[MITgcm-devel] The bad state of "cvs -q update"

[Alistair Adcroft]

Ed,

I'm currently having trouble finding changes I want to check-in on a fully
compiled/tested set-up because the of all the question marks and "C"'s I get
for FC_NAMEMANGLE.h, SIZE.h and CPP_EEOPTIONS.h etc... This happened because
I use testreport.

1) I'd like FC_NAMEMANGLE.h to be created by Makefile and when it is, then
add it to .cvsignore so we don't swamp the output from cvs -q update
2) testreport is currently changing SIZE.h and CPP_EEOPTIONS.h and this
means that everything is being recompiled and is confusing the CVS output.
This REALLY needs to be sorted out.
    i) we uniformly move ALLOW/USE_MPI from CPP_EEOPTIONS.h into the
Makefile and gm_local
       using cmd-line. Most CPP_EEOPTIONS.h can then be moved out of
../code/ into eesupp/inc/
   ii) for SIZE.h, I'm not sure of the best solution but can think of some
dangerous ones
       using CPP and some safer ones using scripts in tools/

I will do it if you can't do it today.

A.
--
Dr Alistair Adcroft            http://www.mit.edu/~adcroft
MIT Climate Modeling Initiative        tel: (617) 253-5938
EAPS 54-1523,  77 Massachusetts Ave,  Cambridge,  MA,  USA