GitLab

#1468 The case without sections at all work. However several other test cases fail; I have to keep robustifying it (all the problematic cases will of course be added to the test suite).

#1315 Apply the rewriting of Lua files to all embedded models. A setprecision has been added to ensure keeping the double precision during the process.

#1315 I have added with the writing of a Lua file the possibility to rejuvenate a Lua file and make sure its comments are up-to-date. It remains to be propagated to all the Lua files.

#1465 Rewriting a Lua file almost works as expected; only heterogeneous vector is not yet worign completely as expected.

#1470 Introduce a policy for PrintContainer, which may handle different cases with the same functions. This will be required to fulfill
#1465.

#1470 Reorganize the Utilities/Containers/Print in several files to conform with the layout of the rest of the code.

#1465 The mechanism to write the input data is now in place. I now must adjust the outputs (add the quots for strings for instance...) and deal with the more complicated cases I have left behind at the moment (containers and variants - especially containers with variants inside...).

#1465 Pave the way to the extension that will enable writing the content of a built object rather than the default values.

#1467 Following #1432, there is no need to keep track of the LuaOptionFile used to interpret input data file.

#1465 Add a test to help implement the writing down of a Lua file from InputData with non-default values inside.

#1466 Make RAII initialization of Petsc and mpi an (internal) independant class - a singleton in fact. This way, more than one instance of MoReFEMData may be run in the same program.

New way to create the Docker images uses up only one Dockerfile, and the call to `docker build` must occur within standard MoReFEM directory.

A README in Docker directory explains precisely how to create images.