[MITgcm-devel] Re: [MITgcm-cvs] MITgcm/pkg/mnc CVS Commit

[Jean-Michel Campin]

Hi Martin,

Oliver started to put some f90 feature, and we tried to 
define some logic rule to get some compatibility and 
testing to work. Oliver, did you put something in the "How_to"
file ?
Regarding this g77 problem, it's likely
that we will switch to gfortran to be the "reference"
compiler for producing our output.txt. But it will take 
some time (also depends on how the upgrade of faulks and others
go). In the mean time, I think that 
tutorial_advection_in_gyre/results/output.txt
was generated with gfortran.

Jean-Michel

On Mon, Jun 23, 2008 at 09:03:08AM +0200, Martin Losch wrote:
> Hi there,
> can I ask this question again?
>
>
> On 23 May 2008, at 14:02, Martin Losch wrote:
>> however I noticed that on eddy.csail.mit.edu  
>> "tutorial_advection_in_gyre" does not compile because of f90-code in 
>> the ptracers-som code that does not compile with g77 (module, pointer, 
>> allocate, etc). That's probably well know to you, but just for my 
>> interest, what is the policy on f90-code in the future?
>>
>> Martin
>> On 23 May 2008, at 08:54, Martin Losch wrote:
>>
>>> Hi Jean-Michel,
>>>
>>> thanks for putting me back on track, for some reason I thought that 
>>> mnc is so much self-contained that I only ran testreport for a few 
>>> experiments that use mnc.
>>> sorry, I'll fix that right away (full testreport running now)
>>> I found (grep mnc_common.h pkg/*/*.[h,F]) occurrences of  
>>> mnc_common.h in
>>> fizhi_mnc_init.F
>>> regrid_mnc_init.F
>>> land_mnc_init.F
>>> can you think of anything else?
>>>
>>> (at least we learned something about makedepend ... )
>>>
>>> Martin
>>> On 23 May 2008, at 02:39, Jean-Michel Campin wrote:
>>>
>>>> Hi Martin,
>>>>
>>>> When you do those kind of changes (renaming files ...),
>>>> can you (at least) run testreport to check that it does
>>>> not break too many experiments ?
>>>>  aim.5l_cs (+ .thSI), fizhi-cs-aqualev20 : fail to compile
>>>>
>>>> And in addition, there are some pkgs that are never tested,
>>>> but with those changes, it's unlikely that they will work
>>>> better now:
>>>>> pkg/regrid/regrid_mnc_init.F:#include "mnc_common.h"
>>>>
>>>> Jean-Michel
>>>>
>>>> On Thu, May 22, 2008 at 08:21:20AM -0400, Martin Losch wrote:
>>>>> Update of /u/gcmpack/MITgcm/pkg/mnc
>>>>> In directory forge:/tmp/cvs-serv31025/pkg/mnc
>>>>>
>>>>> Modified Files:
>>>>> 	mnc_var.F mnc_utils.F mnc_readparms.F mnc_init.F mnc_grid.F
>>>>> 	mnc_file.F mnc_dump.F mnc_dim.F mnc_cwrapper.F
>>>>> 	mnc_cw_write_grid_info.F mnc_cw_udim.F
>>>>> 	mnc_cw_readwrite.template mnc_cw_missingvals.F mnc_cw_init.F
>>>>> 	mnc_cw_cvars.F mnc_cw_citer.F Makefile MNC_OPTIONS.h
>>>>> 	.cvsignore
>>>>> Added Files:
>>>>> 	MNC_COMMON.h
>>>>> Removed Files:
>>>>> 	mnc_common.h
>>>>> Log Message:
>>>>> replace mnc_common.h and mnc_id_header.h with corresponding  
>>>>> upper case
>>>>> versions
>>>>>
>>>>> _______________________________________________
>>>>> MITgcm-cvs mailing list
>>>>> MITgcm-cvs at mitgcm.org
>>>>> http://mitgcm.org/mailman/listinfo/mitgcm-cvs
>>>> _______________________________________________
>>>> MITgcm-devel mailing list
>>>> MITgcm-devel at mitgcm.org
>>>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>>>
>>> _______________________________________________
>>> MITgcm-devel mailing list
>>> MITgcm-devel at mitgcm.org
>>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>>
>> _______________________________________________
>> MITgcm-devel mailing list
>> MITgcm-devel at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel