- Remove initFeet()
- Remove supportfoot_t variables
- Change naming convention of queues

- Replace supportfoot_t by support_state_t

To make sure that -Werror is quite.

They are some slight differences of solution with the new version,
but nothing noticeable.

- Remove: setNbVariables(), clear(), setDimensions(), initialize()
- Hide some methods
- Slight change of members naming style

+ Minor bugfix

- Add automatic adaptation of problem size

- replace raw pointer arrays by the new class
- add construction/destruction
- add public methods: fill,resize, stick_together
- add public members array_, id_, nrows_, ncols_

- Add accessors to the number of equality constraints and the total number
- Add documentation
- Minor changes

- Remove
BuildingConstantPartOfConstraintMatrices
buildConstraintMatricesPLDPHerdt
InitConstants
computeCholeskyOfQ

- other minor cleanings

- Make solver parameters of qp-problem private
- Add solution class to qp-problem
- Add methods resize, dump and print to the solution class
- Improve documentation

- m_Alpha ... m_Gamma
- debugConstructor
- InitializeMatrixPbConstants
- SetAlpha ... GetBeta
- dumpProblem
- buildConstraintMatrices
- ODEBUGS
- debugMatrix
- Matrix/VectorType
- m_PPu ... m_Uc
- computeLinearSystem

- Add documentation

- Dumping of boost matrices now done by IntermedQPMat

- Remove methods addTerm for direct manipulation of problem arrays
(now done by qp-problem itself)
- Remove dumpProblem (same as above)
- Bugfix
- Documentation
- Minor changes

- Add resize() to linear_inequality_t
- Minor changes

- Add methods
buildInequalitiesCoP,
buildInequalitiesFeet,
buildConstraints,
buildConstraintsCoP,
buildConstraintsFeet.
- Use of single sparse matrices instead of queues of small compact ones.
- Add method for initialization of matrices.
- Minor changes

- Add struct dynamics
- Add members for dynamics and linear inequalities
- Add accessors
- Add selection matrices
- Add operator and method for dumping

instead of variables of a custom type

- Compressed matrices x.D and y.D and dense vector dc
- clear() method for zeroing

- Add methods clear(array) and clear(array, range)
- Add size verification of added matrices
- Minor changes

generateIneqFeet(...)
generateIneqCOP(...)

- Introduce structs for zmp and feet convex hulls

- Add type convex_hull_t
- Add method resize(int&)
- Add method reset(void)
- Add method set(double*,double*)
- Add method rotate(double&)

- Use of ublas::compressed_vector to respond to sparsity of constraints