SQLite

** This file contains the C functions that implement various SQL
** functions of SQLite.  
**
** There is only one exported symbol in this file - the function
** sqliteRegisterBuildinFunctions() found at the bottom of the file.
** All other code has file scope.
**
** $Id: func.c,v 1.161 2007/06/22 15:21:16 danielk1977 Exp $
*/
#include "sqliteInt.h"
#include <ctype.h>
/* #include <math.h> */
#include <stdlib.h>
#include <assert.h>
#include "vdbeInt.h"
#include "os.h"


/*
** Return the collating function associated with a function.
*/
static CollSeq *sqlite3GetFuncCollSeq(sqlite3_context *context){
  return context->pColl;
}

/* The correct SQL-92 behavior is for the LIKE operator to ignore
** case.  Thus  'a' LIKE 'A' would be true. */
static const struct compareInfo likeInfoNorm = { '%', '_',   0, 1 };
/* If SQLITE_CASE_SENSITIVE_LIKE is defined, then the LIKE operator
** is case sensitive causing 'a' LIKE 'A' to be false */
static const struct compareInfo likeInfoAlt = { '%', '_',   0, 0 };

/*
** Read a single UTF-8 character and return its value.
*/
u32 sqlite3ReadUtf8(const unsigned char *z){
  u32 c;
  SQLITE_READ_UTF8(z, c);
  return c;
}

/*
** Compare two UTF-8 strings for equality where the first string can
** potentially be a "glob" expression.  Return true (1) if they
** are the same and false (0) if they are different.
**
** Globbing rules:
**

*/
static int patternCompare(
  const u8 *zPattern,              /* The glob pattern */
  const u8 *zString,               /* The string to compare against the glob */
  const struct compareInfo *pInfo, /* Information about how to do the compare */
  const int esc                    /* The escape character */
){
  register int c;
  int invert;
  int seen;
  int c2;
  u8 matchOne = pInfo->matchOne;
  u8 matchAll = pInfo->matchAll;
  u8 matchSet = pInfo->matchSet;
  u8 noCase = pInfo->noCase; 
  int prevEscape = 0;     /* True if the previous character was 'escape' */

  while( (c = *zPattern)!=0 ){
    if( !prevEscape && c==matchAll ){

      while( (c=zPattern[1]) == matchAll || c == matchOne ){
        if( c==matchOne ){
          if( *zString==0 ) return 0;
          SQLITE_SKIP_UTF8(zString);
        }
        zPattern++;
      }


      if( c && esc && sqlite3ReadUtf8(&zPattern[1])==esc ){
        u8 const *zTemp = &zPattern[1];
        SQLITE_SKIP_UTF8(zTemp);

        c = *zTemp;
      }
      if( c==0 ) return 1;
      if( c==matchSet ){
        assert( esc==0 );   /* This is GLOB, not LIKE */

        while( *zString && patternCompare(&zPattern[1],zString,pInfo,esc)==0 ){
          SQLITE_SKIP_UTF8(zString);
        }
        return *zString!=0;
      }else{

        while( (c2 = *zString)!=0 ){
          if( noCase ){
            c2 = sqlite3UpperToLower[c2];
            c = sqlite3UpperToLower[c];
            while( c2 != 0 && c2 != c ){ c2 = sqlite3UpperToLower[*++zString]; }



          }else{
            while( c2 != 0 && c2 != c ){ c2 = *++zString; }

          }

          if( c2==0 ) return 0;
          if( patternCompare(&zPattern[1],zString,pInfo,esc) ) return 1;
          SQLITE_SKIP_UTF8(zString);
        }
        return 0;
      }
    }else if( !prevEscape && c==matchOne ){

      if( *zString==0 ) return 0;
      SQLITE_SKIP_UTF8(zString);
      zPattern++;

    }else if( c==matchSet ){
      int prior_c = 0;
      assert( esc==0 );    /* This only occurs for GLOB, not LIKE */
      seen = 0;
      invert = 0;
      c = sqlite3ReadUtf8(zString);
      if( c==0 ) return 0;
      c2 = *++zPattern;
      if( c2=='^' ){ invert = 1; c2 = *++zPattern; }



      if( c2==']' ){
        if( c==']' ) seen = 1;
        c2 = *++zPattern;
      }
      while( (c2 = sqlite3ReadUtf8(zPattern))!=0 && c2!=']' ){
        if( c2=='-' && zPattern[1]!=']' && zPattern[1]!=0 && prior_c>0 ){
          zPattern++;
          c2 = sqlite3ReadUtf8(zPattern);
          if( c>=prior_c && c<=c2 ) seen = 1;
          prior_c = 0;

        }else if( c==c2 ){
          seen = 1;
          prior_c = c2;
        }else{
          prior_c = c2;
        }


        SQLITE_SKIP_UTF8(zPattern);
      }
      if( c2==0 || (seen ^ invert)==0 ) return 0;
      SQLITE_SKIP_UTF8(zString);
      zPattern++;

    }else if( esc && !prevEscape && sqlite3ReadUtf8(zPattern)==esc){
      prevEscape = 1;
      SQLITE_SKIP_UTF8(zPattern);
    }else{

      if( noCase ){
        if( sqlite3UpperToLower[c] != sqlite3UpperToLower[*zString] ) return 0;
      }else{



        if( c != *zString ) return 0;
      }
      zPattern++;
      zString++;
      prevEscape = 0;
    }
  }
  return *zString==0;
}

/*

    const unsigned char *zEsc = sqlite3_value_text(argv[2]);
    if( zEsc==0 ) return;
    if( sqlite3Utf8CharLen((char*)zEsc, -1)!=1 ){
      sqlite3_result_error(context, 
          "ESCAPE expression must be a single character", -1);
      return;
    }
    escape = sqlite3ReadUtf8(zEsc);
  }
  if( zA && zB ){
    struct compareInfo *pInfo = sqlite3_user_data(context);
#ifdef SQLITE_TEST
    sqlite3_like_count++;
#endif
    

/*
** 2001 September 15
**
** The author disclaims copyright to this source code.  In place of
** a legal notice, here is a blessing:
**
**    May you do good and not evil.
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** Internal interface definitions for SQLite.
**
** @(#) $Id: sqliteInt.h,v 1.578 2007/06/26 10:38:55 danielk1977 Exp $
*/
#ifndef _SQLITEINT_H_
#define _SQLITEINT_H_
#include "sqliteLimit.h"


#if defined(SQLITE_TCL) || defined(TCLSH)

 * This global flag is set for performance testing of triggers. When it is set
 * SQLite will perform the overhead of building new and old trigger references 
 * even when no triggers exist
 */
extern int sqlite3_always_code_trigger_setup;

/*
** A lookup table used by the SQLITE_READ_UTF8 macro.  The definition
** is in utf.c.
*/
extern const unsigned char sqlite3UtfTrans1[];

/*
** Macros for reading UTF8 characters.
**
** SQLITE_READ_UTF8(x,c) reads a single UTF8 value out of x and writes
** that value into c.  The type of x must be unsigned char*.  The type
** of c must be unsigned int.
**
** SQLITE_SKIP_UTF8(x) advances x forward by one character.  The type of
** x must be unsigned char*.
**
** Notes On Invalid UTF-8:
**
**  *  These macros never allow a 7-bit character (0x00 through 0x7f) to
**     be encoded as a multi-byte character.  Any multi-byte character that
**     attempts to encode a value between 0x00 and 0x7f is rendered as 0xfffd.
**
**  *  These macros never allow a UTF16 surrogate value to be encoded.
**     If a multi-byte character attempts to encode a value between
**     0xd800 and 0xe000 then it is rendered as 0xfffd.
**
**  *  Bytes in the range of 0x80 through 0xbf which occur as the first
**     byte of a character are interpreted as single-byte characters
**     and rendered as themselves even though they are technically
**     invalid characters.
**
**  *  These routines accept an infinite number of different UTF8 encodings
**     for unicode values 0x80 and greater.  They do not change over-length
**     encodings to 0xfffd as some systems recommend.
** 
*/
#define SQLITE_READ_UTF8(zIn, c) {                     \
  c = *(zIn++);                                        \
  if( c>=0xc0 ){                                       \
    c = sqlite3UtfTrans1[c-0xc0];                      \
    while( (*zIn & 0xc0)==0x80 ){                      \
      c = (c<<6) + (0x3f & *(zIn++));                  \
    }                                                  \
    if( c<0x80                                         \
        || (c&0xFFFFF800)==0xD800                      \
        || (c&0xFFFFFFFE)==0xFFFE ){  c = 0xFFFD; }    \
  }                                                    \
}
#define SQLITE_SKIP_UTF8(zIn) {                        \
  if( (*(zIn++))>=0xc0 ){                              \
    while( (*zIn & 0xc0)==0x80 ){ zIn++; }             \
  }                                                    \
}




/*
** The SQLITE_CORRUPT_BKPT macro can be either a constant (for production
** builds) or a function call (for debugging).  If it is a function call,
** it allows the operator to set a breakpoint at the spot where database
** corruption is first detected.
*/

int sqlite3FixTriggerStep(DbFixer*, TriggerStep*);
int sqlite3AtoF(const char *z, double*);
char *sqlite3_snprintf(int,char*,const char*,...);
int sqlite3GetInt32(const char *, int*);
int sqlite3FitsIn64Bits(const char *);
int sqlite3Utf16ByteLen(const void *pData, int nChar);
int sqlite3Utf8CharLen(const char *pData, int nByte);
u32 sqlite3ReadUtf8(const unsigned char *);
int sqlite3PutVarint(unsigned char *, u64);
int sqlite3GetVarint(const unsigned char *, u64 *);
int sqlite3GetVarint32(const unsigned char *, u32 *);
int sqlite3VarintLen(u64 v);
void sqlite3IndexAffinityStr(Vdbe *, Index *);
void sqlite3TableAffinityStr(Vdbe *, Table *);
char sqlite3CompareAffinity(Expr *pExpr, char aff2);

**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This file contains routines used to translate between UTF-8, 
** UTF-16, UTF-16BE, and UTF-16LE.
**
** $Id: utf.c,v 1.51 2007/05/23 16:23:09 danielk1977 Exp $
**
** Notes on UTF-8:
**
**   Byte-0    Byte-1    Byte-2    Byte-3    Value
**  0xxxxxxx                                 00000000 00000000 0xxxxxxx
**  110yyyyy  10xxxxxx                       00000000 00000yyy yyxxxxxx
**  1110zzzz  10yyyyyy  10xxxxxx             00000000 zzzzyyyy yyxxxxxx

  0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
  0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,
  0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
  0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
  0x00, 0x01, 0x02, 0x03, 0x00, 0x01, 0x00, 0x00,
};


#define WRITE_UTF8(zOut, c) {                          \
  if( c<0x00080 ){                                     \
    *zOut++ = (c&0xFF);                                \
  }                                                    \
  else if( c<0x00800 ){                                \
    *zOut++ = 0xC0 + ((c>>6)&0x1F);                    \

  if( c>=0xD800 && c<0xE000 ){                                       \
    int c2 = ((*zIn++)<<8);                                           \
    c2 += (*zIn++);                                                   \
    c = (c2&0x03FF) + ((c&0x003F)<<10) + (((c&0x03C0)+0x0040)<<10);   \
    if( (c & 0xFFFF0000)==0 ) c = 0xFFFD;                             \
  }                                                                   \
}

















































/*
** If the TRANSLATE_TRACE macro is defined, the value of each Mem is
** printed on stderr on the way into and out of sqlite3VdbeMemTranslate().
*/ 
/* #define TRANSLATE_TRACE 1 */

    if( !zOut ) return SQLITE_NOMEM;
  }else{
    zOut = zShort;
  }
  z = zOut;

  if( pMem->enc==SQLITE_UTF8 ){
    unsigned int iExtra = 0xD800;

    if( 0==(pMem->flags&MEM_Term) && zTerm>zIn && (zTerm[-1]&0x80) ){
      /* This UTF8 string is not nul-terminated, and the last byte is
      ** not a character in the ascii range (codpoints 0..127). This
      ** means the SQLITE_READ_UTF8() macro might read past the end
      ** of the allocated buffer.
      **
      ** There are four possibilities:
      **
      **   1. The last byte is the first byte of a non-ASCII character,
      **
      **   2. The final N bytes of the input string are continuation bytes
      **      and immediately preceding them is the first byte of a 
      **      non-ASCII character.
      **
      **   3. The final N bytes of the input string are continuation bytes
      **      and immediately preceding them is a byte that encodes a 
      **      character in the ASCII range.
      **
      **   4. The entire string consists of continuation characters.
      **
      ** Cases (3) and (4) require no special handling. The SQLITE_READ_UTF8()
      ** macro will not overread the buffer in these cases.
      */
      unsigned char *zExtra = &zTerm[-1];
      while( zExtra>zIn && (zExtra[0]&0xC0)==0x80 ){
        zExtra--;
      }

      if( (zExtra[0]&0xC0)==0xC0 ){
        /* Make a copy of the last character encoding in the input string.
        ** Then make sure it is nul-terminated and use SQLITE_READ_UTF8()
        ** to decode the codepoint. Store the codepoint in variable iExtra,
        ** it will be appended to the output string later.
        */
        unsigned char *zFree = 0;
        unsigned char zBuf[16];
        int nExtra = (pMem->n+zIn-zExtra);
        zTerm = zExtra;
        if( nExtra>15 ){
          zExtra = sqliteMallocRaw(nExtra+1);
          if( !zExtra ){
            return SQLITE_NOMEM;
          }
          zFree = zExtra;
        }else{
          zExtra = zBuf;
        }
        memcpy(zExtra, zTerm, nExtra);
        zExtra[nExtra] = '\0';
        SQLITE_READ_UTF8(zExtra, iExtra);
        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 */

**
** 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++){