[MITgcm-devel] [MITgcm-cvs] MITgcm/verification/lab_sea/run CVS Commit

[Menemenlis, Dimitris (3248)]

Done.

Patrick, do you want me to implement JM's suggestion,
i.e., add ".txt" at the end of  "costfunction0000" and
"costfunction_seaice0000" or leave as is?
I do not want to unwittingly impact downstream scripts
that may not expect the ".txt".

D.

On Feb 23, 2011, at 3:42 PM, Jean-Michel Campin wrote:

> Dimitris,
> 
>> Do you want me to undo this change?
> I would prefer. The "better" way to fix this would be
> to change the cost S/R to produce output file with a "well defined"
> suffix or prefix (e.g. *.txt) so that this would be easy
> to list in .cvsignore
> 
> Cheers,
> Jean-Michel
> 
> On Wed, Feb 23, 2011 at 02:22:09PM -0800, Menemenlis, Dimitris (3248) wrote:
>> Jean-Michel, you are correct, the check-in description has nothing to
>> do with the .cvsignore changes.  I tried to sneak in this change without
>> you noticing it, buried within other changes ... but obviously I failed :-(
>> These files are created by the adjoint verification tests.
>> I only run the seaice-related adjoint tests in lab_sea and and 
>> offline_exf_seaice.  I assume that similar files would also be
>> created in other adjoint tests but I don't know for sure.
>> 
>> Do you want me to undo this change?
>> 
>> D.
>> 
>> On Feb 23, 2011, at 2:11 PM, Jean-Michel Campin wrote:
>> 
>>> Hi Dimitris,
>>> 
>>> I am not in favor of your 2 ".cvsignore" changes in:
>>>> verification/lab_sea/run
>>> and 
>>>> verification/offline_exf_seaice/run
>>> because we try to use the same ".cvsignore" accross all "run" dir
>>> so that it's simpler to maintain. 
>>> Now I can see why you want to add those files in the list, 
>>> but what about other run dir which are also used for AD test ?
>>> 
>>> And just for confirmation: the check-in description message has nothing
>>> to do with the changes, right ?
>>> 
>>> Cheers,
>>> Jean-Michel