iDynTree 0.11.0 (2018-08-31)

iDynTree 0.11 Release Notes

Summary

• iDynTree now official supports the following compilers: Microsoft Visual Studio 2015/2017, GCC >= 5.3 and clang >= 3.8 .
• The C++14 standard language is now used in the project, including in the public headers.
• libxml2 is now a required dependency. Check the docs to see how to install it.
• A new URDF parser was implemented. A side effect of the new parser is that the serialization of the links and joints of models
  parsed from URDF may change. In your code, please use name string (in place of indices) to identify joints and links, or use the
  iDynTree::ModelLoader class to load a Model with the specified joint serialization.
• CMake 3.5 is now required to build iDynTree.
• This is the last release of iDynTree that supports YARP 2.3.72 . From iDynTree 0.12, the optional part of iDynTree that depends on YARP will require YARP 3.

Important Changes

core

• Improve SpatialInertia documentation ( #435 ).

sensors

• The iDynTree/Sensors/SixAxisFTSensor.h header has been deprecated in favor of the iDynTree/Sensors/SixAxisForceTorqueSensor.h.

model_io

• Added dependency on Gnome libxml2, see next bullet point.
• Added library idyntree-modelio-xml to parse generic XML files. Support for XSD validation (when parsing from file). This library requires Gnome libxml2.
• Rewritten URDF parser to use the new idyntree-modelio-xml library.
• Dropped required dependency on TinyXML1. If TinyXML is found in the system, it is possible to compile the old URDF parser that has been renamed into idyntree-modelio-urdf-legacy.

core

• Added toEigen methods for Span and added typedef for Eigen maps of vectors.
• Added some typedefs to VectorDynsize and VectorFixSize classes to enable the use of make_span with these objects.
• Added copy operator in VectorDynSize and VectorFixSize for Span<double>.

high-level

• Added method to obtain relative Jacobians sparsity pattern
• Added method to obtain free floating Jacobians sparsity pattern

inverse-kinematics

• Constraints Jacobian now exploit sparsity pattern