[Mitgcm-support] RE: I try to follow the procedure

[mitgcm-support at dev.mitgcm.org]

Hi Chris,

> Why do we need check_config in release1?

let's take an example:

In release1_beta1 code, with this set up:
in data:
exeacConserv=.TRUE.
in CPP_OPTIONS.h
#undef EXACT_CONSERV

the model runs, gives wrong results (that do not correspond to 
exeacConserv=FALSE & undef neither to exeacConserv=TRUE & defined)
and gives no warning.

It seems important that for the release we avoid this kind of
dangerous bad use of parameter/option.
And regarding other checks that have been recently added (e.g. 
retired parameters), this config_check appears necessary.

Jean-Michel