[MITgcm-devel] fortran namelist syntax & gfortran

[Martin Losch]

Jean-Michel,

the error message is the same that I got earlier with the f95 syntax 
namelists. it's the line
>    fields(1:3,1) = 'ETAN    ','ETANSQ  ','DETADT2 ',
in aiml.5l_cs/input.thSI/data.diagnostics, which has not been changed 
back since Nov 25. I don't think that it's a problem with testreport.

M.

On Dec 1, 2005, at 3:39 PM, Jean-Michel Campin wrote:

> Hello Ed,
>
> I have 2 remarks:
>
> 1) with a fresh version of the code, on faulks, I did:
> testreport -of ../tools/build_options/linux_ia32_pgf77+authors_fc3 -t 
> aim.5l_cs
>
> and get a nice double pass (in tr_out.txt) for both aim.5l_cs & 
> aim.5l_cs.thSI
> but the 2nd test (aim.5l_cs.thSI) is in fact not doing well:
>
> aim.5l_cs/tr_run.thSI/run.log :
>> running ... PGFIO-F-248/namelist read/unit=11/illegal subscript or 
>> substring specification.
>>  File name = /tmp/FTNkaaaaflkaB    formatted, sequential access   
>> record = 4
>>  In source file diagnostics_readparms.f, at line number 1794
>> make: *** [output.txt] Error 1
>
> and aim.5l_cs/tr_run.thSI/output.txt ends shortly with:
> (PID.TID 0000.0001) S/R DIAGNOSTICS_READPARMS, read namelist 
> "diagnostics_list": start
>
> My impression is that we have a problem in testreport with
> multiple tests.
>
> 2) You changed back some of the namelist but not all of them.
> (I attached here the result of the command:
> egrep '1:' */input*/data* | grep -v 'Header' )
> I would prefer to have only 1 type of namelist (as it was before
> Nov 24), instead of a mix of different syntax, some working with
> one compiler but not with an other, and some with just the reverse.
>> From my point of view, this is just a pain to maintain.
>
> Thanks,
>
> Jean-Michel
> <toto>_______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel