[MITgcm-support] MITgcm on Cygwin and MacOS X

[Chris Hill]

Hi Samar,

 Here are two main reasons (aka excuses) for the current separation of cpp

  1 - meshing well with what TAF/TAMC does - i.e. it wants to see pure
Fortran.
  2 - allowing a special filter that converts the pattern ' _d ' to either D
or E so that
      literal numbers (which are written 1. _d 0 for example in MITgcm) will
be 64-bit on all  
      platforms. This pattern substitution can't be done with cpp because it
requires removing 
      white space (the white space removal is needed to keep the TAF/TAMC
front-end parser happy).

 Does using an alternate suffix e.g. .for or .f77 cause problems or not
work? That was what I thought ought to work and would allow the use of
native DOS and HFS filesystems. Its fairly easy to macro the .f to .for,.f77
etc... in our toolchain so that it can be changed for different compilers if
needed.  

Chris  

> -----Original Message-----
> From: mitgcm-support-bounces at mitgcm.org 
> [mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of samar 
> khatiwala
> Sent: Friday, April 09, 2004 4:42 PM
> To: MITgcm-support
> Subject: Re: [MITgcm-support] MITgcm on Cygwin and MacOS X
> 
> Ed et al
> 
> Could you (or someone else) explain the rationale behind 
> explicitly calling cpp? All compilers I am aware of call cpp 
> before parsing the code. What is so special about compiling 
> this code that makes this necessary?
> 
> Most recently I have been running the MIT model on an a G5 
> XServe with the IBM xlf compiler. xlf (possibly even pgf) 
> gives you the option of saving the "small f" file it creates 
> after running cpp. Optionally, it will also preppend a "F" 
> before the file name so it does not matter whether you have 
> HFS or UFS.
> 
> There are good reasons to stick with HFS+. Unlike UFS it is 
> journaled and with some (minor) effort it too can be made 
> case sensitive.
> 
> Samar
> 
> Oceanography 201B
> Lamont Doherty Earth Observatory
> Columbia University
> Palisades, NY 10964
> voice (845)-365-8454
> Fax   (845)-365-8736
> 
> Department of Applied Physics and Applied Mathematics Mudd 
> 292C Columbia University voice (212)-854-8111
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
>