nkf.c

NKF是网络上写新闻或邮件的时候,汉字的编码转换程序

/** Network Kanji Filter. (PDS Version)
************************************************************************
** Copyright (C) 1987, Fujitsu LTD. (Itaru ICHIKAWA)
** $BO"Mm@h!'(B $B!J3t!KIY;NDL8&5f=j!!%=%U%H#38&!!;T@n!!;j(B 
** $B!J(BE-Mail Address: [e-mail]$B!K(B
** Copyright (C) 1996,1998
** $BO"Mm@h!'(B $BN05eBg3X>pJs9)3X2J(B $B2OLn(B $B??<#(B  mine/X0208 support
** $B!J(BE-Mail Address: [e-mail]$B!K(B
** $BO"Mm@h!'(B COW for DOS & Win16 & Win32 & OS/2
** $B!J(BE-Mail Address: [e-mail].p$B!K(B
**    $B$3$N%=!<%9$N$$$+$J$kJ#<L!$2~JQ!$=$@5$b5vBz$7$^$9!#$?$@$7!"(B
**    $B$=$N:]$K$O!"C/$,9W8%$7$?$r<($9$3$NItJ,$r;D$9$3$H!#(B
**    $B:FG[I[$d;(;o$NIUO?$J$I$NLd$$9g$o$;$bI,MW$"$j$^$;$s!#(B
**    $B$3$N%W%m%0%i%`$K$D$$$F$OFC$K2?$NJ]>Z$b$7$J$$!"0-$7$+$i$:!#(B
**    Everyone is permitted to do anything on this program 
**    including copying, modifying, improving.
**    as long as you don't try to pretend that you wrote it.
**    i.e., the above copyright notice has to appear in all copies.
**    You don't have to ask before copying or publishing.
**    THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE.
***********************************************************************/

static char *CopyRight =
      "Copyright (C) 1987, FUJITSU LTD. (I.Ichikawa),1998 S. Kono, COW";
static char *Version =
      "1.7";
static char *Patchlevel =
      "1/9811/Shinji Kono";

/*
**
**
**
** USAGE:       nkf [flags] [file] 
**
** Flags:
** b    Output is bufferred             (DEFAULT)
** u    Output is unbufferred
**
** t    no operation
**
** j    Outout code is JIS 7 bit        (DEFAULT SELECT) 
** s    Output code is MS Kanji         (DEFAULT SELECT) 
** e    Output code is AT&T JIS         (DEFAULT SELECT) 
** l    Output code is JIS 7bit and ISO8859-1 Latin-1
**
** m    MIME conversion for ISO-2022-JP
** i_ Output sequence to designate JIS-kanji (DEFAULT_J)
** o_ Output sequence to designate single-byte roman characters (DEFAULT_R)
**
** r  {de/en}crypt ROT13/47
**
** v  display Version
**
** T  Text mode output        (for MS-DOS)
**
** x    Do not convert X0201 kana into X0208
** Z    Convert X0208 alphabet to ASCII
**
** f60  fold option
**
** m    MIME decode
** B    try to fix broken JIS, missing Escape
** B[1-9]  broken level
**
** O   Output to 'nkf.out' file 
** d   Delete \r in line feed 
** c   Add \r in line feed 
**/
/******************************/
/* $B%G%U%)%k%H$N=PNO%3!<%IA*Br(B */
/* Select DEFAULT_CODE */
#define DEFAULT_CODE_JIS
/* #define DEFAULT_CODE_SJIS */
/* #define DEFAULT_CODE_EUC */
/******************************/

#if (defined(__TURBOC__) || defined(LSI_C)) && !defined(MSDOS)
#define MSDOS
#endif

#ifndef PERL_XS
#include <stdio.h>
#endif

#if defined(MSDOS) || defined(__OS2__) 
#include <stdlib.h>
#include <fcntl.h>
#include <io.h>
#endif

#ifdef MSDOS
#ifdef LSI_C
#define setbinmode(fp) fsetbin(fp)
#else /* Microsoft C, Turbo C */
#define setbinmode(fp) setmode(fileno(fp), O_BINARY)
#endif
#else /* UNIX,OS/2 */
#define setbinmode(fp)
#endif

#ifdef _IOFBF /* SysV and MSDOS */
#define       setvbuffer(fp, buf, size)       setvbuf(fp, buf, _IOFBF, size)
#else /* BSD */
#define       setvbuffer(fp, buf, size)       setbuffer(fp, buf, size)
#endif

/*Borland C++ 4.5 EasyWin*/
#if defined(__TURBOC__) && defined(_Windows) && !defined(__WIN32__) /*Easy Win */
#define         EASYWIN
#include <windows.h>
#endif

#define         FALSE   0
#define         TRUE    1

/* state of output_mode and input_mode  */

#define         ASCII           0
#define         X0208           1
#define         X0201           2
#define         NO_X0201        3
#define         JIS_INPUT       4
#define         SJIS_INPUT      5
#define         LATIN1_INPUT    6
#define         FIXED_MIME      7
#define         DOUBLE_SPACE    -2

#define         NL      0x0a
#define         ESC     0x1b
#define         SPACE      0x20
#define         AT      0x40
#define         SSP     0xa0
#define         DEL     0x7f
#define         SI      0x0f
#define         SO      0x0e
#define         SSO     0x8e

#define         HOLD_SIZE       32
#define         IOBUF_SIZE      16384

#define         DEFAULT_J       'B'
#define         DEFAULT_R       'B'

#define         SJ0162  0x00e1          /* 01 - 62 ku offset */
#define         SJ6394  0x0161          /* 63 - 94 ku offset */


/* MIME preprocessor */

#define STRICT_MIME       /* do stupid strict mime integrity check */
#define GETC(p) ((!mime_mode)?getc(p):mime_getc(p))
#define UNGETC(c,p)     ((!mime_mode)?ungetc(c,p):mime_ungetc(c))


#ifdef EASYWIN /*Easy Win */
extern POINT _BufferSize;
#endif

/*      function prototype  */

static  int     noconvert(FILE *f);
static  int     kanji_convert(FILE *f);
static  int     h_conv(FILE *f,int c2,int c1);
static  int     push_hold_buf(int c2,int c1);
static  int     s_iconv(int c2,int c1);
static  int     e_oconv(int c2,int c1);
static  int     s_oconv(int c2,int c1);
static  int     j_oconv(int c2,int c1);
static  int     line_fold(int c2,int c1);
static  int     pre_convert(int c1,int c2);
static  int     mime_begin(FILE *f);
static  int     mime_getc(FILE *f);
static  int     mime_ungetc(unsigned int c);
static  int     mime_integrity(FILE *f,unsigned char *p);
static  int     base64decode(int c);
static  int     usage(void);
static  void    arguments(char *c);
static  void    reinit();

/* buffers */

static char            stdibuf[IOBUF_SIZE];
static char            stdobuf[IOBUF_SIZE];
static unsigned char   hold_buf[HOLD_SIZE*2];
static int             hold_count;

/* MIME preprocessor fifo */

#define MIME_BUF_SIZE   (1024)    /* 2^n ring buffer */
#define MIME_BUF_MASK   (MIME_BUF_SIZE-1)   
#define Fifo(n)         mime_buf[(n)&MIME_BUF_MASK]
static unsigned char           mime_buf[MIME_BUF_SIZE];
static unsigned int            mime_top = 0;
static unsigned int            mime_last = 0;  /* decoded */
static unsigned int            mime_input = 0; /* undecoded */

/* flags */
static int             unbuf_f = FALSE;
static int             estab_f = FALSE;
static int             nop_f = FALSE;
static int             binmode_f = TRUE;       /* binary mode */
static int             rot_f = FALSE;          /* rot14/43 mode */
static int             input_f = FALSE;        /* non fixed input code  */
static int             alpha_f = FALSE;        /* convert JIx0208 alphbet to ASCII */
static int             mime_f = TRUE;         /* convert MIME B base64 or Q */
static int             mimebuf_f = FALSE;      /* MIME buffered input */
static int             broken_f = FALSE;       /* convert ESC-less broken JIS */
static int             iso8859_f = FALSE;      /* ISO8859 through */
#if defined(MSDOS) || defined(__OS2__) 
static int             x0201_f = TRUE;         /* Assume JISX0201 kana */
#else
static int             x0201_f = NO_X0201;     /* Assume NO JISX0201 */
#endif

/* X0208 -> ASCII converter */

static int             c1_return;

/* fold parameter */
static int line = 0;    /* chars in line */
static int prev = 0;
static int             fold_f  = FALSE;
static int             fold_len  = 0;

/* options */
static char            kanji_intro = DEFAULT_J,
                ascii_intro = DEFAULT_R;

/* Folding */

int line_fold();
#define FOLD_MARGIN  10
#define DEFAULT_FOLD 60

/* converters */

#ifdef DEFAULT_CODE_JIS
#   define  DEFAULT_CONV j_oconv
#endif
#ifdef DEFAULT_CODE_SJIS
#   define  DEFAULT_CONV s_oconv
#endif
#ifdef DEFAULT_CODE_EUC
#   define  DEFAULT_CONV e_oconv
#endif

static int             (*iconv)(int c2,int c1);   
					/* s_iconv or oconv */
static int             (*oconv)(int c2,int c1) = DEFAULT_CONV; 
					  /* [ejs]_oconv */

/* Global states */
static int             output_mode = ASCII,    /* output kanji mode */
                input_mode =  ASCII,    /* input kanji mode */
                shift_mode =  FALSE;    /* TRUE shift out, or X0201  */
static int             mime_mode =   FALSE;    /* MIME mode B base64, Q hex */

/* X0201 / X0208 conversion tables */

/* X0201 kana conversion table */
/* 90-9F A0-DF */
unsigned char cv[]= {
0x21,0x21,0x21,0x23,0x21,0x56,0x21,0x57,
0x21,0x22,0x21,0x26,0x25,0x72,0x25,0x21,
0x25,0x23,0x25,0x25,0x25,0x27,0x25,0x29,
0x25,0x63,0x25,0x65,0x25,0x67,0x25,0x43,
0x21,0x3c,0x25,0x22,0x25,0x24,0x25,0x26,
0x25,0x28,0x25,0x2a,0x25,0x2b,0x25,0x2d,
0x25,0x2f,0x25,0x31,0x25,0x33,0x25,0x35,
0x25,0x37,0x25,0x39,0x25,0x3b,0x25,0x3d,
0x25,0x3f,0x25,0x41,0x25,0x44,0x25,0x46,
0x25,0x48,0x25,0x4a,0x25,0x4b,0x25,0x4c,
0x25,0x4d,0x25,0x4e,0x25,0x4f,0x25,0x52,
0x25,0x55,0x25,0x58,0x25,0x5b,0x25,0x5e,
0x25,0x5f,0x25,0x60,0x25,0x61,0x25,0x62,
0x25,0x64,0x25,0x66,0x25,0x68,0x25,0x69,
0x25,0x6a,0x25,0x6b,0x25,0x6c,0x25,0x6d,
0x25,0x6f,0x25,0x73,0x21,0x2b,0x21,0x2c,
0x00,0x00};


/* X0201 kana conversion table for daguten */
/* 90-9F A0-DF */
unsigned char dv[]= { 
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x25,0x2c,0x25,0x2e,
0x25,0x30,0x25,0x32,0x25,0x34,0x25,0x36,
0x25,0x38,0x25,0x3a,0x25,0x3c,0x25,0x3e,
0x25,0x40,0x25,0x42,0x25,0x45,0x25,0x47,
0x25,0x49,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x25,0x50,0x25,0x53,
0x25,0x56,0x25,0x59,0x25,0x5c,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00};

/* X0201 kana conversion table for han-daguten */
/* 90-9F A0-DF */
unsigned char ev[]= { 
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x25,0x51,0x25,0x54,
0x25,0x57,0x25,0x5a,0x25,0x5d,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00};


/* X0208 kigou conversion table */
/* 0x8140 - 0x819e */
unsigned char fv[] = {

0x00,0x00,0x00,0x00,0x2c,0x2e,0x00,0x3a,
0x3b,0x3f,0x21,0x00,0x00,0x27,0x60,0x00,
0x5e,0x00,0x5f,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x2d,0x00,0x2f,
0x5c,0x00,0x00,0x7c,0x00,0x00,0x60,0x27,
0x22,0x22,0x28,0x29,0x00,0x00,0x5b,0x5d,
0x7b,0x7d,0x3c,0x3e,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x2b,0x2d,0x00,0x00,
0x00,0x3d,0x00,0x3c,0x3e,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x24,0x00,0x00,0x25,0x23,0x26,0x2a,0x40,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
} ;


static int             file_out = FALSE;
static int             add_cr = FALSE;
static int             del_cr = FALSE;
static int             end_check;

#ifndef PERL_XS
int
main(argc, argv)
    int             argc;
    char          **argv;
{
    FILE  *fin;
    char  *cp;

#ifdef EASYWIN /*Easy Win */
    _BufferSize.y = 400;/*Set Scroll Buffer Size*/
#endif

    for (argc--,argv++; (argc > 0) && **argv == '-'; argc--, argv++) {
        cp = *argv;
	arguments(cp);
    }

    if(iso8859_f && (oconv != j_oconv || !x0201_f )) {
        fprintf(stderr,"Mixed ISO8859/JISX0201/SJIS/EUC output is not allowed.\n");
        exit(1);
    }

    if(binmode_f == TRUE)
#ifdef __OS2__
    if(freopen("","wb",stdout) == NULL) 
        return (-1);
#else
    setbinmode(stdout);
#endif

    if(unbuf_f)
      setbuf(stdout, (char *) NULL);
    else
      setvbuffer(stdout, stdobuf, IOBUF_SIZE);

    if(argc == 0) {
      if(binmode_f == TRUE)
#ifdef __OS2__
      if(freopen("","rb",stdin) == NULL) return (-1);
#else
      setbinmode(stdin);
#endif
      setvbuffer(stdin, stdibuf, IOBUF_SIZE);
      if(nop_f)
          noconvert(stdin);
      else
          kanji_convert(stdin);
    } else {
      while (argc--) {
          if((fin = fopen(*argv++, "r")) == NULL) {
              perror(*--argv);
              return(-1);
          } else {
/* reopen file for stdout */
              if(file_out == TRUE){ 
                  if(argc == 1 ) {
                      if(freopen(*argv++, "w", stdout) == NULL) {
                          perror(*--argv);
                          return (-1);
                      }
                      argc--;
                  } else {
                      if(freopen("nkf.out", "w", stdout) == NULL) {
                         perror(*--argv);
                         return (-1);
                      }
                  }
                  if(binmode_f == TRUE) {
#ifdef __OS2__
                      if(freopen("","wb",stdout) == NULL) 
                           return (-1);
#else
                      setbinmode(stdout);
#endif
                  }
              }
              if(binmode_f == TRUE)
#ifdef __OS2__
                 if(freopen("","rb",fin) == NULL) 
                    return (-1);
#else
                 setbinmode(fin);
#endif 
              setvbuffer(fin, stdibuf, IOBUF_SIZE);
              if(nop_f)
                  noconvert(fin);
              else
                  kanji_convert(fin);
              fclose(fin);
          }
      }
    }
#ifdef EASYWIN /*Easy Win */
    if(file_out == FALSE) 
        scanf("%d",&end_check);
    else 
        fclose(stdout);
#else /* for Other OS */
    if(file_out == TRUE) 
        fclose(stdout);
#endif 
    return (0);
}
#endif

void
arguments(char *cp) 
{
    while (*cp) {
	switch (*cp++) {
	case 'b':           /* buffered mode */
	    unbuf_f = FALSE;
	    continue;
	case 'u':           /* non bufferd mode */
	    unbuf_f = TRUE;
	    continue;
	case 't':           /* transparent mode */
	    nop_f = TRUE;
	    continue;
	case 'j':           /* JIS output */
	case 'n':
	    oconv = j_oconv;
	    continue;
	case 'e':           /* AT&T EUC output */