xdrfile.h File Reference

Interface to read/write portabile binary files using XDR. More...

Namespaces
	PLMD
	PLMD::xdrfile

Macros
#define	_XDRFILE_H_
#define	DIM   3

Typedefs
typedef float	PLMD::xdrfile::matrix[DIM][DIM]
typedef int	PLMD::xdrfile::mybool
typedef float	PLMD::xdrfile::rvec[DIM]
typedef struct XDRFILE	PLMD::xdrfile::XDRFILE
	Abstract datatype for an portable binary file handle. More...

Enumerations
enum	{   PLMD::xdrfile::exdrOK, PLMD::xdrfile::exdrHEADER, PLMD::xdrfile::exdrSTRING, PLMD::xdrfile::exdrDOUBLE,   PLMD::xdrfile::exdrINT, PLMD::xdrfile::exdrFLOAT, PLMD::xdrfile::exdrUINT, PLMD::xdrfile::exdr3DX,   PLMD::xdrfile::exdrCLOSE, PLMD::xdrfile::exdrMAGIC, PLMD::xdrfile::exdrNOMEM, PLMD::xdrfile::exdrENDOFFILE,   PLMD::xdrfile::exdrFILENOTFOUND, PLMD::xdrfile::exdrNR }

Functions
int	PLMD::xdrfile::xdrfile_close (XDRFILE *xfp)
	Close a previously opened portable binary file, just like fclose() More...
int	PLMD::xdrfile::xdrfile_compress_coord_double (double *ptr, int ncoord, double precision, XDRFILE *xfp)
	Compress coordiates in a double array to XDR file. More...
int	PLMD::xdrfile::xdrfile_compress_coord_float (float *ptr, int ncoord, float precision, XDRFILE *xfp)
	Compress coordiates in a float array to XDR file. More...
int	PLMD::xdrfile::xdrfile_decompress_coord_double (double *ptr, int *ncoord, double *precision, XDRFILE *xfp)
	Decompress coordiates from XDR file to array of doubles. More...
int	PLMD::xdrfile::xdrfile_decompress_coord_float (float *ptr, int *ncoord, float *precision, XDRFILE *xfp)
	Decompress coordiates from XDR file to array of floats. More...
XDRFILE *	PLMD::xdrfile::xdrfile_open (const char *path, const char *mode)
	Open a portable binary file, just like fopen() More...
int	PLMD::xdrfile::xdrfile_read_char (char *ptr, int ndata, XDRFILE *xfp)
	Read one or more char type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_double (double *ptr, int ndata, XDRFILE *xfp)
	Read one or more double type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_float (float *ptr, int ndata, XDRFILE *xfp)
	Read one or more float type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_int (int *ptr, int ndata, XDRFILE *xfp)
	Read one or more integer type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_opaque (char *ptr, int nbytes, XDRFILE *xfp)
	Read raw bytes from file (unknown datatype) More...
int	PLMD::xdrfile::xdrfile_read_short (short *ptr, int ndata, XDRFILE *xfp)
	Read one or more short type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_string (char *ptr, int maxlen, XDRFILE *xfp)
	Read a string (array of characters) More...
int	PLMD::xdrfile::xdrfile_read_uchar (unsigned char *ptr, int ndata, XDRFILE *xfp)
	Read one or more unsigned char type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_uint (unsigned int *ptr, int ndata, XDRFILE *xfp)
	Read one or more unsigned integers type variable(s) More...
int	PLMD::xdrfile::xdrfile_read_ushort (unsigned short *ptr, int ndata, XDRFILE *xfp)
	Read one or more unsigned short type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_char (char *ptr, int ndata, XDRFILE *xfp)
	Write one or more characters type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_double (double *ptr, int ndata, XDRFILE *xfp)
	Write one or more double type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_float (float *ptr, int ndata, XDRFILE *xfp)
	Write one or more float type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_int (int *ptr, int ndata, XDRFILE *xfp)
	Write one or more integer type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_opaque (char *ptr, int nbytes, XDRFILE *xfp)
	Write raw bytes to file (unknown datatype) More...
int	PLMD::xdrfile::xdrfile_write_short (short *ptr, int ndata, XDRFILE *xfp)
	Write one or more short type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_string (char *ptr, XDRFILE *xfp)
	Write a string (array of characters) More...
int	PLMD::xdrfile::xdrfile_write_uchar (unsigned char *ptr, int ndata, XDRFILE *xfp)
	Write one or more unsigned characters type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_uint (unsigned int *ptr, int ndata, XDRFILE *xfp)
	Write one or more unsigned integer type variable(s) More...
int	PLMD::xdrfile::xdrfile_write_ushort (unsigned short *ptr, int ndata, XDRFILE *xfp)
	Write one or more unsigned short type variable(s) More...

Detailed Description

Interface to read/write portabile binary files using XDR.

This file provides an interface to read & write portably binary files, using XDR - the external data representation standard defined in RFC 1014.

There are several advantages to the XDR approach:

1. It is portable. And not just portable between big/small integer endian, but truly portable if you have system XDR routines. For example:
   • It doesn't matter if the character representation is ASCII or EBCDIC.
   • Some systems are small endian but use big endian order of the two dword in a double precision floating-point variable. The system XDR libraries will read/write this correctly.
   • Some systems (VAX...) don't use IEEE floating point. Their system XDR libraries will convert to/from this automatically.
2. XDR libraries are required for NFS and lots of other network functions. This means there isn't a single Unix-like system that doesn't have them.
3. There is NO extra metadata whatsoever, and we write plain XDR files. If you write a float, it will take exactly 4 bytes in the file. (All basic datatypes are 4 bytes, double fp 8 bytes).
4. You can read/write the files by calling the system XDR routines directly too - you don't have to use the routines defined in this file.
5. It is no problem if your system doesn't have XDR libraries (MS Windows). We have written our own versions of the necessary routines that work if your system uses ASCII for strings and IEEE floating-point. All types of byte and dword endian for integer and floating-point are supported.
6. You can use these routines for any type of data, but since we designed them for Gromacs we also provide a special routine to write coordinates with (adjustable) lossy compression. The default precision will give you three decimals guaranteed accuracy, and reduces the filesize to 1/10th of normal binary data.

We do not support getting or setting positions in XDR files, since it can break in horrible ways for large (64-bit) files, resulting in silent data corruption. Note that it works great to open/read/write 64-bit files if your system supports it; it is just the random access we cannot trust!

We also provide wrapper routines so this module can be used from FORTRAN - see the file xdrfile_fortran.txt in the Gromacs distribution for documentation on the FORTRAN interface!

Macro Definition Documentation

_XDRFILE_H_

#define _XDRFILE_H_

DIM

#define DIM   3