[MITgcm-devel] switch from CVS to GitHub, next Tuesday

[Martin Losch]

Hi Jean-Michel,
I am trying to understand your test_baudelaire script.
Is it really necessary to manage the git cloning/updating in such complicated way?
according to your script I need to have two copies of the MITgcm on my disk for the tests, right?
(1) the clone from the repository (MITgcm in your script)
(2) a copy of MITgcm (MITgcm_today) which you remove each time for checkOut (along with all testresults)

I would like to do this:
have a copy for ifort (MITgcm_ifort) and one for the cray compiler (MITgcm_cray), and I would like to update thoes copies before running testresport (but never delete them). That way I can also keep all of my testresport results (tr_-directories) which comes in handy, when I want to back to an old problem.
How do I do that properly with git? Can I rename a clone (i.e. my Clone MITgcm into MITgcm_ifort and then cd MITgcm_ifort; git checkout master or git pull)?

Martin

> On 7. Feb 2018, at 15:50, Jean-Michel Campin <jmc at mit.edu> wrote:
> 
> Hi Martin,
> 
> You could start, but it might take some time for PR to get merge
> since the priority here is to get back the doc rst files (+Ocean Sciences
> next week). And we will ask you at some point (but not now) to resubmit your
> changes in pkg/seaice documentation.
> 
> Also, right now, the code is not tested on any cluster, only on 
> baudelaire.mit.edu  and villon.mit.edu
> Working to get it tested on engaging and svante.
> I think the testing on these 2 clusters will be useful example since:
> - on svante: no ssh to outside world and no https on compute nodes,
>   so code update is on head-node.
> - on engaging: both ssh to outside and git https are available from
>   compute node, so everything can be done from inside batch job.
> Now remember that some firewall might not allow: 
>  "git clone https://github.com/MITgcm/MITgcm"
> but might let you do: "git clone git at github.com:MITgcm/MITgcm"
> 
> Cheers,
> Jean-Michel
> 
> On Wed, Feb 07, 2018 at 09:39:13AM +0100, Martin Losch wrote:
>> Hi Jean-Michel, Oliver, etal,
>> great job! Can we start forking/cloning/branching/gitting now?
>> 
>> Martin
>> 
>>> On 6. Feb 2018, at 23:24, Jean-Michel Campin <jmc at mit.edu> wrote:
>>> 
>>> Hi,
>>> 
>>> Oliver finished the conversion of MITgcm code to GitHub (last Friday) and
>>> MITgcm is now available on GitHub. Some updated informations are there:
>>> http://mitgcm.org/public/source_code.html
>>> 
>>> However, the documentation that has been converted from LaTex to Sphinx/ReadTheDocs
>>> (on the prototype git repo: altMITgcm) has not been added yet but will be soon.
>>> 
>>> Cheers,
>>> Jean-Michel
>>> 
>>> On Fri, Jan 26, 2018 at 05:45:53PM -0500, Jean-Michel Campin wrote:
>>>> Hi,
>>>> 
>>>> The conversion from CVS to GitHub is on a good track, so we are planning 
>>>> to "freeze" the MITgcm part (gcmpack/MITgcm) of the CVS repository 
>>>> next Tuesday (Jan 30) morning and start the final conversion from CVS to git.
>>>> 
>>>> The "MITgcm_contrib" part is and will still accept commit after this date
>>>> whereas the "manual" CVS repo has been "frozen" last Tuesday (Jan 23).
>>>> 
>>>> It would make things easier (and more clean) not to make significant modifications
>>>> to main MITgcm code before Tuesday (unless it's a minor/small/limited change)
>>>> but to postpone it to later next week when the GitHub repository is in place.
>>>> 
>>>> Cheers,
>>>> Jean-Michel
>>>> _______________________________________________
>>>> MITgcm-devel mailing list
>>>> MITgcm-devel at mitgcm.org
>>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-devel
>>> _______________________________________________
>>> MITgcm-devel mailing list
>>> MITgcm-devel at mitgcm.org
>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-devel
>> 
>> _______________________________________________
>> MITgcm-devel mailing list
>> MITgcm-devel at mitgcm.org
>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-devel
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-devel