Philippe 26/06/2024: IO_Field_read: 4D, 5D and 6D fields: ensure that data can not be overwritten in case of read error

Philippe 26/06/2024: IO_Field_read: 3D fields: ensure that data can not be overwritten in case of read error + other improvements

Philippe 21/06/2024: IO_Field_read: 2D fields: ensure that data can not be overwritten in case of read error

Philippe 21/06/2024: IO_Field_read: 1D fields: ensure that data can not be overwritten in case of read error

Philippe 21/06/2024: IO_Field_read: 0D fields: ensure that data can not be overwritten in case of read error

Philippe 19/06/2024: outputs: add LOUT_FILESPLIT_DISABLE option in namelist to force disabling Z-split files

Erwan Jezequel authored 10 months ago and RODIER Quentin committed 9 months ago

- /src/MNH/mean_field.f90
- /src/MNH/write_lfin.f90

Quentin 17/06/2024: bugfix ibm_detect, init of Z_NORM_GHOST and Z_NORM_TEMPO before their allocation (this is OK with ifort on Belenos git add ibm_detect.f90 ); crash on gfortran 9.4

Philippe 30/05/2024: move IO_Bakout_struct_prepare from INI_MODEL_n to INIT_MNH after INI_MODEL_n calls to ALL models

Philippe 30/05/2024: IO_Coordvar_write_nc4: fix: take the correct dimensions if the file model number does not correspond to the current model number

Philippe 29/05/2024: IO_Fieldlist_1field_write: manage model number internally (remove corresponding dummy argument)

Philippe 29/05/2024: add more custom constructors for tfielddata: possible to provide metadata + data array or scalar

Philippe 23/05/2024: outputs: do not write box metadata in Z-split files (box data is not written in them)

Philippe 23/05/2024: bugfix: call_rttov13: vector constructor was using character strings with different lengths

Philippe 22/05/2024: IO_Field_header_split_write_nc4: add optional arguments to force size of horizontal dimensions

Jean-Luc R. 07/05/2024: minor update of atm-ocean LCOUPLES option (reading forcings, init and variables names)

Quentin 07/05/2024: bugfix for use of CBL_HEIGHT_DEF=FRI: bl_depth_diag can be called with arrays in the full vertical dimensions
or the les vertical dimensions. The 1D version was also wrong. The FRI option was missing in the user guide (option forgotten since many years)