Adding technique 5 to meshMetric: linear interpolation of h in band around iso, instead of linear interpolation of the metric (1/h^2)

smoothing 1D mesh is now (maybe) correct

No commit message

Add a new metric computation for level set, simply based on ls gradient and hessian eigendirections.
meshAdapt was modified to receive and intersect multiple metrics, preserving the most anisotropic metric.
Also added input and output functions to dgDofContainer, for the "NodeData" format.
Benchmark for anisotropic adaptation to level set iso-zero created in benchmarks_gmsh/LSIsozeroAniso.

No commit message

Worked a lot oncurvilinear meshing : projection of point on surafces work, matching meshes andgeometry works better, added some test examples...

i-This line, and those below, will be ignored--

M    Mesh/CenterlineField.h
M    Mesh/CenterlineField.cpp
M    Geo/discreteFace.h
M    Geo/discreteFace.cpp
M    Geo/GModel.cpp
M    Solver/multiscaleLaplace.cpp
M    benchmarks/stl/mobilette.geo

No commit message