utf.c

sqlite-3.4.1,嵌入式数据库.是一个功能强大的开源数据库,给学习和研发以及小型公司的发展带来了全所未有的好处.

        sqliteFree(zFree);
      }
    }

    if( desiredEnc==SQLITE_UTF16LE ){
      /* UTF-8 -> UTF-16 Little-endian */
      while( zIn<zTerm ){
        SQLITE_READ_UTF8(zIn, c); 
        WRITE_UTF16LE(z, c);
      }
      if( iExtra!=0xD800 ){
        WRITE_UTF16LE(z, iExtra);
      }
    }else{
      assert( desiredEnc==SQLITE_UTF16BE );
      /* UTF-8 -> UTF-16 Big-endian */
      while( zIn<zTerm ){
        SQLITE_READ_UTF8(zIn, c); 
        WRITE_UTF16BE(z, c);
      }
      if( iExtra!=0xD800 ){
        WRITE_UTF16BE(z, iExtra);
      }
    }
    pMem->n = z - zOut;
    *z++ = 0;
  }else{
    assert( desiredEnc==SQLITE_UTF8 );
    if( pMem->enc==SQLITE_UTF16LE ){
      /* UTF-16 Little-endian -> UTF-8 */
      while( zIn<zTerm ){
        READ_UTF16LE(zIn, c); 
        WRITE_UTF8(z, c);
      }
    }else{
      /* UTF-16 Little-endian -> UTF-8 */
      while( zIn<zTerm ){
        READ_UTF16BE(zIn, c); 
        WRITE_UTF8(z, c);
      }
    }
    pMem->n = z - zOut;
  }
  *z = 0;
  assert( (pMem->n+(desiredEnc==SQLITE_UTF8?1:2))<=len );

  sqlite3VdbeMemRelease(pMem);
  pMem->flags &= ~(MEM_Static|MEM_Dyn|MEM_Ephem|MEM_Short);
  pMem->enc = desiredEnc;
  if( zOut==zShort ){
    memcpy(pMem->zShort, zOut, len);
    zOut = (u8*)pMem->zShort;
    pMem->flags |= (MEM_Term|MEM_Short);
  }else{
    pMem->flags |= (MEM_Term|MEM_Dyn);
  }
  pMem->z = (char*)zOut;

translate_out:
#if defined(TRANSLATE_TRACE) && defined(SQLITE_DEBUG)
  {
    char zBuf[100];
    sqlite3VdbeMemPrettyPrint(pMem, zBuf);
    fprintf(stderr, "OUTPUT: %s\n", zBuf);
  }
#endif
  return SQLITE_OK;
}

/*
** This routine checks for a byte-order mark at the beginning of the 
** UTF-16 string stored in *pMem. If one is present, it is removed and
** the encoding of the Mem adjusted. This routine does not do any
** byte-swapping, it just sets Mem.enc appropriately.
**
** The allocation (static, dynamic etc.) and encoding of the Mem may be
** changed by this function.
*/
int sqlite3VdbeMemHandleBom(Mem *pMem){
  int rc = SQLITE_OK;
  u8 bom = 0;

  if( pMem->n<0 || pMem->n>1 ){
    u8 b1 = *(u8 *)pMem->z;
    u8 b2 = *(((u8 *)pMem->z) + 1);
    if( b1==0xFE && b2==0xFF ){
      bom = SQLITE_UTF16BE;
    }
    if( b1==0xFF && b2==0xFE ){
      bom = SQLITE_UTF16LE;
    }
  }
  
  if( bom ){
    /* This function is called as soon as a string is stored in a Mem*,
    ** from within sqlite3VdbeMemSetStr(). At that point it is not possible
    ** for the string to be stored in Mem.zShort, or for it to be stored
    ** in dynamic memory with no destructor.
    */
    assert( !(pMem->flags&MEM_Short) );
    assert( !(pMem->flags&MEM_Dyn) || pMem->xDel );
    if( pMem->flags & MEM_Dyn ){
      void (*xDel)(void*) = pMem->xDel;
      char *z = pMem->z;
      pMem->z = 0;
      pMem->xDel = 0;
      rc = sqlite3VdbeMemSetStr(pMem, &z[2], pMem->n-2, bom, SQLITE_TRANSIENT);
      xDel(z);
    }else{
      rc = sqlite3VdbeMemSetStr(pMem, &pMem->z[2], pMem->n-2, bom, 
          SQLITE_TRANSIENT);
    }
  }
  return rc;
}
#endif /* SQLITE_OMIT_UTF16 */

/*
** pZ is a UTF-8 encoded unicode string. If nByte is less than zero,
** return the number of unicode characters in pZ up to (but not including)
** the first 0x00 byte. If nByte is not less than zero, return the
** number of unicode characters in the first nByte of pZ (or up to 
** the first 0x00, whichever comes first).
*/
int sqlite3Utf8CharLen(const char *zIn, int nByte){
  int r = 0;
  const u8 *z = (const u8*)zIn;
  const u8 *zTerm;
  if( nByte>=0 ){
    zTerm = &z[nByte];
  }else{
    zTerm = (const u8*)(-1);
  }
  assert( z<=zTerm );
  while( *z!=0 && z<zTerm ){
    SQLITE_SKIP_UTF8(z);
    r++;
  }
  return r;
}

#ifndef SQLITE_OMIT_UTF16
/*
** Convert a UTF-16 string in the native encoding into a UTF-8 string.
** Memory to hold the UTF-8 string is obtained from malloc and must be
** freed by the calling function.
**
** NULL is returned if there is an allocation error.
*/
char *sqlite3Utf16to8(const void *z, int nByte){
  Mem m;
  memset(&m, 0, sizeof(m));
  sqlite3VdbeMemSetStr(&m, z, nByte, SQLITE_UTF16NATIVE, SQLITE_STATIC);
  sqlite3VdbeChangeEncoding(&m, SQLITE_UTF8);
  assert( (m.flags & MEM_Term)!=0 || sqlite3MallocFailed() );
  assert( (m.flags & MEM_Str)!=0 || sqlite3MallocFailed() );
  return (m.flags & MEM_Dyn)!=0 ? m.z : sqliteStrDup(m.z);
}

/*
** pZ is a UTF-16 encoded unicode string. If nChar is less than zero,
** return the number of bytes up to (but not including), the first pair
** of consecutive 0x00 bytes in pZ. If nChar is not less than zero,
** then return the number of bytes in the first nChar unicode characters
** in pZ (or up until the first pair of 0x00 bytes, whichever comes first).
*/
int sqlite3Utf16ByteLen(const void *zIn, int nChar){
  unsigned int c = 1;
  char const *z = zIn;
  int n = 0;
  if( SQLITE_UTF16NATIVE==SQLITE_UTF16BE ){
    /* Using an "if (SQLITE_UTF16NATIVE==SQLITE_UTF16BE)" construct here
    ** and in other parts of this file means that at one branch will
    ** not be covered by coverage testing on any single host. But coverage
    ** will be complete if the tests are run on both a little-endian and 
    ** big-endian host. Because both the UTF16NATIVE and SQLITE_UTF16BE
    ** macros are constant at compile time the compiler can determine
    ** which branch will be followed. It is therefore assumed that no runtime
    ** penalty is paid for this "if" statement.
    */
    while( c && ((nChar<0) || n<nChar) ){
      READ_UTF16BE(z, c);
      n++;
    }
  }else{
    while( c && ((nChar<0) || n<nChar) ){
      READ_UTF16LE(z, c);
      n++;
    }
  }
  return (z-(char const *)zIn)-((c==0)?2:0);
}

#if defined(SQLITE_TEST)
/*
** Translate UTF-8 to UTF-8.
**
** This has the effect of making sure that the string is well-formed
** UTF-8.  Miscoded characters are removed.
**
** The translation is done in-place (since it is impossible for the
** correct UTF-8 encoding to be longer than a malformed encoding).
*/
int sqlite3Utf8To8(unsigned char *zIn){
  unsigned char *zOut = zIn;
  unsigned char *zStart = zIn;
  int c;

  while(1){
    SQLITE_READ_UTF8(zIn, c);
    if( c==0 ) break;
    if( c!=0xfffd ){
      WRITE_UTF8(zOut, c);
    }
  }
  *zOut = 0;
  return zOut - zStart;
}
#endif

#if defined(SQLITE_TEST)
/*
** This routine is called from the TCL test function "translate_selftest".
** It checks that the primitives for serializing and deserializing
** characters in each encoding are inverses of each other.
*/
void sqlite3UtfSelfTest(){
  unsigned int i, t;
  unsigned char zBuf[20];
  unsigned char *z;
  int n;
  unsigned int c;

  for(i=0; i<0x00110000; i++){
    z = zBuf;
    WRITE_UTF8(z, i);
    n = z-zBuf;
    z[0] = 0;
    z = zBuf;
    SQLITE_READ_UTF8(z, c);
    t = i;
    if( i>=0xD800 && i<=0xDFFF ) t = 0xFFFD;
    if( (i&0xFFFFFFFE)==0xFFFE ) t = 0xFFFD;
    assert( c==t );
    assert( (z-zBuf)==n );
  }
  for(i=0; i<0x00110000; i++){
    if( i>=0xD800 && i<0xE000 ) continue;
    z = zBuf;
    WRITE_UTF16LE(z, i);
    n = z-zBuf;
    z[0] = 0;
    z = zBuf;
    READ_UTF16LE(z, c);
    assert( c==i );
    assert( (z-zBuf)==n );
  }
  for(i=0; i<0x00110000; i++){
    if( i>=0xD800 && i<0xE000 ) continue;
    z = zBuf;
    WRITE_UTF16BE(z, i);
    n = z-zBuf;
    z[0] = 0;
    z = zBuf;
    READ_UTF16BE(z, c);
    assert( c==i );
    assert( (z-zBuf)==n );
  }
}
#endif /* SQLITE_TEST */
#endif /* SQLITE_OMIT_UTF16 */