utf.c

sqlite 嵌入式数据库的源码

  /* Set len to the maximum number of bytes required in the output buffer. */
  if( desiredEnc==SQLITE_UTF8 ){
    /* When converting from UTF-16, the maximum growth results from
    ** translating a 2-byte character to a 3-byte UTF-8 character (i.e.
    ** code-point 0xFFFC). A single byte is required for the output string
    ** nul-terminator.
    */
    len = (pMem->n/2) * 3 + 1;
  }else{
    /* When converting from UTF-8 to UTF-16 the maximum growth is caused
    ** when a 1-byte UTF-8 character is translated into a 2-byte UTF-16
    ** character. Two bytes are required in the output buffer for the
    ** nul-terminator.
    */
    len = pMem->n * 2 + 2;
  }

  /* Set zIn to point at the start of the input buffer and zTerm to point 1
  ** byte past the end.
  **
  ** Variable zOut is set to point at the output buffer. This may be space
  ** obtained from malloc(), or Mem.zShort, if it large enough and not in
  ** use, or the zShort array on the stack (see above).
  */
  zIn = pMem->z;
  zTerm = &zIn[pMem->n];
  if( len>NBFS ){
    zOut = sqliteMallocRaw(len);
    if( !zOut ) return SQLITE_NOMEM;
  }else{
    zOut = zShort;
  }
  z = zOut;

  if( pMem->enc==SQLITE_UTF8 ){
    if( desiredEnc==SQLITE_UTF16LE ){
      /* UTF-8 -> UTF-16 Little-endian */
      while( zIn<zTerm ){
        READ_UTF8(zIn, c); 
        WRITE_UTF16LE(z, c);
      }
    }else{
      assert( desiredEnc==SQLITE_UTF16BE );
      /* UTF-8 -> UTF-16 Big-endian */
      while( zIn<zTerm ){
        READ_UTF8(zIn, c); 
        WRITE_UTF16BE(z, c);
      }
    }
    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 = pMem->zShort;
    pMem->flags |= (MEM_Term|MEM_Short);
  }else{
    pMem->flags |= (MEM_Term|MEM_Dyn);
  }
  pMem->z = zOut;

translate_out:
#if defined(TRANSLATE_TRACE) && defined(SQLITE_DEBUG)
  {
    char zBuf[100];
    sqlite3VdbeMemPrettyPrint(pMem, zBuf, 100);
    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 *z, int nByte){
  int r = 0;
  const char *zTerm;
  if( nByte>=0 ){
    zTerm = &z[nByte];
  }else{
    zTerm = (const char *)(-1);
  }
  assert( z<=zTerm );
  while( *z!=0 && z<zTerm ){
    SKIP_UTF8(z);
    r++;
  }
  return r;
}

#ifndef SQLITE_OMIT_UTF16
/*
** 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){
  int c = 1;
  char const *z = zIn;
  int n = 0;
  if( SQLITE_UTF16NATIVE==SQLITE_UTF16BE ){
    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);
}

/*
** UTF-16 implementation of the substr()
*/
void sqlite3utf16Substr(
  sqlite3_context *context,
  int argc,
  sqlite3_value **argv
){
  int y, z;
  unsigned char const *zStr;
  unsigned char const *zStrEnd;
  unsigned char const *zStart;
  unsigned char const *zEnd;
  int i;

  zStr = (unsigned char const *)sqlite3_value_text16(argv[0]);
  zStrEnd = &zStr[sqlite3_value_bytes16(argv[0])];
  y = sqlite3_value_int(argv[1]);
  z = sqlite3_value_int(argv[2]);

  if( y>0 ){
    y = y-1;
    zStart = zStr;
    if( SQLITE_UTF16BE==SQLITE_UTF16NATIVE ){
      for(i=0; i<y && zStart<zStrEnd; i++) SKIP_UTF16BE(zStart);
    }else{
      for(i=0; i<y && zStart<zStrEnd; i++) SKIP_UTF16LE(zStart);
    }
  }else{
    zStart = zStrEnd;
    if( SQLITE_UTF16BE==SQLITE_UTF16NATIVE ){
      for(i=y; i<0 && zStart>zStr; i++) RSKIP_UTF16BE(zStart);
    }else{
      for(i=y; i<0 && zStart>zStr; i++) RSKIP_UTF16LE(zStart);
    }
    for(; i<0; i++) z -= 1;
  }

  zEnd = zStart;
  if( SQLITE_UTF16BE==SQLITE_UTF16NATIVE ){
    for(i=0; i<z && zEnd<zStrEnd; i++) SKIP_UTF16BE(zEnd);
  }else{
    for(i=0; i<z && zEnd<zStrEnd; i++) SKIP_UTF16LE(zEnd);
  }

  sqlite3_result_text16(context, zStart, zEnd-zStart, SQLITE_TRANSIENT);
}

#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(){
  int i;
  unsigned char zBuf[20];
  unsigned char *z;
  int n;
  int c;

  for(i=0; i<0x00110000; i++){
    z = zBuf;
    WRITE_UTF8(z, i);
    n = z-zBuf;
    z = zBuf;
    READ_UTF8(z, c);
    assert( c==i );
    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 = 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 = zBuf;
    READ_UTF16BE(z, c);
    assert( c==i );
    assert( (z-zBuf)==n );
  }
}
#endif /* SQLITE_TEST */
#endif /* SQLITE_OMIT_UTF16 */