[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: 960807: C++ netCDF interface
- Subject: Re: 960807: C++ netCDF interface
- Date: Thu, 08 Aug 1996 08:53:06 -0600
>To: address@hidden
>From: Dave Wood <address@hidden>
>Subject: C++ netCDF interface
>Organization: Colorado State University
>Keywords: 199608072100.AA08938
Hi Dave,
> I am trying to write some software using the Win32(MD) library with
> MFC and MSVC 4.0. I must have an incomplete set of header files
> because the build aborts due to excessive errors. I include
> "netcdf.hh" which in turn includes "netcdf.h" and "ncvalues.hh".
> Ncvalues.hh includes several more files including "generic.h". My
> version of generic.h (which I found under the ncgen directory of the
> netCDF 2.4.2 release) contains only a definition of the generic union.
The generic.h in the ncgen directory has nothing to do with the
generic.h file that ncvalues.hh wants; it's just an unfortunate name
collision from different name spaces. I should rename the generic.h
file in ncgen to avoid this confusion.
The generic.h referred to in ncvalues.hh was a header file that was
included with all the C++ compilation systems I knew about, so I thought
we could assume the C++ compiler would find it in a standard place, but
I guess that assumption is wrong.
I've appended one version of the file (that comes with the GNU g++
compiler) so you can see what it does. Our Sun installation has a
copyrighted version in
/opt/SUNWspro/SC4.0/include/CC/generic.h
and we find it in similar places for compilers from other vendors:
/usr/include/CC/generic.h # for HPUX or IRIX
> However, the ncvalues.hh header file makes the following note in a
> comment:
>
> // This is the same as the name2 macro from generic.h, but we need to define
> // our own version since rescanning something generated with the name2 macro
> // won't necessarily cause name2 to be expanded again.
> #define ...
>
> There is no definition of name2 in the generic.h file that I have. This may
> not be that
> important but _somewhere_ there must be a definition of:
>
> declare(NcValues,ncbyte)
> declare(NcValues,char)
> declare(NcValues,short)
> declare(NcValues,nclong)
> declare(NcValues,float)
> declare(NcValues,double)
>
> which is found at the end of ncvalues.hh. Is this also some sort of
> stringizing or token-pasting defined in the lost generic.h? What am I
> missing?
> - --
Yup, it's for token pasting.
--Russ
_____________________________________________________________________
Russ Rew UCAR Unidata Program
address@hidden http://www.unidata.ucar.edu
// This may look like C code, but it is really -*- C++ -*-
/*
Copyright (C) 1988 Free Software Foundation
written by Doug Lea (address@hidden)
This file is part of the GNU C++ Library. This library is free
software; you can redistribute it and/or modify it under the terms of
the GNU Library General Public License as published by the Free
Software Foundation; either version 2 of the License, or (at your
option) any later version. This library is distributed in the hope
that it will be useful, but WITHOUT ANY WARRANTY; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE. See the GNU Library General Public License for more details.
You should have received a copy of the GNU Library General Public
License along with this library; if not, write to the Free Software
Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#ifndef generic_h
#define generic_h 1
/*
* See the CPP manual, argument prescan section for explanation
*/
#define name2(a,b) gEnErIc2(a,b)
#define gEnErIc2(a,b) a ## b
#define name3(a,b,c) gEnErIc3(a,b,c)
#define gEnErIc3(a,b,c) a ## b ## c
#define name4(a,b,c,d) gEnErIc4(a,b,c,d)
#define gEnErIc4(a,b,c,d) a ## b ## c ## d
#define GENERIC_STRING(a) gEnErIcStRiNg(a)
#define gEnErIcStRiNg(a) #a
#define declare(clas,t) name2(clas,declare)(t)
#define declare2(clas,t1,t2) name2(clas,declare2)(t1,t2)
#define implement(clas,t) name2(clas,implement)(t)
#define implement2(clas,t1,t2) name2(clas,implement2)(t1,t2)
//extern genericerror(int,char*);
typedef int (*GPT)(int,char*);
#define set_handler(gen,type,x) name4(set_,type,gen,_handler)(x)
#define errorhandler(gen,type) name3(type,gen,handler)
#define callerror(gen,type,a,b) (*errorhandler(gen,type))(a,b)
#endif /* generic_h */