[MITgcm-devel] check for pkg parameter files (data.ThiPKG)

[Martin Losch]

Hi Jean-Michel,

I agree that your case 3 is a common trap that people (including me) fall into frequently. 
I do not have any ideas better than yours of how to give the user a clue of why the package is not doing anything, but where would this (weak) warning go? Given the enormous amount of out of the model, I expect, that such a warning would be lost in the STDOUT. We could try to make it more visible by printing a summary of all compiled packages and their "used"-status in a table form that is visually different from the surround STDOUT. I don't know how to accomplish that, maybe by indenting such a table or put it at the end of the initialization phase (before the first monitor at myiter=nIter0)

Martin



On Aug 14, 2013, at 9:36 PM, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:

> Hi,
> 
> Generally, for most packages, in order to use "thisPKG", a user
> will have to remember 3 things: compile it (packages.conf), 
> switch on the logical "useThisPKG" in data.pkg, and provide a 
> parameter file "data.thisPKG".
> if the user just do 2 of the 3 steps but forgot the 3rd one:
> - compile + useThisPKG=T but no data.thisPKG           --> stop cleanly
> - useThisPKG=T (+ data.thisPKG exist) but not compiled --> stop cleanly
> - compile thisPKG + data.thisPKG exist but useThisPKG=F --> nothing
> In the 1rst 2 cases, there is a clear error message, but in the 3rd case
> there is no specific warning that thisPKG is not used.
> 
> I am going to try to add a (weak, in lower-case) warning for the case where
> pkg/diagnostics is compiled and data.diagnostics exist but useDiagnostics=F. 
> 
> I am not sure this is really useful; for now, it's just 1 pkg,
> so would not be too much pain to remove this (just 1 file).
> Any suggestion is welcomed.
> 
> Cheers,
> Jean-Michel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel