character.h File Reference

Macros
#define	MAX_UTF8_CHAR_BYTES   6
#define	MAX_UNICODE_CHAR_BYTES   4
#define	USHORT_SIZE   (sizeof (unsigned short))
#define	UINT_SIZE   (sizeof (unsigned int))
#define	UNIT_BYTES(format)
#define	CHAR_UNITS_ASCII(c)   ((c) < 0x80)
#define	CHAR_UNITS_UTF8(c)
#define	CHAR_UNITS_UTF16(c)   ((c) < 0x110000 ? (2 - ((c) < 0x10000)) : 0)
#define	CHAR_UNITS_UTF32(c)   1
#define	CHAR_UNITS(c, format)
#define	CHAR_BYTES   CHAR_UNITS_UTF8
#define	CHAR_UNITS_AT_UTF8(p)
#define	CHAR_UNITS_AT_UTF16(p)
#define	CHAR_UNITS_AT(mt, p)
#define	CHAR_BYTES_AT   CHAR_UNITS_AT_UTF8
#define	CHAR_UNITS_BY_HEAD_UTF8(c)
#define	CHAR_UNITS_BY_HEAD_UTF16(c)    (2 - ((unsigned short) (c) < 0xD800 || (unsigned short) (c) >= 0xDC00))
#define	CHAR_UNITS_BY_HEAD(c, format)
#define	CHAR_BYTES_BY_HEAD   CHAR_UNITS_BY_HEAD_UTF8
#define	STRING_CHAR_UTF8(p)
#define	STRING_CHAR_UTF16(p)
#define	STRING_CHAR   STRING_CHAR_UTF8
#define	STRING_CHAR_ADVANCE_UTF8(p)
#define	STRING_CHAR_ADVANCE_UTF16(p)
#define	STRING_CHAR_ADVANCE   STRING_CHAR_ADVANCE_UTF8
#define	STRING_CHAR_AND_UNITS_UTF8(p, bytes)
#define	STRING_CHAR_AND_UNITS_UTF16(p, units)
#define	STRING_CHAR_AND_UNITS(p, units, format)
#define	STRING_CHAR_AND_BYTES   STRING_CHAR_AND_UNITS_UTF8
#define	CHAR_STRING_UTF8(c, p)
#define	CHAR_STRING_UTF16(c, p)
#define	CHAR_STRING   CHAR_STRING_UTF8
#define	CHAR_HEAD_P_UTF8(p)    ((*(p) & 0xC0) != 0x80)
#define	CHAR_HEAD_P_UTF16(p)
#define	CHAR_HEAD_P   CHAR_HEAD_P_UTF8
#define	TOLOWER(c)   (((c) >= 'A' && (c) <= 'Z') ? (c) + 32 : (c))
#define	TOUPPER(c)   (((c) >= 'a' && (c) <= 'z') ? (c) - 32 : (c))
#define	ISUPPER(c)   ((c) >= 'A' && (c) <= 'Z')
#define	ISALNUM(c)

Functions
void	mchar__define_prop (MSymbol key, MSymbol type, void *mdb)

Macro Definition Documentation

MAX_UTF8_CHAR_BYTES

#define MAX_UTF8_CHAR_BYTES   6

MAX_UNICODE_CHAR_BYTES

#define MAX_UNICODE_CHAR_BYTES   4

USHORT_SIZE

#define USHORT_SIZE   (sizeof (unsigned short))

UINT_SIZE

#define UINT_SIZE   (sizeof (unsigned int))

UNIT_BYTES

#define UNIT_BYTES

(

format

)

Value:

  ((format) <= MTEXT_FORMAT_UTF_8 ? 1           \
   : (format) <= MTEXT_FORMAT_UTF_16BE ? USHORT_SIZE    \
   : UINT_SIZE)

CHAR_UNITS_ASCII

#define CHAR_UNITS_ASCII

(

c

)

((c) < 0x80)

CHAR_UNITS_UTF8

#define CHAR_UNITS_UTF8

(

c

)

Value:

  ((c) < 0x80 ? 1       \
   : (c) < 0x800 ? 2        \
   : (c) < 0x10000 ? 3      \
   : (c) < 0x200000 ? 4     \
   : (c) < 0x4000000 ? 5    \
   : 6)

CHAR_UNITS_UTF16

#define CHAR_UNITS_UTF16

(

c

)

((c) < 0x110000 ? (2 - ((c) < 0x10000)) : 0)

CHAR_UNITS_UTF32

#define CHAR_UNITS_UTF32

(

c

)

1

CHAR_UNITS

#define CHAR_UNITS	(		c,
			format
	)

Value:

  ((format) <= MTEXT_FORMAT_UTF_8 ? CHAR_UNITS_UTF8 (c)     \
   : (format) <= MTEXT_FORMAT_UTF_16BE ? CHAR_UNITS_UTF16 (c)   \
   : CHAR_UNITS_UTF32 (c))

CHAR_BYTES

#define CHAR_BYTES   CHAR_UNITS_UTF8

CHAR_UNITS_AT_UTF8

#define CHAR_UNITS_AT_UTF8

(

p

)

Value:

  (!(*(p) & 0x80) ? 1       \
   : !(*(p) & 0x20) ? 2     \
   : !(*(p) & 0x10) ? 3     \
   : !(*(p) & 0x08) ? 4     \
   : !(*(p) & 0x04) ? 5     \
   : !(*(p) & 0x02) ? 6     \
   : 0)

CHAR_UNITS_AT_UTF16

#define CHAR_UNITS_AT_UTF16

(

p

)

Value:

  (2 - (*(unsigned short *) (p) < 0xD800    \
    || *(unsigned short *) (p) >= 0xDC00))

CHAR_UNITS_AT

#define CHAR_UNITS_AT	(		mt,
			p
	)

Value:

  ((mt)->format <= MTEXT_FORMAT_UTF_8 ? CHAR_UNITS_AT_UTF8 (p)      \
   : (mt)->format <= MTEXT_FORMAT_UTF_16BE ? CHAR_UNITS_AT_UTF16 (p)    \
   : 1)

CHAR_BYTES_AT

#define CHAR_BYTES_AT   CHAR_UNITS_AT_UTF8

CHAR_UNITS_BY_HEAD_UTF8

#define CHAR_UNITS_BY_HEAD_UTF8

(

c

)

Value:

  (!((c) & 0x80) ? 1            \
   : !((c) & 0x20) ? 2          \
   : !((c) & 0x10) ? 3          \
   : !((c) & 0x08) ? 4          \
   : !((c) & 0x04) ? 5          \
   : !((c) & 0x02) ? 6          \
   : 0)

CHAR_UNITS_BY_HEAD_UTF16

#define CHAR_UNITS_BY_HEAD_UTF16

(

c

)

(2 - ((unsigned short) (c) < 0xD800 || (unsigned short) (c) >= 0xDC00))

CHAR_UNITS_BY_HEAD

#define CHAR_UNITS_BY_HEAD	(		c,
			format
	)

Value:

  ((format) <= MTEXT_FORMAT_UTF_8 ? CHAR_UNITS_BY_HEAD_UTF8 (c) \
   : (format) <= MTEXT_FORMAT_UTF_16BE ? CHAR_UNITS_BY_HEAD_UTF16 (c)   \
   : 1)

CHAR_BYTES_BY_HEAD

#define CHAR_BYTES_BY_HEAD   CHAR_UNITS_BY_HEAD_UTF8

STRING_CHAR_UTF8

#define STRING_CHAR_UTF8

(

p

)

Value:

  (!((p)[0] & 0x80) ? (p)[0]                \
   : !((p)[0] & 0x20) ? ((((p)[0] & 0x1F) << 6)     \
             | ((p)[1] & 0x3F))     \
   : !((p)[0] & 0x10) ? ((((p)[0] & 0x0F) << 12)    \
             | (((p)[1] & 0x3F) << 6)   \
             | ((p)[2] & 0x3F))     \
   : !((p)[0] & 0x08) ? ((((p)[0] & 0x07) << 18)    \
             | (((p)[1] & 0x3F) << 12)  \
             | (((p)[2] & 0x3F) << 6)   \
             | ((p)[3] & 0x3F))     \
   : !((p)[0] & 0x04) ? ((((p)[0] & 0x03) << 24)    \
             | (((p)[1] & 0x3F) << 18)  \
             | (((p)[2] & 0x3F) << 12)  \
             | (((p)[3] & 0x3F) << 6)   \
             | ((p)[4] & 0x3F))     \
   : ((((p)[0] & 0x01) << 30)               \
      | (((p)[1] & 0x3F) << 24)             \
      | (((p)[2] & 0x3F) << 18)             \
      | (((p)[3] & 0x3F) << 12)             \
      | (((p)[4] & 0x3F) << 6)              \
      | ((p)[5] & 0x3F)))

STRING_CHAR_UTF16

#define STRING_CHAR_UTF16

(

p

)

Value:

  (((unsigned short) (p)[0] < 0xD800 || (unsigned short) (p)[0] >= 0xDC00) \
   ? (p)[0]                                \
   : ((((p)[0] - 0xD800) << 10) + ((p)[1] - 0xDC00) + 0x10000))

STRING_CHAR

#define STRING_CHAR   STRING_CHAR_UTF8

STRING_CHAR_ADVANCE_UTF8

#define STRING_CHAR_ADVANCE_UTF8

(

p

)

Value:

  (!(*(p) & 0x80) ? ((p)++, (p)[-1])                \
   : !(*(p) & 0x20) ? ((p) += 2, ((((p)[-2] & 0x1F) << 6)   \
                  | ((p)[-1] & 0x3F)))      \
   : !(*(p) & 0x10) ? ((p) += 3, ((((p)[-3] & 0x0F) << 12)  \
                  | (((p)[-2] & 0x3F) << 6) \
                  | ((p)[-1] & 0x3F)))      \
   : !(*(p) & 0x08) ? ((p) += 4, ((((p)[-4] & 0x07) << 18)  \
                  | (((p)[-3] & 0x3F) << 12)    \
                  | (((p)[-2] & 0x3F) << 6) \
                  | ((p)[-1] & 0x3F)))      \
   : !(*(p) & 0x04) ? ((p) += 5, ((((p)[-5] & 0x03) << 24)  \
                  | (((p)[-4] & 0x3F) << 18)    \
                  | (((p)[-3] & 0x3F) << 12)    \
                  | (((p)[-2] & 0x3F) << 6) \
                  | ((p)[-1] & 0x3F)))      \
   : ((p) += 6, ((((p)[-6] & 0x01) << 30)           \
         | (((p)[-5] & 0x3F) << 24)         \
         | (((p)[-4] & 0x3F) << 18)         \
         | (((p)[-3] & 0x3F) << 12)         \
         | (((p)[-2] & 0x3F) << 6)          \
         | ((p)[-1] & 0x3F))))

STRING_CHAR_ADVANCE_UTF16

#define STRING_CHAR_ADVANCE_UTF16

(

p

)

Value:

  (((p)[0] < 0xD800 ||  (p)[0] >= 0xDC00)   \
   ? ((p)++, (p)[-1])               \
   : ((p) += 2, ((((p)[-2] - 0xD800) << 10) + ((p)[-1] - 0xDC00) + 0x10000)))

STRING_CHAR_ADVANCE

#define STRING_CHAR_ADVANCE   STRING_CHAR_ADVANCE_UTF8

STRING_CHAR_AND_UNITS_UTF8

#define STRING_CHAR_AND_UNITS_UTF8	(		p,
			bytes
	)

Value:

  (!((p)[0] & 0x80) ? ((bytes) = 1, (p)[0])     \
   : !((p)[0] & 0x20) ? ((bytes) = 2,           \
             ((((p)[0] & 0x1F) << 6)    \
              | ((p)[1] & 0x3F)))       \
   : !((p)[0] & 0x10) ? ((bytes) = 3,           \
             ((((p)[0] & 0x0F) << 12)   \
              | (((p)[1] & 0x3F) << 6)  \
              | ((p)[2] & 0x3F)))       \
   : !((p)[0] & 0x08) ? ((bytes) = 4,           \
             ((((p)[0] & 0x07) << 18)   \
              | (((p)[1] & 0x3F) << 12) \
              | (((p)[2] & 0x3F) << 6)  \
              | ((p)[3] & 0x3F)))       \
   : !((p)[0] & 0x04) ? ((bytes) = 5,           \
             ((((p)[0] & 0x03) << 24)   \
              | (((p)[1] & 0x3F) << 18) \
              | (((p)[2] & 0x3F) << 12) \
              | (((p)[3] & 0x3F) << 6)  \
              | ((p)[4] & 0x3F)))       \
   : ((bytes) = 6,                  \
      ((((p)[0] & 0x01) << 30)              \
       | (((p)[1] & 0x3F) << 24)            \
       | (((p)[2] & 0x3F) << 18)            \
       | (((p)[3] & 0x3F) << 12)            \
       | (((p)[4] & 0x3F) << 6)             \
       | ((p)[5] & 0x3F))))

STRING_CHAR_AND_UNITS_UTF16

#define STRING_CHAR_AND_UNITS_UTF16	(		p,
			units
	)

Value:

  (((unsigned short) (p)[0] < 0xD800 || (unsigned short) (p)[0] >= 0xDC00) \
   ? ((units) = 1, (p)[0])                         \
   : ((units) = 2,                             \
      (((p)[0] - 0xD800) << 10) + ((p)[1] - 0xDC00) + 0x10000))

STRING_CHAR_AND_UNITS

#define STRING_CHAR_AND_UNITS	(		p,
			units,
			format
	)

Value:

  ((format) <= MTEXT_FORMAT_UTF_8       \
   ? STRING_CHAR_AND_UNITS_UTF8 (p, units)  \
   : (format) <= MTEXT_FORMAT_UTF_16BE      \
   ? STRING_CHAR_AND_UNITS_UTF16 (p, units) \
   : ((units) = 1, ((unsigned) (p))[0]))

STRING_CHAR_AND_BYTES

#define STRING_CHAR_AND_BYTES   STRING_CHAR_AND_UNITS_UTF8

CHAR_STRING_UTF8

#define CHAR_STRING_UTF8	(		c,
			p
	)

Value:

  ((c) < 0x80                               \
   ? ((p)[0] = (c), 1)                          \
   : (c) < 0x800 ? ((p)[0] = (0xC0 | ((c) >> 6)),           \
            (p)[1] = (0x80 | ((c) & 0x3F)),         \
            2)                          \
   : (c) < 0x10000 ? ((p)[0] = (0xE0 | ((c) >> 12)),            \
              (p)[1] = (0x80 | (((c) >> 6) & 0x3F)),        \
              (p)[2] = (0x80 | ((c) & 0x3F)),           \
              3)                        \
   : (c) < 0x200000 ? ((p)[0] = (0xF0 | ((c) >> 18)),           \
               (p)[1] = (0x80 | (((c) >> 12) & 0x3F)),      \
               (p)[2] = (0x80 | (((c) >> 6) & 0x3F)),       \
               (p)[3] = (0x80 | ((c) & 0x3F)),          \
               4)                       \
   : (c) < 0x4000000 ? ((p)[0] = 0xF8,                  \
            (p)[1] = (0x80 | ((c) >> 18)),          \
            (p)[2] = (0x80 | (((c) >> 12) & 0x3F)),     \
            (p)[3] = (0x80 | (((c) >> 6) & 0x3F)),      \
            (p)[4] = (0x80 | ((c) & 0x3F)),         \
            5)                      \
   : ((p)[0] = (0xFC | ((c) >> 30)),                    \
      (p)[1] = (0x80 | (((c) >> 24) & 0x3F)),               \
      (p)[2] = (0x80 | (((c) >> 18) & 0x3F)),               \
      (p)[3] = (0x80 | (((c) >> 12) & 0x3F)),               \
      (p)[4] = (0x80 | (((c) >> 6) & 0x3F)),                \
      (p)[5] = (0x80 | ((c) & 0x3F)),                   \
      6))

CHAR_STRING_UTF16

#define CHAR_STRING_UTF16	(		c,
			p
	)

Value:

  ((c) < 0x10000 ? (p)[0] = (c), 1          \
   : (p[0] = (((c) - 0x10000) >> 10) + 0xD800,      \
      p[1] = (((c) - 0x10000) & 0x3FF) + 0xDC00,    \
      2))

CHAR_STRING

#define CHAR_STRING   CHAR_STRING_UTF8

CHAR_HEAD_P_UTF8

#define CHAR_HEAD_P_UTF8

(

p

)

((*(p) & 0xC0) != 0x80)

CHAR_HEAD_P_UTF16

#define CHAR_HEAD_P_UTF16

(

p

)

Value:

  (*(unsigned short *) (p) < 0xDC00 \
   || *(unsigned short *) (p) >= 0xE000)

CHAR_HEAD_P

#define CHAR_HEAD_P   CHAR_HEAD_P_UTF8

TOLOWER

#define TOLOWER

(

c

)

(((c) >= 'A' && (c) <= 'Z') ? (c) + 32 : (c))

TOUPPER

#define TOUPPER

(

c

)

(((c) >= 'a' && (c) <= 'z') ? (c) - 32 : (c))

ISUPPER

#define ISUPPER

(

c

)

((c) >= 'A' && (c) <= 'Z')

ISALNUM

#define ISALNUM

(

c

)

Value:

  (((c) >= 'A' && (c) <= 'Z')       \
   || ((c) >= 'a' && (c) <= 'z')    \
   || ((c) >= '0' && (c) <= '9'))

Function Documentation

mchar__define_prop()

void mchar__define_prop	(	MSymbol	key,
		MSymbol	type,
		void *	mdb
	)