xref: /sqlite-3.40.0/src/tokenize.c (revision a3fdec71)
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 ** An 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 #include "sqliteInt.h"
19 #include <stdlib.h>
20 
21 /*
22 ** The charMap() macro maps alphabetic characters into their
23 ** lower-case ASCII equivalent.  On ASCII machines, this is just
24 ** an upper-to-lower case map.  On EBCDIC machines we also need
25 ** to adjust the encoding.  Only alphabetic characters and underscores
26 ** need to be translated.
27 */
28 #ifdef SQLITE_ASCII
29 # define charMap(X) sqlite3UpperToLower[(unsigned char)X]
30 #endif
31 #ifdef SQLITE_EBCDIC
32 # define charMap(X) ebcdicToAscii[(unsigned char)X]
33 const unsigned char ebcdicToAscii[] = {
34 /* 0   1   2   3   4   5   6   7   8   9   A   B   C   D   E   F */
35    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 0x */
36    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 1x */
37    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 2x */
38    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 3x */
39    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 4x */
40    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 5x */
41    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0, 95,  0,  0,  /* 6x */
42    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* 7x */
43    0, 97, 98, 99,100,101,102,103,104,105,  0,  0,  0,  0,  0,  0,  /* 8x */
44    0,106,107,108,109,110,111,112,113,114,  0,  0,  0,  0,  0,  0,  /* 9x */
45    0,  0,115,116,117,118,119,120,121,122,  0,  0,  0,  0,  0,  0,  /* Ax */
46    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* Bx */
47    0, 97, 98, 99,100,101,102,103,104,105,  0,  0,  0,  0,  0,  0,  /* Cx */
48    0,106,107,108,109,110,111,112,113,114,  0,  0,  0,  0,  0,  0,  /* Dx */
49    0,  0,115,116,117,118,119,120,121,122,  0,  0,  0,  0,  0,  0,  /* Ex */
50    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  /* Fx */
51 };
52 #endif
53 
54 /*
55 ** The sqlite3KeywordCode function looks up an identifier to determine if
56 ** it is a keyword.  If it is a keyword, the token code of that keyword is
57 ** returned.  If the input is not a keyword, TK_ID is returned.
58 **
59 ** The implementation of this routine was generated by a program,
60 ** mkkeywordhash.h, located in the tool subdirectory of the distribution.
61 ** The output of the mkkeywordhash.c program is written into a file
62 ** named keywordhash.h and then included into this source file by
63 ** the #include below.
64 */
65 #include "keywordhash.h"
66 
67 
68 /*
69 ** If X is a character that can be used in an identifier then
70 ** IdChar(X) will be true.  Otherwise it is false.
71 **
72 ** For ASCII, any character with the high-order bit set is
73 ** allowed in an identifier.  For 7-bit characters,
74 ** sqlite3IsIdChar[X] must be 1.
75 **
76 ** For EBCDIC, the rules are more complex but have the same
77 ** end result.
78 **
79 ** Ticket #1066.  the SQL standard does not allow '$' in the
80 ** middle of identfiers.  But many SQL implementations do.
81 ** SQLite will allow '$' in identifiers for compatibility.
82 ** But the feature is undocumented.
83 */
84 #ifdef SQLITE_ASCII
85 #define IdChar(C)  ((sqlite3CtypeMap[(unsigned char)C]&0x46)!=0)
86 #endif
87 #ifdef SQLITE_EBCDIC
88 const char sqlite3IsEbcdicIdChar[] = {
89 /* x0 x1 x2 x3 x4 x5 x6 x7 x8 x9 xA xB xC xD xE xF */
90     0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0,  /* 4x */
91     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0,  /* 5x */
92     0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0,  /* 6x */
93     0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,  /* 7x */
94     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0,  /* 8x */
95     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0,  /* 9x */
96     1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0,  /* Ax */
97     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /* Bx */
98     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1,  /* Cx */
99     0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1,  /* Dx */
100     0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1,  /* Ex */
101     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0,  /* Fx */
102 };
103 #define IdChar(C)  (((c=C)>=0x42 && sqlite3IsEbcdicIdChar[c-0x40]))
104 #endif
105 
106 
107 /*
108 ** Return the length of the token that begins at z[0].
109 ** Store the token type in *tokenType before returning.
110 */
111 int sqlite3GetToken(const unsigned char *z, int *tokenType){
112   int i, c;
113   switch( *z ){
114     case ' ': case '\t': case '\n': case '\f': case '\r': {
115       testcase( z[0]==' ' );
116       testcase( z[0]=='\t' );
117       testcase( z[0]=='\n' );
118       testcase( z[0]=='\f' );
119       testcase( z[0]=='\r' );
120       for(i=1; sqlite3Isspace(z[i]); i++){}
121       *tokenType = TK_SPACE;
122       return i;
123     }
124     case '-': {
125       if( z[1]=='-' ){
126         for(i=2; (c=z[i])!=0 && c!='\n'; i++){}
127         *tokenType = TK_SPACE;   /* IMP: R-22934-25134 */
128         return i;
129       }
130       *tokenType = TK_MINUS;
131       return 1;
132     }
133     case '(': {
134       *tokenType = TK_LP;
135       return 1;
136     }
137     case ')': {
138       *tokenType = TK_RP;
139       return 1;
140     }
141     case ';': {
142       *tokenType = TK_SEMI;
143       return 1;
144     }
145     case '+': {
146       *tokenType = TK_PLUS;
147       return 1;
148     }
149     case '*': {
150       *tokenType = TK_STAR;
151       return 1;
152     }
153     case '/': {
154       if( z[1]!='*' || z[2]==0 ){
155         *tokenType = TK_SLASH;
156         return 1;
157       }
158       for(i=3, c=z[2]; (c!='*' || z[i]!='/') && (c=z[i])!=0; i++){}
159       if( c ) i++;
160       *tokenType = TK_SPACE;   /* IMP: R-22934-25134 */
161       return i;
162     }
163     case '%': {
164       *tokenType = TK_REM;
165       return 1;
166     }
167     case '=': {
168       *tokenType = TK_EQ;
169       return 1 + (z[1]=='=');
170     }
171     case '<': {
172       if( (c=z[1])=='=' ){
173         *tokenType = TK_LE;
174         return 2;
175       }else if( c=='>' ){
176         *tokenType = TK_NE;
177         return 2;
178       }else if( c=='<' ){
179         *tokenType = TK_LSHIFT;
180         return 2;
181       }else{
182         *tokenType = TK_LT;
183         return 1;
184       }
185     }
186     case '>': {
187       if( (c=z[1])=='=' ){
188         *tokenType = TK_GE;
189         return 2;
190       }else if( c=='>' ){
191         *tokenType = TK_RSHIFT;
192         return 2;
193       }else{
194         *tokenType = TK_GT;
195         return 1;
196       }
197     }
198     case '!': {
199       if( z[1]!='=' ){
200         *tokenType = TK_ILLEGAL;
201         return 2;
202       }else{
203         *tokenType = TK_NE;
204         return 2;
205       }
206     }
207     case '|': {
208       if( z[1]!='|' ){
209         *tokenType = TK_BITOR;
210         return 1;
211       }else{
212         *tokenType = TK_CONCAT;
213         return 2;
214       }
215     }
216     case ',': {
217       *tokenType = TK_COMMA;
218       return 1;
219     }
220     case '&': {
221       *tokenType = TK_BITAND;
222       return 1;
223     }
224     case '~': {
225       *tokenType = TK_BITNOT;
226       return 1;
227     }
228     case '`':
229     case '\'':
230     case '"': {
231       int delim = z[0];
232       testcase( delim=='`' );
233       testcase( delim=='\'' );
234       testcase( delim=='"' );
235       for(i=1; (c=z[i])!=0; i++){
236         if( c==delim ){
237           if( z[i+1]==delim ){
238             i++;
239           }else{
240             break;
241           }
242         }
243       }
244       if( c=='\'' ){
245         *tokenType = TK_STRING;
246         return i+1;
247       }else if( c!=0 ){
248         *tokenType = TK_ID;
249         return i+1;
250       }else{
251         *tokenType = TK_ILLEGAL;
252         return i;
253       }
254     }
255     case '.': {
256 #ifndef SQLITE_OMIT_FLOATING_POINT
257       if( !sqlite3Isdigit(z[1]) )
258 #endif
259       {
260         *tokenType = TK_DOT;
261         return 1;
262       }
263       /* If the next character is a digit, this is a floating point
264       ** number that begins with ".".  Fall thru into the next case */
265     }
266     case '0': case '1': case '2': case '3': case '4':
267     case '5': case '6': case '7': case '8': case '9': {
268       testcase( z[0]=='0' );  testcase( z[0]=='1' );  testcase( z[0]=='2' );
269       testcase( z[0]=='3' );  testcase( z[0]=='4' );  testcase( z[0]=='5' );
270       testcase( z[0]=='6' );  testcase( z[0]=='7' );  testcase( z[0]=='8' );
271       testcase( z[0]=='9' );
272       *tokenType = TK_INTEGER;
273       for(i=0; sqlite3Isdigit(z[i]); i++){}
274 #ifndef SQLITE_OMIT_FLOATING_POINT
275       if( z[i]=='.' ){
276         i++;
277         while( sqlite3Isdigit(z[i]) ){ i++; }
278         *tokenType = TK_FLOAT;
279       }
280       if( (z[i]=='e' || z[i]=='E') &&
281            ( sqlite3Isdigit(z[i+1])
282             || ((z[i+1]=='+' || z[i+1]=='-') && sqlite3Isdigit(z[i+2]))
283            )
284       ){
285         i += 2;
286         while( sqlite3Isdigit(z[i]) ){ i++; }
287         *tokenType = TK_FLOAT;
288       }
289 #endif
290       while( IdChar(z[i]) ){
291         *tokenType = TK_ILLEGAL;
292         i++;
293       }
294       return i;
295     }
296     case '[': {
297       for(i=1, c=z[0]; c!=']' && (c=z[i])!=0; i++){}
298       *tokenType = c==']' ? TK_ID : TK_ILLEGAL;
299       return i;
300     }
301     case '?': {
302       *tokenType = TK_VARIABLE;
303       for(i=1; sqlite3Isdigit(z[i]); i++){}
304       return i;
305     }
306 #ifndef SQLITE_OMIT_TCL_VARIABLE
307     case '$':
308 #endif
309     case '@':  /* For compatibility with MS SQL Server */
310     case '#':
311     case ':': {
312       int n = 0;
313       testcase( z[0]=='$' );  testcase( z[0]=='@' );
314       testcase( z[0]==':' );  testcase( z[0]=='#' );
315       *tokenType = TK_VARIABLE;
316       for(i=1; (c=z[i])!=0; i++){
317         if( IdChar(c) ){
318           n++;
319 #ifndef SQLITE_OMIT_TCL_VARIABLE
320         }else if( c=='(' && n>0 ){
321           do{
322             i++;
323           }while( (c=z[i])!=0 && !sqlite3Isspace(c) && c!=')' );
324           if( c==')' ){
325             i++;
326           }else{
327             *tokenType = TK_ILLEGAL;
328           }
329           break;
330         }else if( c==':' && z[i+1]==':' ){
331           i++;
332 #endif
333         }else{
334           break;
335         }
336       }
337       if( n==0 ) *tokenType = TK_ILLEGAL;
338       return i;
339     }
340 #ifndef SQLITE_OMIT_BLOB_LITERAL
341     case 'x': case 'X': {
342       testcase( z[0]=='x' ); testcase( z[0]=='X' );
343       if( z[1]=='\'' ){
344         *tokenType = TK_BLOB;
345         for(i=2; sqlite3Isxdigit(z[i]); i++){}
346         if( z[i]!='\'' || i%2 ){
347           *tokenType = TK_ILLEGAL;
348           while( z[i] && z[i]!='\'' ){ i++; }
349         }
350         if( z[i] ) i++;
351         return i;
352       }
353       /* Otherwise fall through to the next case */
354     }
355 #endif
356     default: {
357       if( !IdChar(*z) ){
358         break;
359       }
360       for(i=1; IdChar(z[i]); i++){}
361       *tokenType = keywordCode((char*)z, i);
362       return i;
363     }
364   }
365   *tokenType = TK_ILLEGAL;
366   return 1;
367 }
368 
369 /*
370 ** Run the parser on the given SQL string.  The parser structure is
371 ** passed in.  An SQLITE_ status code is returned.  If an error occurs
372 ** then an and attempt is made to write an error message into
373 ** memory obtained from sqlite3_malloc() and to make *pzErrMsg point to that
374 ** error message.
375 */
376 int sqlite3RunParser(Parse *pParse, const char *zSql, char **pzErrMsg){
377   int nErr = 0;                   /* Number of errors encountered */
378   int i;                          /* Loop counter */
379   void *pEngine;                  /* The LEMON-generated LALR(1) parser */
380   int tokenType;                  /* type of the next token */
381   int lastTokenParsed = -1;       /* type of the previous token */
382   u8 enableLookaside;             /* Saved value of db->lookaside.bEnabled */
383   sqlite3 *db = pParse->db;       /* The database connection */
384   int mxSqlLen;                   /* Max length of an SQL string */
385 
386 
387   mxSqlLen = db->aLimit[SQLITE_LIMIT_SQL_LENGTH];
388   if( db->nVdbeActive==0 ){
389     db->u1.isInterrupted = 0;
390   }
391   pParse->rc = SQLITE_OK;
392   pParse->zTail = zSql;
393   i = 0;
394   assert( pzErrMsg!=0 );
395   pEngine = sqlite3ParserAlloc((void*(*)(size_t))sqlite3Malloc);
396   if( pEngine==0 ){
397     db->mallocFailed = 1;
398     return SQLITE_NOMEM;
399   }
400   assert( pParse->pNewTable==0 );
401   assert( pParse->pNewTrigger==0 );
402   assert( pParse->nVar==0 );
403   assert( pParse->nzVar==0 );
404   assert( pParse->azVar==0 );
405   enableLookaside = db->lookaside.bEnabled;
406   if( db->lookaside.pStart ) db->lookaside.bEnabled = 1;
407   while( !db->mallocFailed && zSql[i]!=0 ){
408     assert( i>=0 );
409     pParse->sLastToken.z = &zSql[i];
410     pParse->sLastToken.n = sqlite3GetToken((unsigned char*)&zSql[i],&tokenType);
411     i += pParse->sLastToken.n;
412     if( i>mxSqlLen ){
413       pParse->rc = SQLITE_TOOBIG;
414       break;
415     }
416     switch( tokenType ){
417       case TK_SPACE: {
418         if( db->u1.isInterrupted ){
419           sqlite3ErrorMsg(pParse, "interrupt");
420           pParse->rc = SQLITE_INTERRUPT;
421           goto abort_parse;
422         }
423         break;
424       }
425       case TK_ILLEGAL: {
426         sqlite3DbFree(db, *pzErrMsg);
427         *pzErrMsg = sqlite3MPrintf(db, "unrecognized token: \"%T\"",
428                         &pParse->sLastToken);
429         nErr++;
430         goto abort_parse;
431       }
432       case TK_SEMI: {
433         pParse->zTail = &zSql[i];
434         /* Fall thru into the default case */
435       }
436       default: {
437         sqlite3Parser(pEngine, tokenType, pParse->sLastToken, pParse);
438         lastTokenParsed = tokenType;
439         if( pParse->rc!=SQLITE_OK ){
440           goto abort_parse;
441         }
442         break;
443       }
444     }
445   }
446 abort_parse:
447   if( zSql[i]==0 && nErr==0 && pParse->rc==SQLITE_OK ){
448     if( lastTokenParsed!=TK_SEMI ){
449       sqlite3Parser(pEngine, TK_SEMI, pParse->sLastToken, pParse);
450       pParse->zTail = &zSql[i];
451     }
452     sqlite3Parser(pEngine, 0, pParse->sLastToken, pParse);
453   }
454 #ifdef YYTRACKMAXSTACKDEPTH
455   sqlite3StatusSet(SQLITE_STATUS_PARSER_STACK,
456       sqlite3ParserStackPeak(pEngine)
457   );
458 #endif /* YYDEBUG */
459   sqlite3ParserFree(pEngine, sqlite3_free);
460   db->lookaside.bEnabled = enableLookaside;
461   if( db->mallocFailed ){
462     pParse->rc = SQLITE_NOMEM;
463   }
464   if( pParse->rc!=SQLITE_OK && pParse->rc!=SQLITE_DONE && pParse->zErrMsg==0 ){
465     sqlite3SetString(&pParse->zErrMsg, db, "%s", sqlite3ErrStr(pParse->rc));
466   }
467   assert( pzErrMsg!=0 );
468   if( pParse->zErrMsg ){
469     *pzErrMsg = pParse->zErrMsg;
470     sqlite3_log(pParse->rc, "%s", *pzErrMsg);
471     pParse->zErrMsg = 0;
472     nErr++;
473   }
474   if( pParse->pVdbe && pParse->nErr>0 && pParse->nested==0 ){
475     sqlite3VdbeDelete(pParse->pVdbe);
476     pParse->pVdbe = 0;
477   }
478 #ifndef SQLITE_OMIT_SHARED_CACHE
479   if( pParse->nested==0 ){
480     sqlite3DbFree(db, pParse->aTableLock);
481     pParse->aTableLock = 0;
482     pParse->nTableLock = 0;
483   }
484 #endif
485 #ifndef SQLITE_OMIT_VIRTUALTABLE
486   sqlite3_free(pParse->apVtabLock);
487 #endif
488 
489   if( !IN_DECLARE_VTAB ){
490     /* If the pParse->declareVtab flag is set, do not delete any table
491     ** structure built up in pParse->pNewTable. The calling code (see vtab.c)
492     ** will take responsibility for freeing the Table structure.
493     */
494     sqlite3DeleteTable(db, pParse->pNewTable);
495   }
496 
497   if( pParse->bFreeWith ) sqlite3WithDelete(db, pParse->pWith);
498   sqlite3DeleteTrigger(db, pParse->pNewTrigger);
499   for(i=pParse->nzVar-1; i>=0; i--) sqlite3DbFree(db, pParse->azVar[i]);
500   sqlite3DbFree(db, pParse->azVar);
501   while( pParse->pAinc ){
502     AutoincInfo *p = pParse->pAinc;
503     pParse->pAinc = p->pNext;
504     sqlite3DbFree(db, p);
505   }
506   while( pParse->pZombieTab ){
507     Table *p = pParse->pZombieTab;
508     pParse->pZombieTab = p->pNextZombie;
509     sqlite3DeleteTable(db, p);
510   }
511   if( nErr>0 && pParse->rc==SQLITE_OK ){
512     pParse->rc = SQLITE_ERROR;
513   }
514   return nErr;
515 }
516