rfc2047.c

mutt-1.5.12 源代码。linux 下邮件接受的工具。

/*
 * Copyright (C) 1996-2000 Michael R. Elkins <[e-mail]>
 * Copyright (C) 2000-2001 Edmund Grimley Evans <[e-mail]>
 * 
 *     This program is free software; you can redistribute it and/or modify
 *     it under the terms of the GNU General Public License as published by
 *     the Free Software Foundation; either version 2 of the License, or
 *     (at your option) any later version.
 * 
 *     This program 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 General Public License for more details.
 * 
 *     You should have received a copy of the GNU General Public License
 *     along with this program; if not, write to the Free Software
 *     Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
 */ 

#if HAVE_CONFIG_H
# include "config.h"
#endif

#include "mutt.h"
#include "mime.h"
#include "charset.h"
#include "rfc2047.h"

#include <ctype.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

/* If you are debugging this file, comment out the following line. */
/*#define NDEBUG*/

#ifdef NDEBUG
#define assert(x)
#else
#include <assert.h>
#endif

#define ENCWORD_LEN_MAX 75
#define ENCWORD_LEN_MIN 9 /* strlen ("=?.?.?.?=") */

#define HSPACE(x) ((x) == '\0' || (x) == ' ' || (x) == '\t')

#define CONTINUATION_BYTE(c) (((c) & 0xc0) == 0x80)

extern char RFC822Specials[];

typedef size_t (*encoder_t) (char *, ICONV_CONST char *, size_t,
			     const char *);

static size_t convert_string (ICONV_CONST char *f, size_t flen,
			      const char *from, const char *to,
			      char **t, size_t *tlen)
{
  iconv_t cd;
  char *buf, *ob;
  size_t obl, n;
  int e;

  cd = mutt_iconv_open (to, from, 0);
  if (cd == (iconv_t)(-1))
    return (size_t)(-1);
  obl = 4 * flen + 1;
  ob = buf = safe_malloc (obl);
  n = iconv (cd, &f, &flen, &ob, &obl);
  if (n == (size_t)(-1) || iconv (cd, 0, 0, &ob, &obl) == (size_t)(-1))
  {
    e = errno;
    FREE (&buf);
    iconv_close (cd);
    errno = e;
    return (size_t)(-1);
  }
  *ob = '\0';
  
  *tlen = ob - buf;

  safe_realloc (&buf, ob - buf + 1);
  *t = buf;
  iconv_close (cd);

  return n;
}

char *mutt_choose_charset (const char *fromcode, const char *charsets,
		      char *u, size_t ulen, char **d, size_t *dlen)
{
  char canonical_buff[LONG_STRING];
  char *e = 0, *tocode = 0;
  size_t elen = 0, bestn = 0;
  const char *p, *q;

  for (p = charsets; p; p = q ? q + 1 : 0)
  {
    char *s, *t;
    size_t slen, n;

    q = strchr (p, ':');

    n = q ? q - p : strlen (p);

    if (!n ||
	/* Assume that we never need more than 12 characters of
	   encoded-text to encode a single character. */
	n > (ENCWORD_LEN_MAX - ENCWORD_LEN_MIN + 2 - 12))
      continue;

    t = safe_malloc (n + 1);
    memcpy (t, p, n);
    t[n] = '\0';

    n = convert_string (u, ulen, fromcode, t, &s, &slen);
    if (n == (size_t)(-1))
      continue;

    if (!tocode || n < bestn)
    {
      bestn = n;
      FREE (&tocode);
      tocode = t;
      if (d)
      {
	FREE (&e);
	e = s;
      }
      else
	FREE (&s);
      elen = slen;
      if (!bestn)
	break;
    }
    else
    {
      FREE (&t);
      FREE (&s);
    }
  }
  if (tocode)
  {
    if (d)
      *d = e;
    if (dlen)
      *dlen = elen;
    
    mutt_canonical_charset (canonical_buff, sizeof (canonical_buff), tocode);
    mutt_str_replace (&tocode, canonical_buff);
  }
  return tocode;
}

static size_t b_encoder (char *s, ICONV_CONST char *d, size_t dlen,
			 const char *tocode)
{
  char *s0 = s;

  memcpy (s, "=?", 2), s += 2;
  memcpy (s, tocode, strlen (tocode)), s += strlen (tocode);
  memcpy (s, "?B?", 3), s += 3;
  for (;;)
  {
    if (!dlen)
      break;
    else if (dlen == 1)
    {
      *s++ = B64Chars[(*d >> 2) & 0x3f];
      *s++ = B64Chars[(*d & 0x03) << 4];
      *s++ = '=';
      *s++ = '=';
      break;
    }
    else if (dlen == 2)
    {
      *s++ = B64Chars[(*d >> 2) & 0x3f];
      *s++ = B64Chars[((*d & 0x03) << 4) | ((d[1] >> 4) & 0x0f)];
      *s++ = B64Chars[(d[1] & 0x0f) << 2];
      *s++ = '=';
      break;
    }
    else
    {
      *s++ = B64Chars[(*d >> 2) & 0x3f];
      *s++ = B64Chars[((*d & 0x03) << 4) | ((d[1] >> 4) & 0x0f)];
      *s++ = B64Chars[((d[1] & 0x0f) << 2) | ((d[2] >> 6) & 0x03)];
      *s++ = B64Chars[d[2] & 0x3f];
      d += 3, dlen -= 3;
    }
  }
  memcpy (s, "?=", 2), s += 2;
  return s - s0;
}

static size_t q_encoder (char *s, ICONV_CONST char *d, size_t dlen,
			 const char *tocode)
{
  char hex[] = "0123456789ABCDEF";
  char *s0 = s;

  memcpy (s, "=?", 2), s += 2;
  memcpy (s, tocode, strlen (tocode)), s += strlen (tocode);
  memcpy (s, "?Q?", 3), s += 3;
  while (dlen--)
  {
    unsigned char c = *d++;
    if (c == ' ')
      *s++ = '_';
    else if (c >= 0x7f || c < 0x20 || c == '_' ||  strchr (MimeSpecials, c))
    {
      *s++ = '=';
      *s++ = hex[(c & 0xf0) >> 4];
      *s++ = hex[c & 0x0f];
    }
    else
      *s++ = c;
  }
  memcpy (s, "?=", 2), s += 2;
  return s - s0;
}

/*
 * Return 0 if and set *encoder and *wlen if the data (d, dlen) could
 * be converted to an encoded word of length *wlen using *encoder.
 * Otherwise return an upper bound on the maximum length of the data
 * which could be converted.
 * The data is converted from fromcode (which must be stateless) to
 * tocode, unless fromcode is 0, in which case the data is assumed to
 * be already in tocode, which should be 8-bit and stateless.
 */
static size_t try_block (ICONV_CONST char *d, size_t dlen,
			 const char *fromcode, const char *tocode,
			 encoder_t *encoder, size_t *wlen)
{
  char buf1[ENCWORD_LEN_MAX - ENCWORD_LEN_MIN + 1];
  iconv_t cd;
  ICONV_CONST char *ib;
  char *ob, *p;
  size_t ibl, obl;
  int count, len, len_b, len_q;

  if (fromcode)
  {
    cd = mutt_iconv_open (tocode, fromcode, 0);
    assert (cd != (iconv_t)(-1));
    ib = d, ibl = dlen, ob = buf1, obl = sizeof (buf1) - strlen (tocode);
    if (iconv (cd, &ib, &ibl, &ob, &obl) == (size_t)(-1) ||
	iconv (cd, 0, 0, &ob, &obl) == (size_t)(-1))
    {
      assert (errno == E2BIG);
      iconv_close (cd);
      assert (ib > d);
      return (ib - d == dlen) ? dlen : ib - d + 1;
    }
    iconv_close (cd);
  }
  else
  {
    if (dlen > sizeof (buf1) - strlen (tocode))
      return sizeof (buf1) - strlen (tocode) + 1;
    memcpy (buf1, d, dlen);
    ob = buf1 + dlen;
  }

  count = 0;
  for (p = buf1; p < ob; p++)
  {
    unsigned char c = *p;
    assert (strchr (MimeSpecials, '?'));
    if (c >= 0x7f || c < 0x20 || *p == '_' ||
	(c != ' ' && strchr (MimeSpecials, *p)))
      ++count;
  }

  len = ENCWORD_LEN_MIN - 2 + strlen (tocode);
  len_b = len + (((ob - buf1) + 2) / 3) * 4;
  len_q = len + (ob - buf1) + 2 * count;

  /* Apparently RFC 1468 says to use B encoding for iso-2022-jp. */
  if (!ascii_strcasecmp (tocode, "ISO-2022-JP"))
    len_q = ENCWORD_LEN_MAX + 1;

  if (len_b < len_q && len_b <= ENCWORD_LEN_MAX)
  {
    *encoder = b_encoder;
    *wlen = len_b;
    return 0;
  }
  else if (len_q <= ENCWORD_LEN_MAX)
  {
    *encoder = q_encoder;
    *wlen = len_q;
    return 0;
  }
  else
    return dlen;
}

/*
 * Encode the data (d, dlen) into s using the encoder.
 * Return the length of the encoded word.
 */
static size_t encode_block (char *s, char *d, size_t dlen,
			    const char *fromcode, const char *tocode,
			    encoder_t encoder)
{
  char buf1[ENCWORD_LEN_MAX - ENCWORD_LEN_MIN + 1];
  iconv_t cd;
  ICONV_CONST char *ib;
  char *ob;
  size_t ibl, obl, n1, n2;

  if (fromcode)
  {
    cd = mutt_iconv_open (tocode, fromcode, 0);
    assert (cd != (iconv_t)(-1));
    ib = d, ibl = dlen, ob = buf1, obl = sizeof (buf1) - strlen (tocode);
    n1 = iconv (cd, &ib, &ibl, &ob, &obl);
    n2 = iconv (cd, 0, 0, &ob, &obl);
    assert (n1 != (size_t)(-1) && n2 != (size_t)(-1));
    iconv_close (cd);
    return (*encoder) (s, buf1, ob - buf1, tocode);
  }
  else
    return (*encoder) (s, d, dlen, tocode);
}

/*
 * Discover how much of the data (d, dlen) can be converted into
 * a single encoded word. Return how much data can be converted,
 * and set the length *wlen of the encoded word and *encoder.
 * We start in column col, which limits the length of the word.
 */
static size_t choose_block (char *d, size_t dlen, int col,
			    const char *fromcode, const char *tocode,
			    encoder_t *encoder, size_t *wlen)
{
  size_t n, nn;
  int utf8 = fromcode && !ascii_strcasecmp (fromcode, "UTF-8");

  n = dlen;
  for (;;)
  {
    assert (d + n > d);
    nn = try_block (d, n, fromcode, tocode, encoder, wlen);
    if (!nn && (col + *wlen <= ENCWORD_LEN_MAX + 1 || n <= 1))
      break;
    n = (nn ? nn : n) - 1;
    assert (n > 0);
    if (utf8)
      while (n > 1 && CONTINUATION_BYTE(d[n]))
	--n;
  }
  return n;
}

/*
 * Place the result of RFC-2047-encoding (d, dlen) into the dynamically
 * allocated buffer (e, elen). The input data is in charset fromcode
 * and is converted into a charset chosen from charsets.
 * Return 1 if the conversion to UTF-8 failed, 2 if conversion from UTF-8
 * failed, otherwise 0. If conversion failed, fromcode is assumed to be
 * compatible with us-ascii and the original data is used.
 * The input data is assumed to be a single line starting at column col;
 * if col is non-zero, the preceding character was a space.
 */
static int rfc2047_encode (ICONV_CONST char *d, size_t dlen, int col,
			   const char *fromcode, const char *charsets,
			   char **e, size_t *elen, char *specials)
{
  int ret = 0;
  char *buf;
  size_t bufpos, buflen;
  char *u, *t0, *t1, *t;
  char *s0, *s1;
  size_t ulen, r, n, wlen;
  encoder_t encoder;
  char *tocode1 = 0;
  const char *tocode;
  char *icode = "UTF-8";

  /* Try to convert to UTF-8. */
  if (convert_string (d, dlen, fromcode, icode, &u, &ulen))
  {
    ret = 1; 
    icode = 0;
    u = safe_malloc ((ulen = dlen) + 1);
    memcpy (u, d, dlen);
    u[ulen] = 0;
  }