Modular pkg does not compile

The modular pkg (https://github.com/advrhumanoids/modular_hhcm) cloned inside the container does not compile.
The last commit also creates a consistency problem with the definition at lines https://github.com/ADVRHumanoids/concert_description/blob/cf02fb3476e53f71f9422777c70df769c2051d9e/concert_gazebo/launch/modular.launch#L37 https://github.com/ADVRHumanoids/concert_description/blob/cf02fb3476e53f71f9422777c70df769c2051d9e/concert_gazebo/launch/modular.launch#L38.

Can you try now? I just made some changes to integrate the sensors and I think you tried just before I was able to push.
Is it solved now on branch master?

Now it compiles, but only if I put comments on these two lines of the modular pkg:

modular_hhcm/CMakeLists.txt

Line 5 in 1deaa62

catkin_python_setup()

[catkin_python_setup()](

modular_hhcm/CMakeLists.txt

Line 13 in 1deaa62

roslaunch_add_file_check(launch)

)

However, now it works.

Actually, you shouldn't need to compile anything. The CMakeLists was there in order to use modular with catkin, which we haven't been using in a long time, so it might have broken. We will have a look at this issue.

Anyway, now the preferred way to install modular is to use pip install, as described in the README.

For concert_description we use forest to install everything (which then calls pip install for modular).
This is the forest recipe for concert_description.
To install you should run forest grow concert_description from a terminal.

We will update the README to avoid these misunderstandings.

I have to do catkin_make and compile the whole workspace to add some packages (e.g. the node that converts the robot_description parameter to the robot_description topic)

With commit 5408e1b modular is not a ROS package anymore.
(It was a ROS package before for exposing the meshes, but now we have the package modular_resources for that)
There is no CMakeLists.txt and package.xml anymore so it shouldn't give any more errors when compiling with catkin_make (I managed without errors in a test catkin workspace)
The way to install it's still with pip install.
Does this solve the issue?