[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: netcdf 3.6.1-beta2 for OpenVMS

>To: address@hidden
>From: Russ Rew
>Subject: Re: 20050119: netcdf 3.6.1-beta2 for OpenVMS
>Organization: Technische Universiteit Delft, Kavli Institute of Nanoscience
>Keywords: OpenVMS, supported platforms

Hi Jouk,

> I created patches to get netcdf 3.6.1-beta2 workin on OpenVMS. You can find
> them on my web-page:
>   http://nchrem.tnw.tudelft.nl/openvms/software2.html#NETCDF
> 
> Can they be included in the final version of netcdf 3.6.1?

First, thanks very much for making the changes for OpenVMS available.
It's apparent that you carefully designed the changes to be minimal
and to not affect any non-OpenVMS platforms.  It was a difficult
decision whether to include the changes in netCDF-3.6.1, but we
ultimately decided to make your changes available with a link from our
web site instead.

The problems with including your changes in the source we distribute
include:

 - We can't test the changes here, so we would depend on your updates
   to descrip.mms files if we change the Makefiles.

 - We're aggressively moving toward use of autoconf and automake, and
   the philosophy of autoconf is for #ifdefs in the code to refer to
   features rather than platforms, and to use standard means for
   autoconf to detect features.

 - A static libsrc/ncconfig.h may not capture any differences among
   OpenVMS systems, so we're not sure that distributing a single
   version of ncconfig.h_vms is the right approach.  (This may reflect
   our ignorance of OpenVMS).

 - Putting the changes in the source we distribute, rather than
   offering an external patch, implies our support of the source, and
   we don't feel competent to answer questions about the descrip.mms
   files or the modified open statements, for example.

 - The OpenVMS FAQ website indicates support for Make and a POSIX
   shell that might make it possible for configure to work for OpenVMS
   in the future, although if this were the case, we would still have
   to change the test code to handle features such as
   "NO_MULTIPLE_OPENS".

I realize we have violated some of the principles advocated above with
previous support for Cray platforms, but in the future we would like
to also support Cray platforms with externally supplied patches,
similar to what we're proposing for OpenVMS.

I hope you'll understand that with very limited resources, we have to
be careful to limit taking on new support commitments.  Thanks again
for your expert efforts on behalf of the netCDF community.

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden          http://www.unidata.ucar.edu/staff/russ