I did not understand the suggestion 😊

Sorry, I wanted to write "update" not "merge". I guess you want to update those repos, and the update is something like:

In python code, change any update of import wearables and import hde to import trintrin and any occurence of hde. or wearables. to trintrin.

[traversaro]

Related PRs:

• Add trintrin package robotology-superbuild#1910
• Add recipe for the trintrin C++ and Python package conda-forge/staged-recipes#31251

[S-Dafarra]

Related PRs:

• Add trintrin package robotology-superbuild#1910
• Add recipe for the trintrin C++ and Python package conda-forge/staged-recipes#31251

Once this is done, I can remove the source installation of trintrin in

human-dynamics-estimation/.github/workflows/ci.yml

Lines 49 to 57 in 6a525a7

	- name: Install trintrin from source
	shell: bash -l {0}
	run: |
	git clone https://github.com/ami-iit/trintrin
	cd trintrin
	mkdir build && cd build
	cmake -G Ninja -DCMAKE_BUILD_TYPE=${{ matrix.build_type }} ..
	cmake --build . --config ${{ matrix.build_type }}
	cmake --install . --config ${{ matrix.build_type }}

[traversaro]

As now the robotology-superbuild supports this PR, probably we can merge to unblock the following PR?

[S-Dafarra]

As now the robotology-superbuild supports this PR, probably we can merge to unblock the following PR?

Do you believe that the conda package will land in short time (btw, thanks a lot, it is great!)? If yes, w can wait to remove the source installation from CI and be consistent with the other dependencies

[traversaro]

Do you believe that the conda package will land in short time (btw, thanks a lot, it is great!)? I

No. There is a critical lack of reviewers on staged-recipes, so the PR could take a long time to be merged. I would proceed with the chain of PRs to avoid being blocked, the only thing that I would block on the trintrin conda package being available is the new release of HDE.

[S-Dafarra]

Do you believe that the conda package will land in short time (btw, thanks a lot, it is great!)? I

No. There is a critical lack of reviewers on staged-recipes, so the PR could take a long time to be merged. I would proceed with the chain of PRs to avoid being blocked, the only thing that I would block on the trintrin conda package being available is the new release of HDE.

Ok, then I will edit the CIs to make them happy while waiting for the conda packages to be availalbe

[traversaro]

Do you believe that the conda package will land in short time (btw, thanks a lot, it is great!)? I

No. There is a critical lack of reviewers on staged-recipes, so the PR could take a long time to be merged. I would proceed with the chain of PRs to avoid being blocked, the only thing that I would block on the trintrin conda package being available is the new release of HDE.

Ok, then I will edit the CIs to make them happy while waiting for the conda packages to be availalbe

Ah if the problem is making the CI happy, we can also build a first temporary trintrin package to the robotology channel to unblock the CI, that will be superseded by the conda-forge one once conda-forge/staged-recipes#31251, if you give me the go I can do that easily locally.

[S-Dafarra]

Do you believe that the conda package will land in short time (btw, thanks a lot, it is great!)? I

No. There is a critical lack of reviewers on staged-recipes, so the PR could take a long time to be merged. I would proceed with the chain of PRs to avoid being blocked, the only thing that I would block on the trintrin conda package being available is the new release of HDE.

Ok, then I will edit the CIs to make them happy while waiting for the conda packages to be availalbe

Ah if the problem is making the CI happy, we can also build a first temporary trintrin package to the robotology channel to unblock the CI, that will be superseded by the conda-forge one once conda-forge/staged-recipes#31251, if you give me the go I can do that easily locally.

Yeah I was planning to compile trintrin from source where needed, but your plan would be great! Go for me!

[traversaro]

Yeah I was planning to compile trintrin from source where needed, but your plan would be great! Go for me!

Here you are:

• https://anaconda.org/robotology/libtrintrin
• https://anaconda.org/robotology/trintrin-python
• https://anaconda.org/robotology/trintrin

Use libtrintrin if you just depend on the C++ part, while trintrin-python or trintrin if you also depend on the Python bindings.

I just added them for win-64 and linux-64, feel free to ask if you need more architectures.

[S-Dafarra]

Awesome!

I just added them for win-64 and linux-64, feel free to ask if you need more architectures.

Is it possible also macOS?

[traversaro]

I just added them for win-64 and linux-64, feel free to ask if you need more architectures.

Is it possible also macOS?

I currently disconnected my macOS to connect the StrixHalo, but I can find a solution.

[traversaro]

I will build for osx-arm64, if we have CI still targeting osx-64 we can drop them as the Intel-based macOS systems are quite outdated and I am not aware of anyone using them.

[traversaro]

I will build for osx-arm64, if we have CI still targeting osx-64 we can drop them as the Intel-based macOS systems are quite outdated and I am not aware of anyone using them.

Done:

• https://anaconda.org/robotology/libtrintrin
• https://anaconda.org/robotology/trintrin-python
• https://anaconda.org/robotology/trintrin

[S-Dafarra]

I am having some issues with CI. I added the robotology channel in the env file, but I also wanted to set the channel priority so that when we have a version in both channels I don't have to update the CI. So in b8d306c I moved from setup-micromamba to setup-miniconda since it has the possibility to specify the channel priority and we have been moving out from setup-micromamba in the past. The problem then is that ilammy/msvc-dev-cmd stopped working, so I tried removing it from the CI with a49b944. But then, I had to differentiate the install prefix between Windows and Linux. Moreover, ninja stopped working on Windows. So, I split the configuration phases between Windows and linux/macOs in 1c17d66. At this point cmake-package-check started failing on Windows, since it is using ninja under the hood. So, I tried to:

• reintroduce ilammy/msvc-dev-cmd with an updated version in fcbd6ef, but failed anyway
• using the vs2022_win-64 conda package in windows, hoping it could do the same job of ``ilammy/msvc-dev-cmd`, but it did not.
  Hence, at the moment the error on Windows is

-- The CXX compiler identification is unknown
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - failed
-- Check for working CXX compiler: C:/mingw64/bin/c++.exe
-- Check for working CXX compiler: C:/mingw64/bin/c++.exe - broken
CMake Error at C:/Users/runneradmin/miniconda3/envs/test/Library/share/cmake-4.1/Modules/CMakeTestCXXCompiler.cmake:73 (message):
  The C++ compiler

    "C:/mingw64/bin/c++.exe"

  is not able to compile a simple test program.

  It fails with the following output:

    Change Dir: 'D:/a/human-dynamics-estimation/human-dynamics-estimation/build/CMakeFiles/CMakeScratch/TryCompile-rwg1xo'
    
    Run Build Command(s): C:/Users/runneradmin/miniconda3/envs/test/Library/bin/ninja.exe -v cmTC_1b52a
    [1/2] C:\mingw64\bin\c++.exe    -o CMakeFiles\cmTC_1b52a.dir\testCXXCompiler.cxx.obj -c D:\a\human-dynamics-estimation\human-dynamics-estimation\build\CMakeFiles\CMakeScratch\TryCompile-rwg1xo\testCXXCompiler.cxx
    FAILED: [code=1] CMakeFiles/cmTC_1b52a.dir/testCXXCompiler.cxx.obj 
    C:\mingw64\bin\c++.exe    -o CMakeFiles\cmTC_1b52a.dir\testCXXCompiler.cxx.obj -c D:\a\human-dynamics-estimation\human-dynamics-estimation\build\CMakeFiles\CMakeScratch\TryCompile-rwg1xo\testCXXCompiler.cxx
    ninja: build stopped: subcommand failed.

while on Mac simply I could not get a runner, so I stopped the workflow.

[S-Dafarra]

With 2798a69 I restored the logic using the VCToolsInstallDir env variable, that I had removed in a49b944