Rename file `test_write_eddy.f90` to `test_write_null.f90`. The name
was badly chosen because the program is not a unit test of procedure
`write_eddy`. In fact, it does not even call `write_eddy`.

Add executable in the directoy containing the main program unit. This
is clearer as the executable file will be created in the corresponding
binary directory. Since the directory Overlap is added after
`Inst_eddies`, targets defined in `Overlap/Tests` are not yet defined
when processing `Inst_eddies`, so we cannot use `target_sources` for
these targets in `Inst_eddies`. So we use a relative path to
`Inst_eddies` from `Overlap/Tests`.

Update `CMakeLists.txt` files for `test_get_dispatch_snap`.

Update `CMakeLists.txt` files to take into account commit 3289c070.

Update `CMakeLists.txt` files to take into account commit 3289c070.

Remove procedure `write_aux`. Write to `ishape_last.txt` in procedure
`write_snapshot` and write `grid_nml` in procedure `input_ssh`. In
`input_ssh`, if `grid_nml.txt` already exists, check consistency with
input NetCDF files. Downgrade dummy arguments `corner_deg` and
`step_deg` of `input_ssh` to local variables. Read and write
`grid_nml` at SHPC level, only one file for both orientations and all
slices.

Extract code from program `inst_eddies` into new procedure `input_ssh`.

With `target_sources` rather than referring to their path from another
directory.

It was difficult to find the `CMakeLists.txt` that controlled the the
compilation of a given file from the message of a compilation error.

Since `NR_util` has been integrated into Jumble.

We need this since `inst_eddies_Aviso.py` imports `inst_eddies` which
is in the binary directory.

Rather than with Unix-specific chmod command. This requires CMake `3.20`.

Following convention for the source of `configure_file`. See Scott
2021 k1077, § 19.7.

Following updates of the libraries.

Following updates of the libraries.

Note that programs `test_max_speed_contour_ssh` and `test_weight`
depend on fortranGIS through the module derived_types. Compilation
worked without linking to fortranGIS because objects used by
`test_max_speed_contour_ssh` and `test_weight` in
derived types do not use fortranGIS.

Instead of creating them. Motivation: avoid time spent concatenating
shapefiles in `inst_eddies.py`. This concatenation time, on Ciclad, is
of the same order of magnitude as the time spent by the Fortran
executable.

In case we are a sub-project.

Rename triplet to collection, `shp_tr` to shpc, shpt to shpc in
comments, identifiers and file names.

`read_eddy.f90`, `read_field_indices.f90` and `shp_tr_open.f90` will
become common to Overlap and `Inst_eddies` because they will be used
in `Inst_eddies/Tests/test_set_max_speed.f90`. So there are now enough
common files to justify a new directory.

Preparing for two collections of shapefiles. We create the
procedure `write_nml` so we will be able to call it twice. But it
needs to access the namelist. So the namelist must be at the module level.

Extract writing of `ishape_last.txt` and `grid_nml.txt` to new
procedure `write_aux`. Preparing for two collections of shapefiles.

Preparing for configuration of location of binary executable.

Following update of libraries: NetCDF-Fortran is private in NetCDF95
and MPI-Fortran is private in EZMPI.

Also split the tests, Fortran source files, Python files and JSon
files. Keep at the top level the files used in both `Inst_eddies` and
Overlap.

Motivations for the split:

- The top directory contained many files.

- It may be useful to compile only one of the two sub-projects. For
example, only `Inst_eddies` on Ciclad, where MPI 3 is not available.