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

Re: 19981005: netcdf & POSIX

>To: address@hidden
>From: Paul Wessel <address@hidden>
>Subject: netcdf & POSIX
>Organization: .
>Keywords: 199810050942.DAA25864

Hi Paul,

>       We will soon release GMT 3.1 which uses netcdf 3.4.  GMT is already
> on your list of netcdf usages.  GMT 3.1 is now fully ANSI C and is POSIX
> compliant.
> 
> Question: Is netcdf 3.4 POSIX compliant?

It's our intention that it be POSIX-compliant, and we thought it was.
Anything that's not strict POSIX is a bug.

> I suspect no since when I compiled it under OpenNT (Interix) in WinNT (which
> is a strict POSIX environment) it compiled the libraries fine but failed in
> the ncgen directory since ncgentab.c tries to include values.h which is not
> in POSIX.  Any plans that the entire netcdf package will become
> strictly POSIX?

The ncgentab.c file is generated by yacc, and the #include <values.h>
statement yacc includes in the output doesn't come from our ncgen.y and
ncgen.l input.  This is a bug in the yacc we used.

It turns out that although values.h is included, there's no dependence on
it, so you should be able to just delete that line from ncgentab.c and
compile again.  Please let us know if this doesn't work.

> Great job on 3.4!

Thanks, and keep those bug reports coming.

--Russ

_____________________________________________________________________

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