Silvio Traversaro

> > > Apparently the main build works fine, but it breaks if iDynTree is included in a bigger project via `add_subdirectory` . The CMake version does not matter, what...

> > > That's really complex to spot! I was lucky as I wanted to understand what was going on in the `GenerateExportHeader` file in CMake, so I added some...

For reference, the logic that was added to understand `package://` or `model://` URIs was added in https://github.com/robotology/idyntree/pull/798 .

As it turns out in: * https://github.com/ami-iit/robot-log-visualizer/issues/35 * https://github.com/robotology/whole-body-estimators/issues/147 Probably it could make sense to expose this functionality added in https://github.com/robotology/idyntree/pull/798 of resolving (i.e. transforming in absolute paths) `package://` or...

I thought this was covered by the Octave unit tests.

I have a terrible doubt. Did you loaded casadi in the same workspace in which you are having this problem? If that is the case, can you try to avoid...

Can you run the testsuite? `ctest -VV -R matlab` should be enough.

For an additional data point, it would be interesting to run the test suite also for octave: ~~~ sudo apt install liboctave-dev // Enable ROBOTOLOGY_USES_OCTAVE or IDYNTREE_USES_OCTAVE ctest -VV -R...

Cool, at some point the test suites stopped reading any test, and that is the reason why is passing also for octave, because zero tests were found and zero are...

The problems seems to be that `iDynTree.Twist.Zero()` return `SpatialVector` that is a CRTP-defined type from which Twist is derived, but that probably is not correctly understood by SWIG, and this...