xxd.c

VIM文本编辑器

/* xxd: my hexdump facility. jw
 *
 *  2.10.90 changed to word output
 *  3.03.93 new indent style, dumb bug inserted and fixed.
 *          -c option, mls
 * 26.04.94 better option parser, -ps, -l, -s added.
 *  1.07.94 -r badly needs - as input file.  Per default autoskip over
 *             consequtive lines of zeroes, as unix od does.
 *          -a shows them too.
 *          -i dump as c-style #include "file.h"
 *  1.11.95 if "xxd -i" knows the filename, an 'unsigned char filename_bits[]'
 *          array is written in correct c-syntax.
 *          -s improved, now defaults to absolute seek, relative requires a '+'.
 *          -r improved, now -r -s -0x... is supported.
 *             change/suppress leading '\0' bytes.
 *          -l n improved: stops exactly after n bytes.
 *          -r improved, better handling of partial lines with trailing garbage.
 *          -r improved, now -r -p works again!
 *          -r improved, less flushing, much faster now! (that was silly)
 *  3.04.96 Per repeated request of a single person: autoskip defaults to off.
 * 15.05.96 -v added. They want to know the version.
 *          -a fixed, to show last line inf file ends in all zeros.
 *          -u added: Print upper case hex-letters, as preferred by unix bc.
 *          -h added to usage message. Usage message extended.
 *          Now using outfile if specified even in normal mode, aehem.
 *          No longer mixing of ints and longs. May help doze people.
 *          Added binify ioctl for same reason. (Enough Doze stress for 1996!)
 * 16.05.96 -p improved, removed occasional superfluous linefeed.
 * 20.05.96 -l 0 fixed. tried to read anyway.
 * 21.05.96 -i fixed. now honours -u, and prepends __ to numeric filenames.
 *          compile -DWIN32 for NT or W95. George V. Reilly, * -v improved :-)
 *          support --gnuish-longhorn-options
 * 25.05.96 MAC support added: CodeWarrior already uses ``outline'' in Types.h
 *          which is included by MacHeaders (Axel Kielhorn). Renamed to
 *          xxdline().
 *  7.06.96 -i printed 'int' instead of 'char'. *blush*
 *          added Bram's OS2 ifdefs...
 * 18.07.96 gcc -Wall @ SunOS4 is now slient.
 *          Added osver for MSDOS/DJGPP/WIN32.
 * 29.08.96 Added size_t to strncmp() for Amiga.
 * 24.03.97 Windows NT support (Phil Hanna). Clean exit for Amiga WB (Bram)
 * 02.04.97 Added -E option, to have EBCDIC translation instead of ASCII
 *          ([e-mail])
 * 22.05.97 added -g (group octets) option ([e-mail]).
 *
 * (c) 1990-1997 by Juergen Weigert ([e-mail])
 *
 * Distribute freely and credit me,
 * make money and share with me,
 * lose money and don't ask me.
 */
#include <stdio.h>
#include <fcntl.h>
#ifdef __TSC__
# define MSDOS
#endif
#if !defined(OS2) && defined(__EMX__)
# define OS2
#endif
#if defined(MSDOS) || defined(WIN32) || defined(OS2)
# include <io.h>	/* for setmode() */
#else
# ifdef UNIX
#  include <unistd.h>
# endif
#endif
#include <stdlib.h>
#include <string.h>	/* for strncmp() */
#include <ctype.h>	/* for isalnum() */
#if __MWERKS__ && !defined(__BEOS__)
# include <unix.h>	/* for fdopen() on MAC */
#endif

/*
 * This corrects the problem of missing prototypes for certain functions
 * in some GNU installations (e.g. SunOS 4.1.x).
 */
#if defined(__GNUC__) && defined(__STDC__)
# ifndef __USE_FIXED_PROTOTYPES__
#  define __USE_FIXED_PROTOTYPES__
# endif
#endif

#ifndef __USE_FIXED_PROTOTYPES__
/*
 * This is historic and works only if the compiler really has no prototypes:
 *
 * Include prototypes for Sun OS 4.x, when using an ANSI compiler.
 * FILE is defined on OS 4.x, not on 5.x (Solaris).
 * if __SVR4 is defined (some Solaris versions), don't include this.
 */
#if defined(sun) && defined(FILE) && !defined(__SVR4) && defined(__STDC__)
#  define __P(a) a
/* excerpt from my sun_stdlib.h */
extern int fprintf __P((FILE *, char *, ...));
extern int fputs   __P((char *, FILE *));
extern int _flsbuf __P((unsigned char, FILE *));
extern int _filbuf __P((FILE *));
extern int fflush  __P((FILE *));
extern int fclose  __P((FILE *));
extern int fseek   __P((FILE *, long, int));
extern int rewind  __P((FILE *));

extern void perror __P((char *));
# endif
#endif

extern long int strtol();
extern long int ftell();

char version[] = "xxd V1.8 22may97 by Juergen Weigert";
#ifdef WIN32
char osver[] = " (Win32)";
#else
# ifdef DJGPP
char osver[] = " (dos 32 bit)";
# else
#  ifdef MSDOS
char osver[] = " (dos 16 bit)";
#  else
char osver[] = "";
#  endif
# endif
#endif

#if defined(MSDOS) || defined(WIN32) || defined(OS2)
# define BIN_READ(yes)  ((yes) ? "rb" : "rt")
# define BIN_WRITE(yes) ((yes) ? "wb" : "wt")
# define BIN_CREAT(yes) ((yes) ? (O_CREAT|O_BINARY) : O_CREAT)
# define BIN_ASSIGN(fp, yes) setmode(fileno(fp), (yes) ? O_BINARY : O_TEXT)
# define PATH_SEP '\\'
#else
# define BIN_READ(dummy)  "r"
# define BIN_WRITE(dummy) "w"
# define BIN_CREAT(dummy) O_CREAT
# define BIN_ASSIGN(fp, dummy) fp
# define PATH_SEP '/'
#endif

/* open has only to arguments on the Mac */
#if __MWERKS__
# define OPEN(name, mode, umask) open(name, mode)
#else
# define OPEN(name, mode, umask) open(name, mode, umask)
#endif

#ifdef AMIGA
# define STRNCMP(s1, s2, l) strncmp(s1, s2, (size_t)l)
#else
# define STRNCMP(s1, s2, l) strncmp(s1, s2, l)
#endif

#ifndef __P
# ifdef __STDC__
#  define __P(a) a
# else
#  define __P(a) ()
# endif
#endif

/* Let's collect some prototypes */
/* CodeWarrior is really picky about missing prototypes */
static void exit_with_usage __P((char *));
static int huntype __P((FILE *, FILE *, FILE *, char *, int, int, long));
static void xxdline __P((FILE *, char *, int));

#define TRY_SEEK	/* attempt to use lseek, or skip forward by reading */
#define COLS 256	/* change here, if you ever need more columns */
#define LLEN (9 + 3*COLS + 2 + COLS)

char hexxa[] = "0123456789abcdef0123456789ABCDEF", *hexx = hexxa;

/* the different hextypes known by this program: */
#define HEX_NORMAL 0
#define HEX_POSTSCRIPT 1
#define HEX_CINCLUDE 2

static void
exit_with_usage(pname)
char *pname;
{
  fprintf(stderr, "Usage:\n       %s [options] [infile [outfile]]\n", pname);
  fprintf(stderr, "    or\n       %s -r [-s [-]offset] [-c cols] [-ps] [infile [outfile]]\n", pname);
  fprintf(stderr, "Options:\n");
  fprintf(stderr, "    -a          toggle autoskip: A single '*' replaces nul-lines. Default off.\n");
  fprintf(stderr, "    -c cols     format <cols> octets per line. Default 16 (-i: 12, -ps: 30).\n");
  fprintf(stderr, "    -E          show characters in EBCDIC. Default ASCII.\n");
  fprintf(stderr, "    -g          number of octets per group in normal output. Default 2.\n");
  fprintf(stderr, "    -h          print this summary.\n");
  fprintf(stderr, "    -i          output in C include file style.\n");
  fprintf(stderr, "    -l len      stop after <len> octets.\n");
  fprintf(stderr, "    -ps         output in postscript continuous hexdump style.\n");
  fprintf(stderr, "    -r          reverse operation: convert (or patch) hexdump into binary.\n");
  fprintf(stderr, "    -r -s off   revert with <off> added to file positions found in hexdump.\n");
  fprintf(stderr, "    -s %sseek  start at <seek> bytes abs. %sinfile offset.\n",
#ifdef TRY_SEEK
	  "[+][-]", "(or +: rel.) ");
#else
	  "", "");
#endif
  fprintf(stderr, "    -u          use upper case hex letters.\n");
  fprintf(stderr, "    -v          show version: \"%s%s\".\n", version, osver);
  exit(1);
}

/*
 * Max. cols binary characters are decoded from the input stream per line.
 * Two adjacent garbage characters after evaluated data delimit valid data.
 * Everything up to the next newline is discarded.
 *
 * The name is historic and came from 'undo type opt h'.
 */
static int
huntype(fpi, fpo, fperr, pname, cols, hextype, base_off)
FILE *fpi, *fpo, *fperr;
char *pname;
int cols, hextype;
long base_off;
{
  int c, ign_garb = 1, n1 = -1, n2 = 0, n3, p = cols;
  long have_off = 0, want_off = 0;

  rewind(fpi);

  while ((c = getc(fpi)) != EOF)
    {
      if (c == '\r')	/* Doze style input file? */
        continue;

      n3 = n2;
      n2 = n1;

      if (c >= '0' && c <= '9')
        n1 = c - '0';
      else if (c >= 'a' && c <= 'f')
        n1 = c - 'a' + 10;
      else if (c >= 'A' && c <= 'F')
        n1 = c - 'A' + 10;
      else
        {
          n1 = -1;
	  if (ign_garb)
	    continue;
	}

      ign_garb = 0;

      if (p >= cols)
	{
	  if (!hextype)
	    {
	      if (n1 < 0)
		{
		  p = 0;
		  continue;
		}
	      want_off = (want_off << 4) | n1;
	      continue;
	    }
	  else
	    p = 0;
	}

      if (base_off + want_off != have_off)
        {
	  fflush(fpo);
#ifdef TRY_SEEK
	  c = fseek(fpo, base_off + want_off - have_off, 1);
	  if (c >= 0)
	    have_off = base_off + want_off;
#endif
	  if (base_off + want_off < have_off)
	    {
	      fprintf(fperr, "%s: sorry, cannot seek backwards.\n", pname);
	      return 5;
	    }
	  for (; have_off < base_off + want_off; have_off++)
	    putc(0, fpo);
	}

      if (n2 >= 0 && n1 >= 0)
        {
	  putc((n2 << 4) | n1, fpo);
	  have_off++;
	  want_off++;
	  n1 = -1;
	  if (++p >= cols)
	    {
	      if (!hextype)
	        want_off = 0;
	      while ((c = getc(fpi)) != '\n' && c != EOF)
	        ;
	      ign_garb = 1;
	    }
	}
      else if (n1 < 0 && n2 < 0 && n3 < 0)
        {
	  if (!hextype)
	    want_off = 0;
	  while ((c = getc(fpi)) != '\n' && c != EOF)
	    ;
	  ign_garb = 1;
	}
    }
  fflush(fpo);
#ifdef TRY_SEEK
  fseek(fpo, 0L, 2);
#endif
  fclose(fpo);
  fclose(fpi);
  return 0;
}

/*
 * Print line l. If nz is false, xxdline regards the line a line of
 * zeroes. If there are three or more consecutive lines of zeroes,
 * they are replaced by a single '*' character.
 *
 * If the output ends with more than two lines of zeroes, you
 * should call xxdline again with l being the last line and nz
 * negative. This ensures that the last line is shown even when
 * it is all zeroes.
 *
 * If nz is always positive, lines are never suppressed.
 */
static void
xxdline(fp, l, nz)
FILE *fp;
char *l;
int nz;
{
  static char z[LLEN+1];
  static int zero_seen = 0;

  if (!nz && zero_seen == 1)
    strcpy(z, l);

  if (nz || !zero_seen++)
    {
      if (nz)
	{