[exim.git] / src / src / base64.c

/*************************************************
*     Exim - an Internet mail transport agent    *
*************************************************/

/* Copyright (c) Tom Kistner <tom@duncanthrax.net> 2004, 2015 */
/* License: GPL */

/* Copyright (c) University of Cambridge 1995 - 2018 */
/* See the file NOTICE for conditions of use and distribution. */


#include "exim.h"
#ifdef WITH_CONTENT_SCAN	/* file-IO specific decode function */
# include "mime.h"

/* BASE64 decoder matrix */
static unsigned char mime_b64[256]={
/*   0 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/*  16 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/*  32 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,   62,  128,  128,  128,   63,
/*  48 */   52,   53,   54,   55,   56,   57,   58,   59,   60,   61,  128,  128,  128,  255,  128,  128,
/*  64 */  128,    0,    1,    2,    3,    4,    5,    6,    7,    8,    9,   10,   11,   12,   13,   14,
/*  80 */   15,   16,   17,   18,   19,   20,   21,   22,   23,   24,   25,  128,  128,  128,  128,  128,
/*  96 */  128,   26,   27,   28,   29,   30,   31,   32,   33,   34,   35,   36,   37,   38,   39,   40,
/* 112 */   41,   42,   43,   44,   45,   46,   47,   48,   49,   50,   51,  128,  128,  128,  128,  128,
/* 128 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 144 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 160 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 176 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 192 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 208 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 224 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,
/* 240 */  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128,  128
};

/* decode base64 MIME part */
ssize_t
mime_decode_base64(FILE * in, FILE * out, uschar * boundary)
{
uschar ibuf[MIME_MAX_LINE_LENGTH], obuf[MIME_MAX_LINE_LENGTH];
uschar *opos;
ssize_t len, size = 0;
int bytestate = 0;

opos = obuf;

while (Ufgets(ibuf, MIME_MAX_LINE_LENGTH, in) != NULL)
  {
  if (boundary != NULL
     && Ustrncmp(ibuf, "--", 2) == 0
     && Ustrncmp((ibuf+2), boundary, Ustrlen(boundary)) == 0
     )
    break;

  for (uschar * ipos = ibuf ; *ipos != '\r' && *ipos != '\n' && *ipos; ++ipos)
    if (*ipos == '=')			/* skip padding */
      ++bytestate;

    else if (mime_b64[*ipos] == 128)	/* skip bad characters */
      mime_set_anomaly(MIME_ANOMALY_BROKEN_BASE64);

    /* simple state-machine */
    else switch((bytestate++) & 3)
      {
      case 0:
	*opos = mime_b64[*ipos] << 2; break;
      case 1:
	*opos++ |= mime_b64[*ipos] >> 4;
	*opos = mime_b64[*ipos] << 4; break;
      case 2:
	*opos++ |= mime_b64[*ipos] >> 2;
	*opos = mime_b64[*ipos] << 6; break;
      case 3:
	*opos++ |= mime_b64[*ipos]; break;
      }

  /* something to write? */
  len = opos - obuf;
  if (len > 0)
    {
    if (fwrite(obuf, 1, len, out) != len) return -1; /* error */
    size += len;
    /* copy incomplete last byte to start of obuf, where we continue */
    if ((bytestate & 3) != 0)
      *obuf = *opos;
    opos = obuf;
    }
  } /* while */

/* write out last byte if it was incomplete */
if (bytestate & 3)
  {
  if (fwrite(obuf, 1, 1, out) != 1) return -1;
  ++size;
  }

return size;
}

#endif	/*WITH_CONTENT_SCAN*/

/*************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************
 *************************************************/


/*************************************************
*          Decode byte-string in base 64         *
*************************************************/

/* This function decodes a string in base 64 format as defined in RFC 2045
(MIME) and required by the SMTP AUTH extension (RFC 2554). The decoding
algorithm is written out in a straightforward way. Turning it into some kind of
compact loop is messy and would probably run more slowly.

Arguments:
  code        points to the coded string, zero-terminated
  ptr         where to put the pointer to the result, which is in
              allocated store, and zero-terminated

Returns:      the number of bytes in the result,
              or -1 if the input was malformed

Whitespace in the input is ignored.
A zero is added on to the end to make it easy in cases where the result is to
be interpreted as text. This is not included in the count. */

static uschar dec64table[] = {
  255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, /*  0-15 */
  255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, /* 16-31 */
  255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, /* 32-47 */
   52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255,255,255,255, /* 48-63 */
  255,  0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, /* 64-79 */
   15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255, /* 80-95 */
  255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, /* 96-111 */
   41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255  /* 112-127*/
};

int
b64decode(const uschar *code, uschar **ptr)
{

int x, y;
uschar *result;

{
  int l = Ustrlen(code);
  *ptr = result = store_get(1 + l/4 * 3 + l%4, is_tainted(code));
}

/* Each cycle of the loop handles a quantum of 4 input bytes. For the last
quantum this may decode to 1, 2, or 3 output bytes. */

while ((x = *code++) != 0)
  {
  if (isspace(x)) continue;
  /* debug_printf("b64d: '%c'\n", x); */

  if (x > 127 || (x = dec64table[x]) == 255) return -1;

  while (isspace(y = *code++)) ;
  /* debug_printf("b64d: '%c'\n", y); */
  if (y > 127 || (y = dec64table[y]) == 255)
    return -1;

  *result++ = (x << 2) | (y >> 4);
  /* debug_printf("b64d:      -> %02x\n", result[-1]); */

  while (isspace(x = *code++)) ;
  /* debug_printf("b64d: '%c'\n", x); */
  if (x == '=')		/* endmarker, but there should be another */
    {
    while (isspace(x = *code++)) ;
    /* debug_printf("b64d: '%c'\n", x); */
    if (x != '=') return -1;
    while (isspace(y = *code++)) ;
    if (y != 0) return -1;
    /* debug_printf("b64d: DONE\n"); */
    break;
    }
  else
    {
    if (x > 127 || (x = dec64table[x]) == 255) return -1;
    *result++ = (y << 4) | (x >> 2);
    /* debug_printf("b64d:      -> %02x\n", result[-1]); */

    while (isspace(y = *code++)) ;
    /* debug_printf("b64d: '%c'\n", y); */
    if (y == '=')
      {
      while (isspace(y = *code++)) ;
      if (y != 0) return -1;
      /* debug_printf("b64d: DONE\n"); */
      break;
      }
    else
      {
      if (y > 127 || (y = dec64table[y]) == 255) return -1;
      *result++ = (x << 6) | y;
      /* debug_printf("b64d:      -> %02x\n", result[-1]); */
      }
    }
  }

*result = 0;
return result - *ptr;
}


/*************************************************
*          Encode byte-string in base 64         *
*************************************************/

/* This function encodes a string of bytes, containing any values whatsoever,
in base 64 as defined in RFC 2045 (MIME) and required by the SMTP AUTH
extension (RFC 2554). The encoding algorithm is written out in a
straightforward way. Turning it into some kind of compact loop is messy and
would probably run more slowly.

Arguments:
  clear       points to the clear text bytes
  len         the number of bytes to encode

Returns:      a pointer to the zero-terminated base 64 string, which
              is in working store
*/

static uschar *enc64table =
  US"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";

uschar *
b64encode(const uschar * clear, int len)
{
uschar *code = store_get(4*((len+2)/3) + 1, is_tainted(clear));
uschar *p = code;

while (len-- >0)
  {
  int x, y;

  x = *clear++;
  *p++ = enc64table[(x >> 2) & 63];

  if (len-- <= 0)
    {
    *p++ = enc64table[(x << 4) & 63];
    *p++ = '=';
    *p++ = '=';
    break;
    }

  y = *clear++;
  *p++ = enc64table[((x << 4) | ((y >> 4) & 15)) & 63];

  if (len-- <= 0)
    {
    *p++ = enc64table[(y << 2) & 63];
    *p++ = '=';
    break;
    }

  x = *clear++;
  *p++ = enc64table[((y << 2) | ((x >> 6) & 3)) & 63];

  *p++ = enc64table[x & 63];
  }

*p = 0;

return code;
}


/* End of base64.c */
/* vi: sw ai sw=2
*/
Commit	Line	Data
f4d091fb JH	1	/*************************************************
	2	* Exim - an Internet mail transport agent *
	3	*************************************************/
	4
	5	/* Copyright (c) Tom Kistner <tom@duncanthrax.net> 2004, 2015 */
	6	/* License: GPL */
	7
f9ba5e22	8	/* Copyright (c) University of Cambridge 1995 - 2018 */
f4d091fb JH	9	/* See the file NOTICE for conditions of use and distribution. */
	10
	11
	12	#include "exim.h"
	13	#ifdef WITH_CONTENT_SCAN /* file-IO specific decode function */
	14	# include "mime.h"
	15
	16	/* BASE64 decoder matrix */
	17	static unsigned char mime_b64[256]={
	18	/* 0 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	19	/* 16 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	20	/* 32 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 62, 128, 128, 128, 63,
	21	/* 48 */ 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 128, 128, 128, 255, 128, 128,
	22	/* 64 */ 128, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
	23	/* 80 */ 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 128, 128, 128, 128, 128,
	24	/* 96 */ 128, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
	25	/* 112 */ 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 128, 128, 128, 128, 128,
	26	/* 128 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	27	/* 144 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	28	/* 160 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	29	/* 176 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	30	/* 192 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	31	/* 208 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	32	/* 224 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
	33	/* 240 */ 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128
	34	};
	35
	36	/* decode base64 MIME part */
	37	ssize_t
	38	mime_decode_base64(FILE * in, FILE * out, uschar * boundary)
	39	{
	40	uschar ibuf[MIME_MAX_LINE_LENGTH], obuf[MIME_MAX_LINE_LENGTH];
d7978c0f	41	uschar *opos;
f4d091fb JH	42	ssize_t len, size = 0;
	43	int bytestate = 0;
	44
	45	opos = obuf;
	46
	47	while (Ufgets(ibuf, MIME_MAX_LINE_LENGTH, in) != NULL)
	48	{
	49	if (boundary != NULL
	50	&& Ustrncmp(ibuf, "--", 2) == 0
	51	&& Ustrncmp((ibuf+2), boundary, Ustrlen(boundary)) == 0
	52	)
	53	break;
	54
d7978c0f	55	for (uschar * ipos = ibuf ; ipos != '\r' && ipos != '\n' && *ipos; ++ipos)
f4d091fb JH	56	if (ipos == '=') / skip padding */
	57	++bytestate;
	58
	59	else if (mime_b64[ipos] == 128) / skip bad characters */
	60	mime_set_anomaly(MIME_ANOMALY_BROKEN_BASE64);
	61
	62	/* simple state-machine */
	63	else switch((bytestate++) & 3)
	64	{
	65	case 0:
	66	opos = mime_b64[ipos] << 2; break;
	67	case 1:
	68	opos++ \|= mime_b64[ipos] >> 4;
	69	opos = mime_b64[ipos] << 4; break;
	70	case 2:
	71	opos++ \|= mime_b64[ipos] >> 2;
	72	opos = mime_b64[ipos] << 6; break;
	73	case 3:
	74	opos++ \|= mime_b64[ipos]; break;
	75	}
	76
	77	/* something to write? */
	78	len = opos - obuf;
	79	if (len > 0)
	80	{
	81	if (fwrite(obuf, 1, len, out) != len) return -1; /* error */
	82	size += len;
	83	/* copy incomplete last byte to start of obuf, where we continue */
	84	if ((bytestate & 3) != 0)
	85	obuf = opos;
	86	opos = obuf;
	87	}
	88	} /* while */
	89
	90	/* write out last byte if it was incomplete */
	91	if (bytestate & 3)
	92	{
	93	if (fwrite(obuf, 1, 1, out) != 1) return -1;
	94	++size;
	95	}
	96
	97	return size;
	98	}
	99
	100	#endif /WITH_CONTENT_SCAN/
	101
	102	/*************************************************
3c51463e JH	103	*************************************************
	104	*************************************************
	105	*************************************************
	106	*************************************************
	107	*************************************************
	108	*************************************************
	109	*************************************************
	110	*************************************************
	111	*************************************************
	112	*************************************************
	113	*************************************************
	114	*************************************************
	115	*************************************************
	116	*************************************************
	117	*************************************************/
f4d091fb JH	118
	119
	120	/*************************************************
	121	* Decode byte-string in base 64 *
	122	*************************************************/
	123
	124	/* This function decodes a string in base 64 format as defined in RFC 2045
	125	(MIME) and required by the SMTP AUTH extension (RFC 2554). The decoding
	126	algorithm is written out in a straightforward way. Turning it into some kind of
	127	compact loop is messy and would probably run more slowly.
	128
	129	Arguments:
	130	code points to the coded string, zero-terminated
	131	ptr where to put the pointer to the result, which is in
df3def24	132	allocated store, and zero-terminated
f4d091fb JH	133
	134	Returns: the number of bytes in the result,
	135	or -1 if the input was malformed
	136
69a70afa	137	Whitespace in the input is ignored.
f4d091fb JH	138	A zero is added on to the end to make it easy in cases where the result is to
	139	be interpreted as text. This is not included in the count. */
	140
	141	static uschar dec64table[] = {
	142	255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, /* 0-15 */
	143	255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, /* 16-31 */
	144	255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, /* 32-47 */
	145	52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255,255,255,255, /* 48-63 */
	146	255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, /* 64-79 */
	147	15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255, /* 80-95 */
	148	255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, /* 96-111 */
	149	41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255 /* 112-127*/
	150	};
	151
	152	int
35cf75e9	153	b64decode(const uschar code, uschar *ptr)
f4d091fb	154	{
cf3cd306	155
f4d091fb	156	int x, y;
cf3cd306	157	uschar *result;
f4d091fb	158
cf3cd306 HSHR	159	{
cf3cd306 HSHR	160	int l = Ustrlen(code);
f3ebb786	161	ptr = result = store_get(1 + l/4 3 + l%4, is_tainted(code));
cf3cd306	162	}
f4d091fb JH	163
	164	/* Each cycle of the loop handles a quantum of 4 input bytes. For the last
	165	quantum this may decode to 1, 2, or 3 output bytes. */
	166
	167	while ((x = *code++) != 0)
	168	{
69a70afa	169	if (isspace(x)) continue;
df3def24	170	/* debug_printf("b64d: '%c'\n", x); */
69a70afa	171
f4d091fb	172	if (x > 127 \|\| (x = dec64table[x]) == 255) return -1;
69a70afa JH	173
69a70afa JH	174	while (isspace(y = *code++)) ;
df3def24	175	/* debug_printf("b64d: '%c'\n", y); */
889d293b	176	if (y > 127 \|\| (y = dec64table[y]) == 255)
f4d091fb JH	177	return -1;
	178
	179	*result++ = (x << 2) \| (y >> 4);
df3def24	180	/* debug_printf("b64d: -> %02x\n", result[-1]); */
f4d091fb	181
69a70afa	182	while (isspace(x = *code++)) ;
df3def24 JH	183	/* debug_printf("b64d: '%c'\n", x); */
df3def24 JH	184	if (x == '=') /* endmarker, but there should be another */
f4d091fb	185	{
69a70afa	186	while (isspace(x = *code++)) ;
df3def24 JH	187	/* debug_printf("b64d: '%c'\n", x); */
	188	if (x != '=') return -1;
	189	while (isspace(y = *code++)) ;
	190	if (y != 0) return -1;
	191	/* debug_printf("b64d: DONE\n"); */
	192	break;
f4d091fb JH	193	}
	194	else
	195	{
	196	if (x > 127 \|\| (x = dec64table[x]) == 255) return -1;
	197	*result++ = (y << 4) \| (x >> 2);
df3def24	198	/* debug_printf("b64d: -> %02x\n", result[-1]); */
69a70afa JH	199
69a70afa JH	200	while (isspace(y = *code++)) ;
df3def24	201	/* debug_printf("b64d: '%c'\n", y); */
69a70afa	202	if (y == '=')
f4d091fb	203	{
df3def24 JH	204	while (isspace(y = *code++)) ;
	205	if (y != 0) return -1;
	206	/* debug_printf("b64d: DONE\n"); */
	207	break;
f4d091fb JH	208	}
	209	else
	210	{
	211	if (y > 127 \|\| (y = dec64table[y]) == 255) return -1;
	212	*result++ = (x << 6) \| y;
df3def24	213	/* debug_printf("b64d: -> %02x\n", result[-1]); */
f4d091fb JH	214	}
	215	}
	216	}
	217
	218	*result = 0;
	219	return result - *ptr;
	220	}
	221
	222
f4d091fb JH	223	/*************************************************
	224	* Encode byte-string in base 64 *
	225	*************************************************/
	226
	227	/* This function encodes a string of bytes, containing any values whatsoever,
	228	in base 64 as defined in RFC 2045 (MIME) and required by the SMTP AUTH
	229	extension (RFC 2554). The encoding algorithm is written out in a
	230	straightforward way. Turning it into some kind of compact loop is messy and
	231	would probably run more slowly.
	232
	233	Arguments:
	234	clear points to the clear text bytes
	235	len the number of bytes to encode
	236
	237	Returns: a pointer to the zero-terminated base 64 string, which
	238	is in working store
	239	*/
	240
	241	static uschar *enc64table =
	242	US"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
	243
	244	uschar *
1f20760b	245	b64encode(const uschar * clear, int len)
f4d091fb	246	{
f3ebb786	247	uschar code = store_get(4((len+2)/3) + 1, is_tainted(clear));
f4d091fb JH	248	uschar *p = code;
	249
	250	while (len-- >0)
	251	{
286b9d5f	252	int x, y;
f4d091fb JH	253
	254	x = *clear++;
	255	*p++ = enc64table[(x >> 2) & 63];
	256
	257	if (len-- <= 0)
	258	{
	259	*p++ = enc64table[(x << 4) & 63];
	260	*p++ = '=';
	261	*p++ = '=';
	262	break;
	263	}
	264
	265	y = *clear++;
	266	*p++ = enc64table[((x << 4) \| ((y >> 4) & 15)) & 63];
	267
	268	if (len-- <= 0)
	269	{
	270	*p++ = enc64table[(y << 2) & 63];
	271	*p++ = '=';
	272	break;
	273	}
	274
	275	x = *clear++;
	276	*p++ = enc64table[((y << 2) \| ((x >> 6) & 3)) & 63];
	277
	278	*p++ = enc64table[x & 63];
	279	}
	280
	281	*p = 0;
	282
	283	return code;
	284	}
	285
	286
	287	/* End of base64.c */
	288	/* vi: sw ai sw=2
	289	*/