Various_stepons/stepon_{eul,homme,sld} have not yet  been modified for DART.

----------------
stepon_fv.F90:
Where is functionality of inicFile_Mod.F90 now?
   subroutines which I modified for ENDOFRUN, are not in this module anymore.
 x inicfile had change of last argument of putfil changed from .true. to .false.
       It looks like this function is now executed in main/histFileMod.F90.
       The logical flag is set in function do_disp, which sets it to .false.
       for the end of run instance.  So I think that I don't need this
       change anymore.   I'll see when the clminput.nc file does or doesn't
       appear when needed.
   inicfields
 x do_inicwrite() (generic function) had is_last_step added to 'use time_manager' list
       and 'if' test for ENDOFRUN added in.
       do_inicwrite() is now in lnd/clm2/src/main/driver.F90, which I've added to _mods

   So I'll try removing inicFile_Mod.F90 from this directory; it appears that I don't
      need a modified version of the new one.


stepon[_fv].f90
   I modified 'use time_manager' to include line
     nelapse, nestep, stop_ymd, stop_tod
     but none of these are used in 3.1 stepon.F90


10/9/07
Several subroutines were added, hopefully temporarily, to deal with CAM 3.5,
which hadn't been tested for using the GREGORIAN calendar (instead of NO_LEAP).

clm_time_manager.F90; There was confusion between the gregorian flag CLM wanted
to see (ESMF_CAL_GREGORIAN) and the one coming from CAM (GREGORIAN).  I made
it look for GREGORIAN.

ESMF.inc needs to have the contents commented out in order for gregorian to be used.
See directory Gregorian.

There's also a problem with the ocean model interacting with the gregorian
calendar in leap years.  I put in debug statement in Gregorian/ocn_time_manager.F90.
See e-mails from 10/2007.
I decided to do the 20 year spin-up run with NO_LEAP and no ocean cycling,
then hopefully CCSM will fix this before I need to assimilate in (or just after)
a leap year.

3/26/08
leapfix; added lines to 
   clm_time_manager.F90  
   ice_time_manager.F90
   ocn_time_manager.F90  
   time_manager.F90
to work around leap year problem.
Also brought back in ESMF.inc into directory to specify GREG instead of NOLEAP.
(I believe I compiled with NOLEAP the last time(s) I built CAM, for version 28,
and IC generation).
==============================
6/5/08
csm_share/unit_testers/make.Macros:  CPPDEF    += -DSEQ_ESMF

lnd/clm2/src/main/controlMod.F90:#if (defined SEQ_MCT) || (defined SEQ_ESMF)
lnd/clm2/src/main/controlMod.F90:#elif (defined SEQ_MCT) || (defined SEQ_ESMF)
lnd/clm2/src/main/do_close_dispose.F90:#if (defined SEQ_MCT) || (defined SEQ_ESMF)
lnd/clm2/src/main/driver.F90:#if (defined SEQ_MCT) || (defined SEQ_ESMF)
lnd/clm2/src/main/restFileMod.F90:#elif (defined SEQ_MCT) || (defined SEQ_ESMF)
lnd/clm2/src/main/restFileMod.F90:         call endrun('restFile_archive error: must pass nlend as argument for SEQ_MCT or SEQ_ESMF')

utils/esmf_wrf_timemgr/ESMF_CalendarMod.F90:  ESMF_CAL_GREGORIAN =  ESMF_CalendarType(1), &


atm/cam/src/utils/time_manager.F90:      cal_type = ESMF_CAL_GREGORIAN
lnd/clm2/src/main/clm_time_manager.F90:!        else if ( trim(cal) == 'ESMF_CAL_GREGORIAN' ) then
lnd/clm2/src/main/clm_time_manager.F90:           calendar = 'ESMF_CAL_GREGORIAN'
lnd/clm2/src/main/clm_time_manager.F90:     cal_type = ESMF_CAL_GREGORIAN

====================================
6/26/08

CAM 3.5.06 won't run correctly on bluefire under default(11.1) xlf90_r compiler;
seg-faults at various places depending on optimization.
Work-arounds were committed to CAM 3.5.45; 
   one in models/ocn/dom/ocn_comp.F90 seems irrelevant for v06.
<        frac(c)%land(:ncols) = arr2d(:ncols,c)
---
>        frac(c)%land(:ncols) = 1._r8 - arr2d(:ncols,c)
   2 in atm/cam/src/physics/cam/cam_diagnostics.F90
955a931,935
>       !following line added temporarily in cam3_5_45 as workaround to 
>       !bluevista compiler problems
>       nullify(cwat_var, conv_var_3d, conv_var_2d)
962a943,949
>       !following if block added temporarily in cam3_5_45 as workaround to 
>       !bluevista compiler problems
>       if(.not. associated(cwat_var)) then
>          call endrun('Could not associate to pbuf fld_ptr')
>       end if


