Re: Math/udunits update

[Brook Milligan <brook%nmsu.edu@localhost>]

> On Mar 6, 2020, at 2:24 PM, Dr. Thomas Orgis <thomas.orgis%uni-hamburg.de@localhost> wrote:
> 
> Am Fri, 6 Mar 2020 14:06:10 -0700
> schrieb Brook Milligan <brook%nmsu.edu@localhost>:
> 
>> I think the safe thing is likely to make a new udunits2 package and leave the other 
> 
> 1. Udunits and udunits2 are two distinct packages for me. I see them
> referred to as such. So ACK here.
> 
> 2. We have wip/udunits2 already, which I do use. I did not see it
> mentioned in the thread, did I miss it?
> 
> I do see that they conflict, though. Both provide include/udunits.h .
> 
> So far I needed udunits2 as a dependency for GIS software. I did not
> need udunits.

From what I understand, the current version of udunits2 installs udunits1 headers and includes relevant stuff in libudunits2.  Thus, it seems that other than the Fortran API, which is not in udunits2, and a disagreement about the interpretation of “g”, everything is in udunits2.

Are you finding this not to be the case?  Do we really need two versions?  Are there examples of packages that cannot use the v.1 stuff in udunits2?

As a check, both of the existing packages that depend on udunits (graphics/ncview and math/R-RNetCDF) can compile against udunits2 (the former requires a minor patch to switch from libudunits to libudunits2).

Cheers,
Brook

--
Brook Milligan
Department of Biology
New Mexico State University
Las Cruces, New Mexico 88003
+1 575-646-7980
brook%nmsu.edu@localhost