big5.c File Reference

#include "postgres_fe.h"
#include "mb/pg_wchar.h"

Include dependency graph for big5.c:

Go to the source code of this file.

Data Structures
struct	codes_t

Functions
static unsigned short	BinarySearchRange (const codes_t *array, int high, unsigned short code)
unsigned short	BIG5toCNS (unsigned short big5, unsigned char *lc)
unsigned short	CNStoBIG5 (unsigned short cns, unsigned char lc)

Variables
static const codes_t	big5Level1ToCnsPlane1 [25]
static const codes_t	cnsPlane1ToBig5Level1 [26]
static const codes_t	big5Level2ToCnsPlane2 [48]
static const codes_t	cnsPlane2ToBig5Level2 [49]
static const unsigned short	b1c4 [][2]
static const unsigned short	b2c3 [][2]

Function Documentation

BIG5toCNS()

unsigned short BIG5toCNS	(	unsigned short	big5,
		unsigned char *	lc
	)

Definition at line 292 of file big5.c.

{
    unsigned short cns = 0;
    int         i;
 
    if (big5 < 0xc940U)
    {
        /* level 1 */
 
        for (i = 0; i < sizeof(b1c4) / (sizeof(unsigned short) * 2); i++)
        {
            if (b1c4[i][0] == big5)
            {
                *lc = LC_CNS11643_4;
                return (b1c4[i][1] | 0x8080U);
            }
        }
 
        if (0 < (cns = BinarySearchRange(big5Level1ToCnsPlane1, 23, big5)))
            *lc = LC_CNS11643_1;
    }
    else if (big5 == 0xc94aU)
    {
        /* level 2 */
        *lc = LC_CNS11643_1;
        cns = 0x4442;
    }
    else
    {
        /* level 2 */
        for (i = 0; i < sizeof(b2c3) / (sizeof(unsigned short) * 2); i++)
        {
            if (b2c3[i][0] == big5)
            {
                *lc = LC_CNS11643_3;
                return (b2c3[i][1] | 0x8080U);
            }
        }
 
        if (0 < (cns = BinarySearchRange(big5Level2ToCnsPlane2, 46, big5)))
            *lc = LC_CNS11643_2;
    }
 
    if (0 == cns)
    {                           /* no mapping Big5 to CNS 11643-1992 */
        *lc = 0;
        return (unsigned short) '?';
    }
 
    return cns | 0x8080;
}

References b1c4, b2c3, big5Level1ToCnsPlane1, big5Level2ToCnsPlane2, BinarySearchRange(), i, LC_CNS11643_1, LC_CNS11643_2, LC_CNS11643_3, and LC_CNS11643_4.

Referenced by big52euc_tw(), and big52mic().

BinarySearchRange()

static unsigned short BinarySearchRange ( const codes_t *  array, int  high, unsigned short  code  )

static

Definition at line 207 of file big5.c.

{
    int         low,
                mid,
                distance,
                tmp;
 
    low = 0;
    mid = high >> 1;
 
    for (; low <= high; mid = (low + high) >> 1)
    {
        if ((array[mid].code <= code) && (array[mid + 1].code > code))
        {
            if (0 == array[mid].peer)
                return 0;
            if (code >= 0xa140U)
            {
                /* big5 to cns */
                tmp = ((code & 0xff00) - (array[mid].code & 0xff00)) >> 8;
                high = code & 0x00ff;
                low = array[mid].code & 0x00ff;
 
                /*
                 * NOTE: big5 high_byte: 0xa1-0xfe, low_byte: 0x40-0x7e,
                 * 0xa1-0xfe (radicals: 0x00-0x3e, 0x3f-0x9c) big5 radix is
                 * 0x9d.                     [region_low, region_high] We
                 * should remember big5 has two different regions (above).
                 * There is a bias for the distance between these regions.
                 * 0xa1 - 0x7e + bias = 1 (Distance between 0xa1 and 0x7e is
                 * 1.) bias = - 0x22.
                 */
                distance = tmp * 0x9d + high - low +
                    (high >= 0xa1 ? (low >= 0xa1 ? 0 : -0x22)
                     : (low >= 0xa1 ? +0x22 : 0));
 
                /*
                 * NOTE: we have to convert the distance into a code point.
                 * The code point's low_byte is 0x21 plus mod_0x5e. In the
                 * first, we extract the mod_0x5e of the starting code point,
                 * subtracting 0x21, and add distance to it. Then we calculate
                 * again mod_0x5e of them, and restore the final codepoint,
                 * adding 0x21.
                 */
                tmp = (array[mid].peer & 0x00ff) + distance - 0x21;
                tmp = (array[mid].peer & 0xff00) + ((tmp / 0x5e) << 8)
                    + 0x21 + tmp % 0x5e;
                return tmp;
            }
            else
            {
                /* cns to big5 */
                tmp = ((code & 0xff00) - (array[mid].code & 0xff00)) >> 8;
 
                /*
                 * NOTE: ISO charsets ranges between 0x21-0xfe (94charset).
                 * Its radix is 0x5e. But there is no distance bias like big5.
                 */
                distance = tmp * 0x5e
                    + ((int) (code & 0x00ff) - (int) (array[mid].code & 0x00ff));
 
                /*
                 * NOTE: Similar to big5 to cns conversion, we extract
                 * mod_0x9d and restore mod_0x9d into a code point.
                 */
                low = array[mid].peer & 0x00ff;
                tmp = low + distance - (low >= 0xa1 ? 0x62 : 0x40);
                low = tmp % 0x9d;
                tmp = (array[mid].peer & 0xff00) + ((tmp / 0x9d) << 8)
                    + (low > 0x3e ? 0x62 : 0x40) + low;
                return tmp;
            }
        }
        else if (array[mid].code > code)
            high = mid - 1;
        else
            low = mid + 1;
    }
 
    return 0;
}

References codes_t::code, and codes_t::peer.

Referenced by BIG5toCNS(), and CNStoBIG5().

CNStoBIG5()

unsigned short CNStoBIG5	(	unsigned short	cns,
		unsigned char	lc
	)

Definition at line 345 of file big5.c.

{
    int         i;
    unsigned int big5 = 0;
 
    cns &= 0x7f7f;
 
    switch (lc)
    {
        case LC_CNS11643_1:
            big5 = BinarySearchRange(cnsPlane1ToBig5Level1, 24, cns);
            break;
        case LC_CNS11643_2:
            big5 = BinarySearchRange(cnsPlane2ToBig5Level2, 47, cns);
            break;
        case LC_CNS11643_3:
            for (i = 0; i < sizeof(b2c3) / (sizeof(unsigned short) * 2); i++)
            {
                if (b2c3[i][1] == cns)
                    return b2c3[i][0];
            }
            break;
        case LC_CNS11643_4:
            for (i = 0; i < sizeof(b1c4) / (sizeof(unsigned short) * 2); i++)
            {
                if (b1c4[i][1] == cns)
                    return b1c4[i][0];
            }
        default:
            break;
    }
    return big5;
}

References b1c4, b2c3, BinarySearchRange(), cnsPlane1ToBig5Level1, cnsPlane2ToBig5Level2, i, LC_CNS11643_1, LC_CNS11643_2, LC_CNS11643_3, and LC_CNS11643_4.

Referenced by euc_tw2big5(), and mic2big5().

Variable Documentation

b1c4

const unsigned short b1c4[][2]

static

Initial value:

= {
    {0xC879, 0x2123},
    {0xC87B, 0x2124},
    {0xC87D, 0x212A},
    {0xC8A2, 0x2152}
}

Definition at line 189 of file big5.c.

Referenced by BIG5toCNS(), and CNStoBIG5().

b2c3

const unsigned short b2c3[][2]

static

Initial value:

= {
    {0xF9D6, 0x4337},
    {0xF9D7, 0x4F50},
    {0xF9D8, 0x444E},
    {0xF9D9, 0x504A},
    {0xF9DA, 0x2C5D},
    {0xF9DB, 0x3D7E},
    {0xF9DC, 0x4B5C}
}

Definition at line 197 of file big5.c.

Referenced by BIG5toCNS(), and CNStoBIG5().

big5Level1ToCnsPlane1

const codes_t big5Level1ToCnsPlane1[25]

static

Initial value:

= { 
    {0xA140, 0x2121},
    {0xA1F6, 0x2258},
    {0xA1F7, 0x2257},
    {0xA1F8, 0x2259},
    {0xA2AF, 0x2421},
    {0xA3C0, 0x4221},
    {0xa3e1, 0x0000},
    {0xA440, 0x4421},
    {0xACFE, 0x5753},
    {0xacff, 0x0000},
    {0xAD40, 0x5323},
    {0xAFD0, 0x5754},
    {0xBBC8, 0x6B51},
    {0xBE52, 0x6B50},
    {0xBE53, 0x6F5C},
    {0xC1AB, 0x7536},
    {0xC2CB, 0x7535},
    {0xC2CC, 0x7737},
    {0xC361, 0x782E},
    {0xC3B9, 0x7865},
    {0xC3BA, 0x7864},
    {0xC3BB, 0x7866},
    {0xC456, 0x782D},
    {0xC457, 0x7962},
    {0xc67f, 0x0000}
}

Definition at line 25 of file big5.c.

Referenced by BIG5toCNS().

big5Level2ToCnsPlane2

const codes_t big5Level2ToCnsPlane2[48]

static

Definition at line 84 of file big5.c.

Referenced by BIG5toCNS().

cnsPlane1ToBig5Level1

const codes_t cnsPlane1ToBig5Level1[26]

static

Initial value:

= { 
    {0x2121, 0xA140},
    {0x2257, 0xA1F7},
    {0x2258, 0xA1F6},
    {0x2259, 0xA1F8},
    {0x234f, 0x0000},
    {0x2421, 0xA2AF},
    {0x2571, 0x0000},
    {0x4221, 0xA3C0},
    {0x4242, 0x0000},
    {0x4421, 0xA440},
    {0x5323, 0xAD40},
    {0x5753, 0xACFE},
    {0x5754, 0xAFD0},
    {0x6B50, 0xBE52},
    {0x6B51, 0xBBC8},
    {0x6F5C, 0xBE53},
    {0x7535, 0xC2CB},
    {0x7536, 0xC1AB},
    {0x7737, 0xC2CC},
    {0x782D, 0xC456},
    {0x782E, 0xC361},
    {0x7864, 0xC3BA},
    {0x7865, 0xC3B9},
    {0x7866, 0xC3BB},
    {0x7962, 0xC457},
    {0x7d4c, 0x0000}
}

Definition at line 54 of file big5.c.

Referenced by CNStoBIG5().

cnsPlane2ToBig5Level2

const codes_t cnsPlane2ToBig5Level2[49]

static

Definition at line 136 of file big5.c.

Referenced by CNStoBIG5().