(_hexdigits): Must be zero-terminated,

otherwise the native Darwin linker refuses to load it.
(qp_decode): Bugfix: in accordance with rfc1892 a sequence
of whitespace is allowed in an encoded string provided that
it is followed by a printable character.
(qp_encode): Preserve trailing whitespace.

1 files changed, 137 insertions, 93 deletions

diff --git a/mailbox/filter_trans.c b/mailbox/filter_trans.c
index f8b2dc039..ab1f196f8 100644
--- a/mailbox/filter_trans.c
+++ b/mailbox/filter_trans.c
@@ -251,7 +251,7 @@ base64_init (filter_t filter)
     return ENOMEM;
 
   ts->min_size = 4;
-  ts->s_buf = calloc (4, 1);
+  ts->s_buf = calloc (ts->min_size, 1);
   if (ts->s_buf == NULL)
     {
       free (ts);
@@ -348,7 +348,10 @@ base64_encode (const char *iptr, size_t isize, char *optr, size_t osize,
 /*------------------------------------------------------
  * quoted-printable decoder/encoder
  *------------------------------------------------------*/
-static const char _hexdigits[16] = "0123456789ABCDEF";
+static const char _hexdigits[] = "0123456789ABCDEF";
+
+#define QP_LINE_MAX	76
+#define ISWS(c) ((c)==' ' || (c)=='\t')
 
 static int
 qp_init (filter_t filter)
@@ -358,8 +361,8 @@ qp_init (filter_t filter)
   if (ts == NULL)
     return ENOMEM;
 
-  ts->min_size = 4;
-  ts->s_buf = calloc (4, 1);
+  ts->min_size = QP_LINE_MAX;
+  ts->s_buf = calloc (ts->min_size, 1);
   if (ts->s_buf == NULL)
     {
       free (ts);
@@ -380,80 +383,121 @@ qp_decode (const char *iptr, size_t isize, char *optr, size_t osize,
   char c;
   int last_char = 0;
   size_t consumed = 0;
-
+  size_t wscount = 0;
+  
   *nbytes = 0;
   while (consumed < isize && *nbytes < osize)
     {
       c = *iptr++;
-      if (c == '=')
+
+      if (ISWS (c))
 	{
-	  /* There must be 2 more characters before I consume this.  */
-	  if ((consumed + 2) >= isize)
-	    return consumed;
-	  else
+	  wscount++;
+	  consumed++;
+	}
+      else
+	{
+	  /* Octets with values of 9 and 32 MAY be
+	     represented as US-ASCII TAB (HT) and SPACE characters,
+	     respectively, but MUST NOT be so represented at the end
+	     of an encoded line.  Any TAB (HT) or SPACE characters
+	     on an encoded line MUST thus be followed on that line
+	     by a printable character. */
+	  
+	  if (wscount)
 	    {
-	      /* you get =XX where XX are hex characters.  */
-	      char 	chr[3];
-	      int 	new_c;
-
-	      chr[3] = 0;
-	      chr[0] = *iptr++;
-	      /* Ignore LF.  */
-	      if (chr[0] != '\n')
+	      if (c != '\r' && c != '\n')
 		{
-		  chr[1] = *iptr++;
-		  new_c = strtoul (chr, NULL, 16);
-		  *optr++ = new_c;
-		  (*nbytes)++;
-		  consumed += 3;
+		  size_t sz;
+		  
+		  if (consumed >= isize)
+		    break;
+
+		  if (*nbytes + wscount > osize)
+		    sz = osize - *nbytes;
+		  else
+		    sz = wscount;
+		  memcpy (optr, iptr - wscount - 1, sz);
+		  optr += sz;
+		  (*nbytes) += sz;
+		  if (wscount > sz)
+		    {
+		      wscount -= sz;
+		      break;
+		    }
 		}
+	      wscount = 0;
+	    }
+		
+	  if (c == '=')
+	    {
+	      /* There must be 2 more characters before I consume this.  */
+	      if (consumed + 2 >= isize)
+		break;
 	      else
-		consumed += 2;
+		{
+		  /* you get =XX where XX are hex characters.  */
+		  char 	chr[3];
+		  int 	new_c;
+
+		  chr[3] = 0;
+		  chr[0] = *iptr++;
+		  /* Ignore LF.  */
+		  if (chr[0] != '\n')
+		    {
+		      chr[1] = *iptr++;
+		      new_c = strtoul (chr, NULL, 16);
+		      *optr++ = new_c;
+		      (*nbytes)++;
+		      consumed += 3;
+		    }
+		  else
+		    consumed += 2;
+		}
 	    }
-	}
-      /* CR character.  */
-      else if (c == '\r')
-	{
-	  /* There must be at least 1 more character before I consume this.  */
-	  if ((consumed + 1) >= isize )
-	    return (consumed);
-	  else
+	  /* CR character.  */
+	  else if (c == '\r')
 	    {
-	      iptr++; /* Skip the CR character.  */
-	      *optr++ = '\n';
-	      (*nbytes)++;
-	      consumed += 2;
+	      /* There must be at least 1 more character before
+		 I consume this.  */
+	      if (consumed + 1 >= isize)
+		break;
+	      else
+		{
+		  iptr++; /* Skip the CR character.  */
+		  *optr++ = '\n';
+		  (*nbytes)++;
+		  consumed += 2;
+		}
 	    }
-	}
-      else if ((c == 9) || (c == 32))
-	{
-	  if ((last_char == 9) || (last_char == 32))
-	    consumed++;
 	  else
 	    {
 	      *optr++ = c;
 	      (*nbytes)++;
 	      consumed++;
 	    }
-	}
-      else
-	{
-	  *optr++ = c;
-	  (*nbytes)++;
-	  consumed++;
-	}
+	}	  
       last_char = c;
     }
-  return consumed;
+  return consumed - wscount;
 }
 
+#define SOFTBRK() \
+      /* check if we have enough room for the soft linebreak */\
+      if (*nbytes + 2 > osize) \
+  	break;\
+      *optr++ = '=';\
+      *optr++ = '\n';\
+      (*nbytes) += 2;\
+      *line_len = 0;
+
 static int
 qp_encode (const char *iptr, size_t isize, char *optr, size_t osize,
 	   size_t *nbytes, int *line_len)
 {
-#define QP_LINE_MAX	76
-  int c;
+  unsigned int c;
   size_t consumed = 0;
+  size_t wscount = 0;
 
   *nbytes = 0;
 
@@ -464,28 +508,37 @@ qp_encode (const char *iptr, size_t isize, char *optr, size_t osize,
 
   while (consumed < isize)
     {
-      if (*line_len == QP_LINE_MAX)
+      int simple_char;
+      
+      /* candidate byte to convert */
+      c = *(unsigned char*) iptr;
+      simple_char = (c >= 32 && c <= 60)
+     	             || (c >= 62 && c <= 126)
+	             || c == '\t'
+	             || c == '\n';
+
+      if (*line_len == QP_LINE_MAX
+	  || (c == '\n' && consumed && ISWS (optr[-1]))
+	  || (!simple_char && *line_len >= (QP_LINE_MAX - 3)))
 	{
 	  /* to cut a qp line requires two bytes */
-	  if ((*nbytes) + 2 > osize) 
-	    return consumed;
+	  if (*nbytes + 2 > osize) 
+	    break;
 
 	  *optr++ = '=';
 	  *optr++ = '\n';
 	  (*nbytes) += 2;
 	  *line_len = 0;
 	}
-
-      /* candidate byte to convert */
-      c = *iptr;
-
-      if (((c >= 32) && (c <= 60))  || 
-	  ((c >= 62) && (c <= 126)) || 
-	  (c == 9))
+	  
+      if (ISWS (c))
+	wscount++;
+	
+      if (simple_char)
 	{
 	  /* a non-quoted character uses up one byte */
 	  if (*nbytes + 1 > osize) 
-	    return consumed;
+	    break;
 
 	  *optr++ = c;
 	  (*nbytes)++;
@@ -493,43 +546,34 @@ qp_encode (const char *iptr, size_t isize, char *optr, size_t osize,
 
 	  iptr++;
 	  consumed++;
+
+	  if (!ISWS (c))
+	    wscount = 0;
+	  
+	  if (c == '\n')
+	    *line_len = 0;
 	}
       else
 	{
-	  /* can we store the quoted character in the remaining of the
-	     line ? */
-	  if (*line_len >= (QP_LINE_MAX - 3))
-	    {
-	      /* check if we have enough room to store the padding */
-	      if (*nbytes + *line_len - QP_LINE_MAX + 3 > osize) 
-		return consumed;
-
-	      /* add spaces.  */
-	      while (*line_len < QP_LINE_MAX)
-		{
-		  *optr++ = ' ';
-		  (*nbytes)++;
-		  (*line_len)++;
-		}
-	    }
-	  else
-	    {
-	      /* a quoted character uses up three bytes */
-	      if ((*nbytes) + 3 > osize) 
-		return consumed;
+	  /* a quoted character uses up three bytes */
+	  if ((*nbytes) + 3 > osize) 
+	    break;
 
-	      *optr++ = '=';
-	      *optr++ = _hexdigits[(c >> 4) & 0xf];
-	      *optr++ = _hexdigits[c & 0xf];
+	  *optr++ = '=';
+	  *optr++ = _hexdigits[(c >> 4) & 0xf];
+	  *optr++ = _hexdigits[c & 0xf];
 
-	      (*nbytes)   += 3;
-	      (*line_len) += 3;
+	  (*nbytes) += 3;
+	  (*line_len) += 3;
 
-	      /* we've actuall used up one byte of input */
-	      iptr     ++;
-	      consumed ++;
-	    }
+	  /* we've actuall used up one byte of input */
+	  iptr++;
+	  consumed++;
+	  wscount = 0;
 	}
     }
-  return consumed;
+  if (wscount)
+    wscount--;
+  (*nbytes) -= wscount;
+  return consumed - wscount;
 }