sdfopen

[Jennifer Adams]

On Dec 16, 2009, at 5:58 PM, Dan Schwarzkopf wrote:

> what was the rationale for removing the template option in grads 2.0 ?
The I/O layer that used to be in sdfopen was removed because it was  
leaky and led to seg faults, particularly when using templating. Hoop  
has written a GrADS script that very nearly duplicates the behavior of  
the templating feature in sdfopen. Please see:
ftp://grads.iges.org/grads/scripts/sdfopent.gs

> the need to use control files instead of sdfopen for our  (GFDL)  
> timeseries files (which use
> a 365 day calendar) make the 2.0 version essentially useless.
You would likely find GrADS to be even more useless if you discovered  
it was opening a file and setting the initial time incorrectly. There  
was a problem with using the udunits library to determine the initial  
time for 365_day_calendar files (udunits doesn't support no_leap  
calendars). So, rather than have it be wrong, we ask the user to  
provide the accurate time axis info. This is a known problem for many  
of the netcdf files in the IPCC AR4 archive.

Jennifer

--
Jennifer M. Adams
IGES/COLA
4041 Powder Mill Road, Suite 302
Calverton, MD 20705
jma at cola.iges.org



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20091216/878387b2/attachment.html