lack of interactivity in compiling latest version of the GCM

[aslmd]

in revision 2328 things were fine

but when moving to most recent (up-to-date) the GCM is no longer able to compile without asking the question below, which prevent compiling in scripts with redirection to be successful

ATTENTION: vous etes sans doute en train de compiler le modele par ailleurs
Attendez que la premiere compilation soit terminee pour relancer la suivante.

Si vous etes sur que vous ne compilez pas le modele par ailleurs,
vous pouvez continuer en repondant oui.
Voulez-vous vraiment continuer?

WARNING: you are probably already compiling the model somewhere else.
Wait until the first compilation is finished before launching this one.
If you are sure that you are not compiling elsewhere, just answer
yes (or 'oui') to the question below to proceed.
Do you wish to continue?

[aslmd]

When answering "yes", it says anyway:

./makelmdz_fcm: line 557: cd: /data/spiga/2020_MESOSCALE_sandbox/code/LMDZ.COMMON/libf/phymars/scatterers: Aucun fichier ou dossier de ce type

Then quite immediate compilation failure

[aslmd]

I was trying on ciclad using the following command in the LMDZ.COMMON folder

./makelmdz_fcm -cpp MESOINI -j 8 -s 2 -d 64x48x29 -arch CICLADifort -parallel mpi -p mars gcm

[aslmd]

After a closer investigation by commit dichotomy exploration, the problem starts with revision r2339. GCM does not compile on ciclad with that version (same error as reported above), while it does compile beautifully for revision r2338

[emillour]

makelmdz_fcm stops because of "-e" option introduced in r2339 (makes script stop as soon as an error is encountered.
The error in this case is that the "-s 2" option triggers buiding "scatterers.h" via a dedicated script. However the Mars physics don't require this scatterers.h since r1246.

Removing the "-s 2" from the arguments will solve the problem. Number of scatterers should be set in callphys.def (naerkind = 2).
It would be advisable to add checks in makelmdz_fcm to yield a more relevant error message if users wrongly use option "-s ..."

[aslmd]

Thanks Ehouarn! You just have added a check in r2441
I will modify soon my script meso_install.sh to remove the use of "-s 2" in compiling command
So I am closing this!