dlls/msi/tokenize.c

   1 /*
   2 ** 2001 September 15
   3 **
   4 ** The author disclaims copyright to this source code.  In place of
   5 ** a legal notice, here is a blessing:
   6 **
   7 **    May you do good and not evil.
   8 **    May you find forgiveness for yourself and forgive others.
   9 **    May you share freely, never taking more than you give.
  10 **
  11 *************************************************************************
  12 ** A tokenizer for SQL
  13 **
  14 ** This file contains C code that splits an SQL input string up into
  15 ** individual tokens and sends those tokens one-by-one over to the
  16 ** parser for analysis.
  17 */
  18
  19 #include <ctype.h>
  20 #include <stdarg.h>
  21 #include <stdlib.h>
  22
  23 #include "windef.h"
  24 #include "winbase.h"
  25 #include "query.h"
  26 #include "sql.tab.h"
  27
  28 /*
  29 ** All the keywords of the SQL language are stored as in a hash
  30 ** table composed of instances of the following structure.
  31 */
  32 struct keyword
  33 {
  34   const WCHAR *name;             /* The keyword name */
  35   unsigned int len;
  36   int tokenType;           /* The token value for this keyword */
  37 };
  38
  39 #define MAX_TOKEN_LEN 11
  40
  41 /*
  42 ** These are the keywords
  43 ** They MUST be in alphabetical order
  44 */
  45 #define X(str)  str, ARRAY_SIZE(str) - 1
  46 static const struct keyword aKeywordTable[] = {
  47   { X(L"ADD"),         TK_ADD },
  48   { X(L"ALTER"),       TK_ALTER },
  49   { X(L"AND"),         TK_AND },
  50   { X(L"BY"),          TK_BY },
  51   { X(L"CHAR"),        TK_CHAR },
  52   { X(L"CHARACTER"),   TK_CHAR },
  53   { X(L"CREATE"),      TK_CREATE },
  54   { X(L"DELETE"),      TK_DELETE },
  55   { X(L"DISTINCT"),    TK_DISTINCT },
  56   { X(L"DROP"),        TK_DROP },
  57   { X(L"FREE"),        TK_FREE },
  58   { X(L"FROM"),        TK_FROM },
  59   { X(L"HOLD"),        TK_HOLD },
  60   { X(L"INSERT"),      TK_INSERT },
  61   { X(L"INT"),         TK_INT },
  62   { X(L"INTEGER"),     TK_INT },
  63   { X(L"INTO"),        TK_INTO },
  64   { X(L"IS"),          TK_IS },
  65   { X(L"KEY"),         TK_KEY },
  66   { X(L"LIKE"),        TK_LIKE },
  67   { X(L"LOCALIZABLE"), TK_LOCALIZABLE },
  68   { X(L"LONG"),        TK_LONG },
  69   { X(L"LONGCHAR"),    TK_LONGCHAR },
  70   { X(L"NOT"),         TK_NOT },
  71   { X(L"NULL"),        TK_NULL },
  72   { X(L"OBJECT"),      TK_OBJECT },
  73   { X(L"OR"),          TK_OR },
  74   { X(L"ORDER"),       TK_ORDER },
  75   { X(L"PRIMARY"),     TK_PRIMARY },
  76   { X(L"SELECT"),      TK_SELECT },
  77   { X(L"SET"),         TK_SET },
  78   { X(L"SHORT"),       TK_SHORT },
  79   { X(L"TABLE"),       TK_TABLE },
  80   { X(L"TEMPORARY"),   TK_TEMPORARY },
  81   { X(L"UPDATE"),      TK_UPDATE },
  82   { X(L"VALUES"),      TK_VALUES },
  83   { X(L"WHERE"),       TK_WHERE },
  84 };
  85 #undef X
  86
  87 /*
  88 ** Comparison function for binary search.
  89 */
  90 static int __cdecl compKeyword(const void *m1, const void *m2){
  91   const struct keyword *k1 = m1, *k2 = m2;
  92   int ret, len = min( k1->len, k2->len );
  93
  94   if ((ret = wcsnicmp( k1->name, k2->name, len ))) return ret;
  95   if (k1->len < k2->len) return -1;
  96   else if (k1->len > k2->len) return 1;
  97   return 0;
  98 }
  99
 100 /*
 101 ** This function looks up an identifier to determine if it is a
 102 ** keyword.  If it is a keyword, the token code of that keyword is
 103 ** returned.  If the input is not a keyword, TK_ID is returned.
 104 */
 105 static int sqliteKeywordCode(const WCHAR *z, int n){
 106   struct keyword key, *r;
 107
 108   if( n>MAX_TOKEN_LEN )
 109     return TK_ID;
 110
 111   key.tokenType = 0;
 112   key.name = z;
 113   key.len = n;
 114   r = bsearch( &key, aKeywordTable, ARRAY_SIZE(aKeywordTable), sizeof(struct keyword), compKeyword );
 115   if( r )
 116     return r->tokenType;
 117   return TK_ID;
 118 }
 119
 120
 121 /*
 122 ** If X is a character that can be used in an identifier then
 123 ** isIdChar[X] will be 1.  Otherwise isIdChar[X] will be 0.
 124 **
 125 ** In this implementation, an identifier can be a string of
 126 ** alphabetic characters, digits, and "_" plus any character
 127 ** with the high-order bit set.  The latter rule means that
 128 ** any sequence of UTF-8 characters or characters taken from
 129 ** an extended ISO8859 character set can form an identifier.
 130 */
 131 static const char isIdChar[] = {
 132 /* x0 x1 x2 x3 x4 x5 x6 x7 x8 x9 xA xB xC xD xE xF */
 133     0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /* 0x */
 134     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /* 1x */
 135     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,  /* 2x */
 136     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0,  /* 3x */
 137     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* 4x */
 138     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1,  /* 5x */
 139     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* 6x */
 140     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0,  /* 7x */
 141     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* 8x */
 142     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* 9x */
 143     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* Ax */
 144     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* Bx */
 145     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* Cx */
 146     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* Dx */
 147     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* Ex */
 148     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* Fx */
 149 };
 150
 151 /*
 152 ** WCHAR safe version of isdigit()
 153 */
 154 static inline int isDigit(WCHAR c)
 155 {
 156     return c >= '0' && c <= '9';
 157 }
 158
 159 /*
 160 ** WCHAR safe version of isspace(), except '\r'
 161 */
 162 static inline int isSpace(WCHAR c)
 163 {
 164     return c == ' ' || c == '\t' || c == '\n' || c == '\f';
 165 }
 166
 167 /*
 168 ** Return the length of the token that begins at z[0].  Return
 169 ** -1 if the token is (or might be) incomplete.  Store the token
 170 ** type in *tokenType before returning.
 171 */
 172 int sqliteGetToken(const WCHAR *z, int *tokenType, int *skip){
 173   int i;
 174
 175   *skip = 0;
 176   switch( *z ){
 177     case ' ': case '\t': case '\n': case '\f':
 178       for(i=1; isSpace(z[i]); i++){}
 179       *tokenType = TK_SPACE;
 180       return i;
 181     case '-':
 182       if( z[1]==0 ) return -1;
 183       *tokenType = TK_MINUS;
 184       return 1;
 185     case '(':
 186       *tokenType = TK_LP;
 187       return 1;
 188     case ')':
 189       *tokenType = TK_RP;
 190       return 1;
 191     case '*':
 192       *tokenType = TK_STAR;
 193       return 1;
 194     case '=':
 195       *tokenType = TK_EQ;
 196       return 1;
 197     case '<':
 198       if( z[1]=='=' ){
 199         *tokenType = TK_LE;
 200         return 2;
 201       }else if( z[1]=='>' ){
 202         *tokenType = TK_NE;
 203         return 2;
 204       }else{
 205         *tokenType = TK_LT;
 206         return 1;
 207       }
 208     case '>':
 209       if( z[1]=='=' ){
 210         *tokenType = TK_GE;
 211         return 2;
 212       }else{
 213         *tokenType = TK_GT;
 214         return 1;
 215       }
 216     case '!':
 217       if( z[1]!='=' ){
 218         *tokenType = TK_ILLEGAL;
 219         return 2;
 220       }else{
 221         *tokenType = TK_NE;
 222         return 2;
 223       }
 224     case '?':
 225       *tokenType = TK_WILDCARD;
 226       return 1;
 227     case ',':
 228       *tokenType = TK_COMMA;
 229       return 1;
 230     case '`': case '\'': {
 231       int delim = z[0];
 232       for(i=1; z[i]; i++){
 233         if( z[i]==delim )
 234           break;
 235       }
 236       if( z[i] ) i++;
 237       if( delim == '`' )
 238         *tokenType = TK_ID;
 239       else
 240         *tokenType = TK_STRING;
 241       return i;
 242     }
 243     case '.':
 244       if( !isDigit(z[1]) ){
 245         *tokenType = TK_DOT;
 246         return 1;
 247       }
 248       /* Fall through */
 249     case '0': case '1': case '2': case '3': case '4':
 250     case '5': case '6': case '7': case '8': case '9':
 251       *tokenType = TK_INTEGER;
 252       for(i=1; isDigit(z[i]); i++){}
 253       return i;
 254     case '[':
 255       for(i=1; z[i] && z[i-1]!=']'; i++){}
 256       *tokenType = TK_ID;
 257       return i;
 258     default:
 259       if( !isIdChar[*z] ){
 260         break;
 261       }
 262       for(i=1; isIdChar[z[i]]; i++){}
 263       *tokenType = sqliteKeywordCode(z, i);
 264       if( *tokenType == TK_ID && z[i] == '`' ) *skip = 1;
 265       return i;
 266   }
 267   *tokenType = TK_ILLEGAL;
 268   return 1;
 269 }