Re: 20030903:NCL and NC_MAX_VARS

>To: address@hidden (Russ Rew)
>From: Mary Haley <address@hidden>
>Subject: Re: NCL and NC_MAX_VARS
>Organization: UCAR/SCD
>Keywords: 200309031508.h83F8SLd024439

Mary,

> It just occurred to me that the version of NCL this guy is using
> was linked against 3.5.1-beta, and I see that the limit is bumped
> up to 4096, so I think something else is going on. I'm trying to get
> his netCDF file so I can investigate further.

There have been several 3.5.1-betaxx releases, and I neglected to
determine for which "xx" the "betaxx" had the limit bumped up to 4096.
The change was made on Oct 13, 2001, so it should be in the
3.5.1-beta5, or 3.5.1-beta10 or later releases, but not in anything
before beta4 or in the beta8 release which turned out to be a mistake
that reverted to an earlier version (don't ask!) ...

> Meanwhile, on a related note, I saw your note about beta12 being out a
> few weeks ago, and had actually started building and testing it awhile
> back.  I have a question: when I run "make test", how do I know that
> all the tests passed?  I see things like "Number of failures: 0" and
> :"***Success***".  Is it enough to look for these to assure myself
> that the tests ran successfully?

Yes, that's all you need to look for.  Unfortunately, we've never come
up with a way to automatically indicate that all the tests passed.  If
there is a failure, the "make test" should stop in the directory in
which the failure occurred, so it won't complete.

--Russ