In program `test_mean_speed`. To make it work with a non-uniform grid,
we would have to read a contour in projection coordinates.

Use `uniform_lon_lat` from modules `input_ssh_m` or `read_grid_m` to
compute actual argument of `shpc_create`.

In program `examine_eddy`.

Replace dummy argument `grid_lon_lat` of procedure `shpc_create` by
`with_proj`.

Specify whether to read shapefile `extr_proj` in procedure `shpc_open`
instead of just trying to read this shapefile. Use `uniform_lon_lat`
to decide whether to read the shapefile. As `shpc_open` is called in
both programs `inst_eddies` and `eddy_graph`, we cannot access
variable `uniform_lon_lat` of module `input_ssh_m`, which is not used
in program `eddy_graph`, nor variable `uniform_lon_lat` of module
`read_grid_m`, which is not used in program `inst_eddies`. So we add a
dummy argument `with_proj` to procedure `shpc_open`.

Rename variable `corner` of module `input_ssh_m` to `corner_whole`.

I did not like the difference of treatment between corner and step.

Preparing to read a second input variable from stdin in procedure
`input_ssh`.

Instead of deducing it from the longitude coordinate. The deduction
could not be generalized to a non-uniform longitude-latitude
grid. Also, the test on `step(1)` and nlon was not perfectly robust.

We are trying to simplify the logic in program `inst_eddies`. Also,
the program `inst_eddies` should not write `grid_nml.txt` to the top
level of the SHPC directory since several instances could run in
parallel for several slices. So we are leaving it as a task for a
script above `inst_eddies` to create `grid_nml.txt` at the top level
of the SHPC directory. In tests, we have to move `grid_nml.txt` to
SHPC because we use `SHPC/grid_nml.txt` in other tests.

We are reverting commit d4d05bc2. It was not true that we needed two
input fluxes in the Fortran program. `inst_eddies.py` can read the
file `config_nml.txt` and merge its content to stdin for the Fortran
program.

Add argument SHPC-dir to script `inst_eddies_Aviso.py`. We want to be
able to run `inst_eddies_Aviso` in parallel for several slices. Since
we write `used_config_nml.txt` to the current directory, we need to
run the parallel instances from different directories. So the current
directory cannot always be the directory containing the SHPC to be
appended to.

Add argument `SHPC_dir` to function `loop_inst_eddies`. We want to be
able to run `loop_inst_eddies` in parallel for several slices. Since
we write `used_config_nml.txt` to the current directory, we need to
run the parallel instances from different directories. So the current
directory cannot always be the directory containing the SHPC to be
appended to.

We are trying to simplify the logic in program `inst_eddies`. Also,
the program `inst_eddies` should not write `used_config_nml.txt` to
the top level of the SHPC directory since several instances could run
in parallel for several slices. So we are leaving it as a task for a
script over `inst_eddies` to create `used_config_nml.txt` at the top
level of the SHPC directory.

We want to be able to run `inst_eddies` from a directory other than
the directory containing the SHPC to be appended to. Preparing to
write `grid_nml.txt` and `used_config_nml.txt` in the current
directory. To be able to do this in parallel for several slices, we
need to be running in different directories.

In program `examine_eddy`.

Change order of command line arguments of program
`examine_eddy`. Preparing to allow a variable number of NetCDF file
arguments.

Do not always specify uv-file when it is the same as ssh-file.

Instead of reading command line arguments in procedure `input_ssh`,
read them outside and pass file names to `input_ssh`. This allows
`input_ssh` to be called by programs with different number of command
line arguments, in different positions. Preparing to make the
specification of uv-file optional.

Replace argument `s` of `set_all_extr` by its components. The idea is
to remove the structure `snapshot` from program `inst_edddies`.

We are circumventing a bug of the Intel Fortran compiler. This was
tested with both ifx `2023.1.0` and ifort `2021.9.0`. The bug occurs
on return from procedure `set_max_speed`: `free(): double free
detected in tcache 2` and crashes execution. The bug is revealed in
commit cc13fd5b. The GNU compiler does not have this bug.

To circumvent the Intel compiler bug, instead of an ssh component of
`cont_list_proj`, we have to use a separate array:
`cont_list_ssh`. This is maybe for the best because the component
`area` of `cont_list_proj` was not defined.

`MAX_RADIUS_deg` is not general enough for a non longitude latitude
grid. The variable `max_radius` moves from module `input_ssh_m` to
module `config_m`. The variable `MAX_RADIUS_deg` disappears. The
constraint `2 * max_radius_deg(1) < 180` becomes `4 * max_radius(1) <
nlon` but is only applied to the periodic case because we do not have
the general equivalent of 180° in number of points.

Promote nlon and nlat to variables of module `input_ssh_m`, avoiding
their repeated computation.

`Region_1_20454_cyclone_3/config_nml.txt` is a symlink to
`Extraction_eddies_region_1_noise/config_nml.txt`.

In main program units `inst_eddies`, `examine_eddy` and
`test_get_1_outerm`, redefine `outside_points` as coordinates in
projection space, avoiding a round trip to longitude and latitude. In
`examine_eddy`, we therefore store `coord_proj` instead of coord.

From `nearby_extr`.

Instead of `outside_points`.

Instead of `outside_points`.

Avoiding conversion back and forth to real.