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

960315: Building NETCDF on VMS



Kevin,

>Date: Fri, 15 Mar 1996 14:32:29 -0500 (EST) 
>From: address@hidden
>Organization: Nova University
>Subject: RE: 960315: Building NETCDF on VMS 
>Keywords: 199603151320.AA15784

In the above message you wrote:

> Thanks for your reply (and yes, I'm the Kevin from RSMAS 
> stilling hanging
> around in sunny South Fl.)

Good to hear from you.

> Well, I transferred the file fills.nc to a Sun workstation and got the
> following:
>
> kevin-43> ncdump fills.nc
> netcdf fills {
> dimensions:
>         n = 2 ;
> 
> variables:
>         double d(n) ;
>         float f(n) ;
>         long l(n) ;
>         short s(n) ;
>         byte b(n) ;
> 
> data:
> 
>  d = 1, 9.969209968386869e+36 ;
> 
>  f = 1, 9.96921e+36 ;
> 
>  l = 1, -2147483647 ;
> 
>  s = 1, -32767 ;
> 
>  b = 01, 0201 ;
> }
>
> First off, is this correct?

The above values are correct.  (It appears that you used the
version 2.3 ncdump utility rather than the one in the distribution --
but this doesn't matter)

> Also,am I correct in my thinking that fills.nc is NETCDF-supplied,
> rather than being writing by one of the make files?

That's correct.

It appears that the nctest failure is due to some problem other than
corruption of the fills.nc file.  Unfortunately, I don't have access to
an Alpha OpenVMS system, so I can't track this problem down.

Would you be willing to investigate this problem?  Do you have a
debugger that you could use to examine the contents of the relevant
variables just prior to the tests that fail?

> Regarding the ncdump isprint problem, here's the result of the
> pre_processing:

> if (((decc$$gl___ctypea)?(*decc$$ga___ctypet)[((int)uc = 
> *gp.cp++ & 0377)]&0x80:decc$isprint(uc = *gp.cp++ & 
> 0377)))

Well, isprint() is definitely a macro and not a function (as the
standard requires).  Pity.

Try changing the line in ncdump.c from

    if (isprint(uc = *gp.cp++ & 0377))

to

    uc = *gp.cp++ & 0377;
    if (isprint(uc))


Please let me know if this helps.

--------
Steve Emmerson   <address@hidden>