[MITgcm-devel] changes in testreport

[Jean-Michel Campin]

Hi Constantinos,

It seems that my modifications on testreport do not work very well
on td194 (hp-ux_ia64).
An other thing is that we did not have any test with hp-ux_parisc
on td192 since Thursday (when I did the changes), and wandering
if there is a problem also with those ones.
Can we look at this on Monday (or if I can get an acces to 
this machine) ?
Thanks,
Jean-Michel

On Thu, Sep 06, 2007 at 09:12:53PM -0400, Jean-Michel Campin wrote:
> Hi,
> 
> I changed a little bit testreport, and might not be 100%
> ok on all platforms (I am sure Martin will find something).
> 
> The idea is (cvs check-in msg):
> > Allow to choose (per experiment) the list of tested output variables
> >    and the main variable (used to issue a pass/FAIL ) ;
> > - for now, done by adding the specific list in file: [exp]/input/tr_checklist ;
> > - default is unchanged: >cg2d< , (T,S,U,V,5 tracers)_[min,max,mean,StD] ;
> > Also modify printing of summary:
> > a) ">" NN "<" indicates the "main" variable (used for pass/FAIL)
> > b) echo the list of tested variables in the individual summary.txt of
> >  each experiment.
> 
> I put an example of tr_checklist in /advect_cs/input
> (same default list of variables, but the "main" one is changed
> to Tsd (for Theta Std.Dev) instead of PS (for Pressure Solver = cg2d)
> 
> I submitted a standard test on faulks and it's now on the webb page
> (faulks :        not_explicitly_specified )
> so that every one can see this "printing improvement".
> If you have some comments, welcome !
> 
> Unfortunately, it has a problem with the score: "1:1"
> (something in the testing web-page building) and would need
> to be change to the Number of tests that "pass".
> I disabled for now the print changes to get normal score for
> tonight testing.
> 
> Did not touch to the AD testing part, but there is a good chance
> (with little modification to the code & dir structure)
> to re-use, for the AD test, the same functions used for the forward
> testing (Patrick, we can talk about this when you are back).
> 
> Jean-Michel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel