GitLab

Reason: using --unicode-literals everywhere is not a good approach.

This reverts commit 323836e9.

futurize --stage1 --unicode-literals `echo $(find -name "*.py*") ./core/main/yade-batch.in ./core/main/yade-oar.in`

- Add the possibility to calculate bending and twisting moments using the incremental formulation in both CFLaw and HMLaw;
- Add maximum elastic torsional moment in CFLaw (it should be computed using normal force, I will see how to compute it somehow);
- Fix mistake in mindlin.py example script.

2. Fix a bug in post2d

1. Rename InteractionDispatchers to InteractionLoop; backwards-compat for python as usual, typedef with deprecated warning in c++; update all scripts and docs to reflect that.
2. Add NozzleFactory, which might be in the future generalized to a generic factory object and a numebr of derived classes with different parameters (such as factory area shape, specification of sphere parameters etc); see scripts/test/shots.py
3. Fix many crashers in the OpenGL code; unbreak particle selection
4. Finalize the MatchMaker class, with convertor from python floats to specify fixed value.
5. MatchMaker is used to specify different coefficients of restitution for different material couples in Ip2_FrictMat_FrictMat_MindlinPhys; viscous constants moved away from Law2_ScGeom_MindlinPhys_Mindlin into MindlinPhys instead.
6. Fix compilation errors caused by unintelligedn 3rd-party Combine builder to scons -- save sombined files to an external file, force refresh of all combined files if the md5 changes. Does not increase compilation time.
7. Add DomainLimiter engine for deleting particles that go ouside some domain (see scripts/test/shots.py)
8. Add DragForceApplier (not yet tested); thans to Stefano for the suggestion.

1. Add Engine::dead attribute; if set, the engine will not be run at all (unless explicitly called via () from python or ::action from c++). Useful for temporarily (de)activating engines without having to modify O.engines
2. PeriodicPythonRunner renamed to PyRunner as it is used so often.
3. Add Cell.getVolume() and utils.porosity
4. Check for force&torque NaNs in NewtonIntegrator in debug builds
5. Add --rebuild option which will launch scons before running the simulation itself
6. Some fixes in the HertzMindlin law
7. The qt4 interface will try to open packaged docs, then docs in the source dir and online docs as the last resort.
8. Some fixes in PeriTriaxController (the servo-mechanism) and probably a bug fix in TriaxialStressController (will ask Bruno on the list to confirm the correctness)

1. Remove BoundDispatcher from O.engines in examples, scripts and sphinx (accidental typos possible, sorry)

- Add of a script NormalInelasticityTest to test ... the NormalInelasticityLaw (Law2_ScGeom_NormalInelasticityPhys_NormalInelasticity)
Moreover my python problems I had difficulties to obtain "perfect results". I think it is linked to the way un is computed in Ig2_Sphere_Sphere_ScGeom (see for
example this "shift2").
Why not making penetrationDepth directly accessible through python ??

- As discussed with Chiara directly, typos in mindlin_test