trio.c

一个C语言写的快速贝叶斯垃圾邮件过滤工具

# define CLASS_ALPHA "[:alpha:]"
# define CLASS_BLANK "[:blank:]"
# define CLASS_CNTRL "[:cntrl:]"
# define CLASS_DIGIT "[:digit:]"
# define CLASS_GRAPH "[:graph:]"
# define CLASS_LOWER "[:lower:]"
# define CLASS_PRINT "[:print:]"
# define CLASS_PUNCT "[:punct:]"
# define CLASS_SPACE "[:space:]"
# define CLASS_UPPER "[:upper:]"
# define CLASS_XDIGIT "[:xdigit:]"
#endif

/*
 * SPECIFIERS:
 *
 *
 * a  Hex-float
 * A  Hex-float
 * c  Character
 * C  Widechar character (wint_t)
 * d  Decimal
 * e  Float
 * E  Float
 * F  Float
 * F  Float
 * g  Float
 * G  Float
 * i  Integer
 * m  Error message
 * n  Count
 * o  Octal
 * p  Pointer
 * s  String
 * S  Widechar string (wchar_t *)
 * u  Unsigned
 * x  Hex
 * X  Hex
 * [] Group
 * <> User-defined
 *
 * Reserved:
 *
 * D  Binary Coded Decimal %D(length,precision) (OS/390)
 */
#define SPECIFIER_CHAR 'c'
#define SPECIFIER_STRING 's'
#define SPECIFIER_DECIMAL 'd'
#define SPECIFIER_INTEGER 'i'
#define SPECIFIER_UNSIGNED 'u'
#define SPECIFIER_OCTAL 'o'
#define SPECIFIER_HEX 'x'
#define SPECIFIER_HEX_UPPER 'X'
#if TRIO_FEATURE_FLOAT
# define SPECIFIER_FLOAT_E 'e'
# define SPECIFIER_FLOAT_E_UPPER 'E'
# define SPECIFIER_FLOAT_F 'f'
# define SPECIFIER_FLOAT_F_UPPER 'F'
# define SPECIFIER_FLOAT_G 'g'
# define SPECIFIER_FLOAT_G_UPPER 'G'
#endif
#define SPECIFIER_POINTER 'p'
#if TRIO_FEATURE_SCANF
# define SPECIFIER_GROUP '['
# define SPECIFIER_UNGROUP ']'
#endif
#define SPECIFIER_COUNT 'n'
#if TRIO_UNIX98
# define SPECIFIER_CHAR_UPPER 'C'
# define SPECIFIER_STRING_UPPER 'S'
#endif
#define SPECIFIER_HEXFLOAT 'a'
#define SPECIFIER_HEXFLOAT_UPPER 'A'
#define SPECIFIER_ERRNO 'm'
#if TRIO_FEATURE_BINARY
# define SPECIFIER_BINARY 'b'
# define SPECIFIER_BINARY_UPPER 'B'
#endif
#if TRIO_FEATURE_USER_DEFINED
# define SPECIFIER_USER_DEFINED_BEGIN '<'
# define SPECIFIER_USER_DEFINED_END '>'
# define SPECIFIER_USER_DEFINED_SEPARATOR ':'
#endif

/*
 * QUALIFIERS:
 *
 *
 * Numbers = d,i,o,u,x,X
 * Float = a,A,e,E,f,F,g,G
 * String = s
 * Char = c
 *
 *
 * 9$ Position
 *      Use the 9th parameter. 9 can be any number between 1 and
 *      the maximal argument
 *
 * 9 Width
 *      Set width to 9. 9 can be any number, but must not be postfixed
 *      by '$'
 *
 * h  Short
 *    Numbers:
 *      (unsigned) short int
 *
 * hh Short short
 *    Numbers:
 *      (unsigned) char
 *
 * l  Long
 *    Numbers:
 *      (unsigned) long int
 *    String:
 *      as the S specifier
 *    Char:
 *      as the C specifier
 *
 * ll Long Long
 *    Numbers:
 *      (unsigned) long long int
 *
 * L  Long Double
 *    Float
 *      long double
 *
 * #  Alternative
 *    Float:
 *      Decimal-point is always present
 *    String:
 *      non-printable characters are handled as \number
 *
 *    Spacing
 *
 * +  Sign
 *
 * -  Alignment
 *
 * .  Precision
 *
 * *  Parameter
 *    print: use parameter
 *    scan: no parameter (ignore)
 *
 * q  Quad
 *
 * Z  size_t
 *
 * w  Widechar
 *
 * '  Thousands/quote
 *    Numbers:
 *      Integer part grouped in thousands
 *    Binary numbers:
 *      Number grouped in nibbles (4 bits)
 *    String:
 *      Quoted string
 *
 * j  intmax_t
 * t  prtdiff_t
 * z  size_t
 *
 * !  Sticky
 * @  Parameter (for both print and scan)
 *
 * I  n-bit Integer
 *    Numbers:
 *      The following options exists
 *        I8  = 8-bit integer
 *        I16 = 16-bit integer
 *        I32 = 32-bit integer
 *        I64 = 64-bit integer
 */
#define QUALIFIER_POSITION '$'
#define QUALIFIER_SHORT 'h'
#define QUALIFIER_LONG 'l'
#define QUALIFIER_LONG_UPPER 'L'
#define QUALIFIER_ALTERNATIVE '#'
#define QUALIFIER_SPACE ' '
#define QUALIFIER_PLUS '+'
#define QUALIFIER_MINUS '-'
#define QUALIFIER_DOT '.'
#define QUALIFIER_STAR '*'
#define QUALIFIER_CIRCUMFLEX '^' /* For scanlists */
#define QUALIFIER_SIZE_T 'z'
#define QUALIFIER_PTRDIFF_T 't'
#define QUALIFIER_INTMAX_T 'j'
#define QUALIFIER_QUAD 'q'
#define QUALIFIER_SIZE_T_UPPER 'Z'
#if TRIO_MISC
# define QUALIFIER_WIDECHAR 'w'
#endif
#define QUALIFIER_FIXED_SIZE 'I'
#define QUALIFIER_QUOTE '\''
#define QUALIFIER_STICKY '!'
#define QUALIFIER_VARSIZE '&' /* This should remain undocumented */
#define QUALIFIER_ROUNDING_UPPER 'R'
#if TRIO_EXTENSION
# define QUALIFIER_PARAM '@' /* Experimental */
# define QUALIFIER_COLON ':' /* For scanlists */
# define QUALIFIER_EQUAL '=' /* For scanlists */
#endif


/*************************************************************************
 *
 * Internal Structures
 *
 *************************************************************************/

/* Parameters */
typedef struct {
  /* An indication of which entry in the data union is used */
  int type;
  /* The flags */
  trio_flags_t flags;
  /* The width qualifier */
  int width;
  /* The precision qualifier */
  int precision;
  /* The base qualifier */
  int base;
  /* The size for the variable size qualifier */
  int varsize;
  /* The marker of the end of the specifier */
  int indexAfterSpecifier;
  /* The data from the argument list */
  union {
    char *string;
#if TRIO_FEATURE_WIDECHAR
    trio_wchar_t *wstring;
#endif
    trio_pointer_t pointer;
    union {
      trio_intmax_t as_signed;
      trio_uintmax_t as_unsigned;
    } number;
#if TRIO_FEATURE_FLOAT
    double doubleNumber;
    double *doublePointer;
    trio_long_double_t longdoubleNumber;
    trio_long_double_t *longdoublePointer;
#endif
    int errorNumber;
  } data;
#if TRIO_FEATURE_USER_DEFINED
  /* For the user-defined specifier */
  char user_name[MAX_USER_NAME];
  char user_data[MAX_USER_DATA];
#endif
} trio_parameter_t;

/* Container for customized functions */
typedef struct {
  union {
    trio_outstream_t out;
    trio_instream_t in;
  } stream;
  trio_pointer_t closure;
} trio_custom_t;

/* General trio "class" */
typedef struct _trio_class_t {
  /*
   * The function to write characters to a stream.
   */
  void (*OutStream) TRIO_PROTO((struct _trio_class_t *, int));
  /*
   * The function to read characters from a stream.
   */
  void (*InStream) TRIO_PROTO((struct _trio_class_t *, int *));
  /*
   * The current location in the stream.
   */
  trio_pointer_t location;
  /*
   * The character currently being processed.
   */
  int current;
  /*
   * The number of characters that would have been written/read
   * if there had been sufficient space.
   */
  int processed;
  /*
   * The number of characters that are actually written/read.
   * Processed and committed will only differ for the *nprintf
   * and *nscanf functions.
   */
  int committed;
  /*
   * The upper limit of characters that may be written/read.
   */
  int max;
  /*
   * The last output error that was detected.
   */
  int error;
} trio_class_t;

/* References (for user-defined callbacks) */
typedef struct _trio_reference_t {
  trio_class_t *data;
  trio_parameter_t *parameter;
} trio_reference_t;

#if TRIO_FEATURE_USER_DEFINED
/* Registered entries (for user-defined callbacks) */
typedef struct _trio_userdef_t {
  struct _trio_userdef_t *next;
  trio_callback_t callback;
  char *name;
} trio_userdef_t;
#endif

/*************************************************************************
 *
 * Internal Variables
 *
 *************************************************************************/

static TRIO_CONST char rcsid[] = "@(#)$Id: trio.c,v 1.95 2005/12/26 17:15:21 breese Exp $";

#if TRIO_FEATURE_FLOAT
/*
 * Need this to workaround a parser bug in HP C/iX compiler that fails
 * to resolves macro definitions that includes type 'long double',
 * e.g: va_arg(arg_ptr, long double)
 */
# if defined(TRIO_PLATFORM_MPEIX)
static TRIO_CONST trio_long_double_t ___dummy_long_double = 0;
# endif
#endif

static TRIO_CONST char internalNullString[] = "(nil)";

#if defined(USE_LOCALE)
static struct lconv *internalLocaleValues = NULL;
#endif

/*
 * UNIX98 says "in a locale where the radix character is not defined,
 * the radix character defaults to a period (.)"
 */
#if TRIO_FEATURE_FLOAT || TRIO_FEATURE_LOCALE || defined(USE_LOCALE)
static int internalDecimalPointLength = 1;
static char internalDecimalPoint = '.';
static char internalDecimalPointString[MAX_LOCALE_SEPARATOR_LENGTH + 1] = ".";
#endif
#if TRIO_FEATURE_QUOTE || TRIO_FEATURE_LOCALE || TRIO_EXTENSION
static int internalThousandSeparatorLength = 1;
static char internalThousandSeparator[MAX_LOCALE_SEPARATOR_LENGTH + 1] = ",";
static char internalGrouping[MAX_LOCALE_GROUPS] = { (char)NO_GROUPING };
#endif

static TRIO_CONST char internalDigitsLower[] = "0123456789abcdefghijklmnopqrstuvwxyz";
static TRIO_CONST char internalDigitsUpper[] = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ";
#if TRIO_FEATURE_SCANF
static BOOLEAN_T internalDigitsUnconverted = TRUE;
static int internalDigitArray[128];
# if TRIO_EXTENSION
static BOOLEAN_T internalCollationUnconverted = TRUE;
static char internalCollationArray[MAX_CHARACTER_CLASS][MAX_CHARACTER_CLASS];
# endif
#endif

#if TRIO_FEATURE_USER_DEFINED
static TRIO_VOLATILE trio_callback_t internalEnterCriticalRegion = NULL;
static TRIO_VOLATILE trio_callback_t internalLeaveCriticalRegion = NULL;
static trio_userdef_t *internalUserDef = NULL;
#endif


/*************************************************************************
 *
 * Internal Functions
 *
 ************************************************************************/

#if defined(TRIO_EMBED_NAN)
# include "trionan.c"
#endif

#if defined(TRIO_EMBED_STRING)
# include "triostr.c"
#endif

/*************************************************************************
 * TrioIsQualifier
 *
 * Description:
 *  Remember to add all new qualifiers to this function.
 *  QUALIFIER_POSITION must not be added.
 */
TRIO_PRIVATE BOOLEAN_T
TrioIsQualifier
TRIO_ARGS1((character),
	   TRIO_CONST char character)
{
  /* QUALIFIER_POSITION is not included */
  switch (character)
    {
    case '0': case '1': case '2': case '3': case '4':
    case '5': case '6': case '7': case '8': case '9':
    case QUALIFIER_PLUS:
    case QUALIFIER_MINUS:
    case QUALIFIER_SPACE:
    case QUALIFIER_DOT:
    case QUALIFIER_STAR:
    case QUALIFIER_ALTERNATIVE:
    case QUALIFIER_SHORT:
    case QUALIFIER_LONG:
    case QUALIFIER_CIRCUMFLEX:
    case QUALIFIER_LONG_UPPER:
    case QUALIFIER_SIZE_T:
    case QUALIFIER_PTRDIFF_T:
    case QUALIFIER_INTMAX_T:
    case QUALIFIER_QUAD:
    case QUALIFIER_SIZE_T_UPPER:
#if defined(QUALIFIER_WIDECHAR)
    case QUALIFIER_WIDECHAR:
#endif
    case QUALIFIER_QUOTE:
    case QUALIFIER_STICKY:
    case QUALIFIER_VARSIZE:
#if defined(QUALIFIER_PARAM)
    case QUALIFIER_PARAM:
#endif
    case QUALIFIER_FIXED_SIZE:
    case QUALIFIER_ROUNDING_UPPER:
      return TRUE;
    default:
      return FALSE;
    }
}

/*************************************************************************
 * TrioSetLocale
 */
#if defined(USE_LOCALE)
TRIO_PRIVATE void
TrioSetLocale(TRIO_NOARGS)
{
  internalLocaleValues = (struct lconv *)localeconv();
  if (internalLocaleValues)
    {
      if ((internalLocaleValues->decimal_point) &&
	  (internalLocaleValues->decimal_point[0] != NIL))
	{
	  internalDecimalPointLength = trio_length(internalLocaleValues->decimal_point);
	  if (internalDecimalPointLength == 1)
	    {
	      internalDecimalPoint = internalLocaleValues->decimal_point[0];
	    }
	  else
	    {
	      internalDecimalPoint = NIL;
	      trio_copy_max(internalDecimalPointString,
			    sizeof(internalDecimalPointString),
			    internalLocaleValues->decimal_point);
	    }
	}
# if TRIO_EXTENSION
      if ((internalLocaleValues->thousands_sep) &&
	  (internalLocaleValues->thousands_sep[0] != NIL))
	{
	  trio_copy_max(internalThousandSeparator,
			sizeof(internalThousandSeparator),
			internalLocaleValues->thousands_sep);
	  internalThousandSeparatorLength = trio_length(internalThousandSeparator);
	}
# endif
# if TRIO_EXTENSION
      if ((internalLocaleValues->grouping) &&
	  (internalLocaleValues->grouping[0] != NIL))
	{
	  trio_copy_max(internalGrouping,
			sizeof(internalGrouping),
			internalLocaleValues->grouping);
	}
# endif
    }
}
#endif /* defined(USE_LOCALE) */

#if TRIO_FEATURE_FLOAT && TRIO_FEATURE_QUOTE
TRIO_PRIVATE int
TrioCalcThousandSeparatorLength
TRIO_ARGS1((digits),
	   int digits)
{
  int count = 0;
  int step = NO_GROUPING;
  char *groupingPointer = internalGrouping;

  while (digits > 0)
    {
      if (*groupingPointer == CHAR_MAX)
	{
	  /* Disable grouping */
	  break; /* while */
	}
      else if (*groupingPointer == 0)
	{
	  /* Repeat last group */
	  if (step == NO_GROUPING)
	    {
	      /* Error in locale */
	      break; /* while */
	    }
	}
      else
	{
	  step = *groupingPointer++;
	}
      if (digits > step)
	count += internalThousandSeparatorLength;
      digits -= step;
    }
  return count;