mbctype.c

C语言库函数的原型,有用的拿去

/***
*mbctype.c - MBCS table used by the functions that test for types of char
*
*       Copyright (c) Microsoft Corporation.  All rights reserved.
*
*Purpose:
*       table used to determine the type of char
*
*******************************************************************************/

#ifdef _MBCS

#include <windows.h>
#include <sect_attribs.h>
#include <cruntime.h>
#include <dbgint.h>
#include <mbdata.h>
#include <mbctype.h>
#include <mtdll.h>
#include <stdlib.h>
#include <stdio.h>
#include <internal.h>
#include <locale.h>
#include <setlocal.h>
#include <awint.h>
#include <rterr.h>

#ifndef CRTDLL

extern "C" int __cdecl __initmbctable(void);
/*
 * Flag to ensure multibyte ctype table is only initialized once
 */
extern "C" int __mbctype_initialized;

_CRTALLOC(".CRT$XIC") static _PIFV pinit = __initmbctable;

#endif  /* CRTDLL */

#define _CHINESE_SIMP_CP    936
#define _KOREAN_WANGSUNG_CP 949
#define _CHINESE_TRAD_CP    950
#define _KOREAN_JOHAB_CP    1361

#define NUM_CHARS 257 /* -1 through 255 */

#define NUM_CTYPES 4 /* table contains 4 types of info */
#define MAX_RANGES 8 /* max number of ranges needed given languages so far */

/* character type info in ranges (pair of low/high), zeros indicate end */
typedef struct
{
    int             code_page;
    unsigned short  mbulinfo[NUM_ULINFO];
    unsigned char   rgrange[NUM_CTYPES][MAX_RANGES];
} code_page_info;

extern "C"
{
threadmbcinfo __initialmbcinfo =
{
    0,
    _CLOCALECP, /* _MB_CP_ANSI */
    0,
    0,
    { 0, 0, 0, 0, 0, 0 },
    {
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
    0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
    0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
    0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
    0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0x00 /* rest is zero */
    },
    {
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
    0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x73,
    0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7a, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x4a, 0x4b,
    0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57,
    0x58, 0x59, 0x5a, 0x00, 0x00, 0x00, 0x00, 0x00 /* rest is zero */
    }
};

/* MBCS ctype array */
unsigned char _mbctype[NUM_CHARS] =
{
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
    0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
    0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
    0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
    0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0x00 /* rest is zero */
};
unsigned char _mbcasemap[256] =
{
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
    0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x73,
    0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7a, 0x00, 0x00, 0x00, 0x00, 0x00,
    0x00, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x4a, 0x4b,
    0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57,
    0x58, 0x59, 0x5a, 0x00, 0x00, 0x00, 0x00, 0x00 /* rest is zero */
};

/* global variable to indicate current code page */
int __mbcodepage = _CLOCALECP; /* __initialmbcinfo.mbcodepage; */

/* global flag indicating if current code page is a multibyte code page */
int __ismbcodepage = 0; /*__initialmbcinfo.ismbcodepage; */

/* global variable to indicate current LCID */
int __mblcid = 0; /* __initialmbcinfo.mblcid; */

/* global variable to indicate current full-width-latin upper/lower info */
unsigned short __mbulinfo[NUM_ULINFO]; /* __initialmbcinfo.mbulinfo */

/* global pointer to the current per-thread mbc information structure. */
pthreadmbcinfo __ptmbcinfo = &__initialmbcinfo;
}

static int fSystemSet;

static char __rgctypeflag[NUM_CTYPES] = { _MS, _MP, _M1, _M2 };

static code_page_info __rgcode_page_info[] =
{
    {
      _KANJI_CP, /* Kanji (Japanese) Code Page */
      { 0x8260, 0x8279,   /* Full-Width Latin Upper Range 1 */
        0x8281 - 0x8260,  /* Full-Width Latin Case Difference 1 */

        0x0000, 0x0000,   /* Full-Width Latin Upper Range 2 */
        0x0000            /* Full-Width Latin Case Difference 2 */
#ifndef _WIN32
        ,
        0x8281, 0x829A,   /* Full-Width Latin Lower Range 1 */

        0x0000, 0x0000,   /* Full-Width Latin Lower Range 2 */

        0x824F, 0x8258    /* Full-Width Latin Digit Range */
#endif  /* _WIN32 */
      },
      {
        { 0xA6, 0xDF, 0,    0,    0,    0,    0, 0, }, /* Single Byte Ranges */
        { 0xA1, 0xA5, 0,    0,    0,    0,    0, 0, }, /* Punctuation Ranges */
        { 0x81, 0x9F, 0xE0, 0xFC, 0,    0,    0, 0, }, /* Lead Byte Ranges */
        { 0x40, 0x7E, 0x80, 0xFC, 0,    0,    0, 0, }, /* Trail Byte Ranges */
      }
    },
    {
      _CHINESE_SIMP_CP, /* Chinese Simplified (PRC) Code Page */
      { 0xA3C1, 0xA3DA,   /* Full-Width Latin Upper Range 1 */
        0xA3E1 - 0xA3C1,  /* Full-Width Latin Case Difference 1 */

        0x0000, 0x0000,   /* Full-Width Latin Upper Range 2 */
        0x0000            /* Full-Width Latin Case Difference 2 */
#ifndef _WIN32
        ,
        0xA3E1, 0xA3FA,   /* Full-Width Latin Lower Range 1 */

        0x0000, 0x0000,   /* Full-Width Latin Lower Range 2 */

        0xA3B0, 0xA3B9    /* Full-Width Latin Digit Range */
#endif  /* _WIN32 */
      },
      {
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Single Byte Ranges */
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Punctuation Ranges */
        { 0x81, 0xFE, 0,    0,    0,    0,    0, 0, }, /* Lead Byte Ranges */
        { 0x40, 0xFE, 0,    0,    0,    0,    0, 0, }, /* Trail Byte Ranges */
      }
    },
    {
      _KOREAN_WANGSUNG_CP, /* Wangsung (Korean) Code Page */
      { 0xA3C1, 0xA3DA,   /* Full-Width Latin Upper Range 1 */
        0xA3E1 - 0xA3C1,  /* Full-Width Latin Case Difference 1 */

        0x0000, 0x0000,   /* Full-Width Latin Upper Range 2 */
        0x0000            /* Full-Width Latin Case Difference 2 */
#ifndef _WIN32
        ,
        0xA3E1, 0xA3FA,   /* Full-Width Latin Lower Range 1 */

        0x0000, 0x0000,   /* Full-Width Latin Lower Range 2 */

        0xA3B0, 0xA3B9    /* Full-Width Latin Digit Range */
#endif  /* _WIN32 */
      },
      {
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Single Byte Ranges */
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Punctuation Ranges */
        { 0x81, 0xFE, 0,    0,    0,    0,    0, 0, }, /* Lead Byte Ranges */
        { 0x41, 0xFE, 0,    0,    0,    0,    0, 0, }, /* Trail Byte Ranges */
      }
    },
    {
      _CHINESE_TRAD_CP, /* Chinese Traditional (Taiwan) Code Page */
      { 0xA2CF, 0xA2E4,   /* Full-Width Latin Upper Range 1 */
        0xA2E9 - 0xA2CF,  /* Full-Width Latin Case Difference 1 */

        0xA2E5, 0xA2E8,   /* Full-Width Latin Upper Range 2 */
        0xA340 - 0XA2E5   /* Full-Width Latin Case Difference 2 */
#ifndef _WIN32
        ,
        0xA2E9, 0xA2FE,   /* Full-Width Latin Lower Range 1 */

        0xA340, 0xA343,   /* Full-Width Latin Lower Range 2 */

        0xA2AF, 0xA2B8    /* Full-Width Latin Digit Range */
#endif  /* _WIN32 */
      },
      {
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Single Byte Ranges */
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Punctuation Ranges */
        { 0x81, 0xFE, 0,    0,    0,    0,    0, 0, }, /* Lead Byte Ranges */
        { 0x40, 0x7E, 0xA1, 0xFE, 0,    0,    0, 0, }, /* Trail Byte Ranges */
      }
    },
    {
      _KOREAN_JOHAB_CP, /* Johab (Korean) Code Page */
      { 0xDA51, 0xDA5E,   /* Full-Width Latin Upper Range 1 */
        0xDA71 - 0xDA51,  /* Full-Width Latin Case Difference 1 */

        0xDA5F, 0xDA6A,   /* Full-Width Latin Upper Range 2 */
        0xDA91 - 0xDA5F   /* Full-Width Latin Case Difference 2 */
#ifndef _WIN32
        ,
        0xDA71, 0xDA7E,   /* Full-Width Latin Lower Range 1 */

        0xDA91, 0xDA9C,   /* Full-Width Latin Lower Range 2 */

        0xDA40, 0xDA49    /* Full-Width Latin Digit Range */
#endif  /* _WIN32 */
      },
      {
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Single Byte Ranges */
        { 0,    0,    0,    0,    0,    0,    0, 0, }, /* Punctuation Ranges */
        { 0x81, 0xD3, 0xD8, 0xDE, 0xE0, 0xF9, 0, 0, }, /* Lead Byte Ranges */
        { 0x31, 0x7E, 0x81, 0xFE, 0,    0,    0, 0, }, /* Trail Byte Ranges */
      }
    }
};

extern "C" int __cdecl _setmbcp_nolock(int, pthreadmbcinfo);

/***
*getSystemCP - Get system default CP if requested.
*
*Purpose:
*       Get system default CP if requested.
*
*Entry:
*       codepage - user requested code page/world script
*Exit:
*       requested code page
*
*Exceptions:
*
*******************************************************************************/

static int getSystemCP (int codepage)
{
    _locale_t plocinfo = NULL;
    _LocaleUpdate _loc_update(plocinfo);
    fSystemSet = 0;

    /* get system code page values if requested */

    if (codepage == _MB_CP_OEM)
    {
        fSystemSet = 1;
        return GetOEMCP();
    }

    else if (codepage == _MB_CP_ANSI)
    {
        fSystemSet = 1;
        return GetACP();
    }

    else
    if (codepage == _MB_CP_LOCALE)
    {
        fSystemSet = 1;
        return __LC_CODEPAGE(_loc_update.GetLocaleT()->locinfo);
    }

    return codepage;
}


/***
*CPtoLCID() - Code page to LCID.
*
*Purpose:
*       Some API calls want an LCID, so convert MB CP to appropriate LCID,
*       and then API converts back to ANSI CP for that LCID.
*
*Entry:
*   codepage - code page to convert
*Exit:
*       returns appropriate LCID
*
*Exceptions:
*
*******************************************************************************/

static int CPtoLCID (int codepage)
{
    switch (codepage) {
    case 932:
        return MAKELCID(MAKELANGID(LANG_JAPANESE,SUBLANG_DEFAULT),
                        SORT_DEFAULT);
    case 936:
        return MAKELCID(MAKELANGID(LANG_CHINESE,SUBLANG_CHINESE_SIMPLIFIED),
                        SORT_DEFAULT);
    case 949:
        return MAKELCID(MAKELANGID(LANG_KOREAN,SUBLANG_DEFAULT),
                        SORT_DEFAULT);
    case 950:
        return MAKELCID(MAKELANGID(LANG_CHINESE,SUBLANG_CHINESE_TRADITIONAL),
                        SORT_DEFAULT);
    }

    return 0;
}