xref: /sqlite-3.40.0/src/tokenize.c (revision 8773b858)
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 '$' in the
80 ** middle of identifiers.  But many SQL implementations do.
81 ** SQLite will allow '$' 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 int sqlite3IsIdChar(u8 c){ return IdChar(c); }
106 
107 
108 /*
109 ** Return the length of the token that begins at z[0].
110 ** Store the token type in *tokenType before returning.
111 */
112 int sqlite3GetToken(const unsigned char *z, int *tokenType){
113   int i, c;
114   switch( *z ){
115     case ' ': case '\t': case '\n': case '\f': case '\r': {
116       testcase( z[0]==' ' );
117       testcase( z[0]=='\t' );
118       testcase( z[0]=='\n' );
119       testcase( z[0]=='\f' );
120       testcase( z[0]=='\r' );
121       for(i=1; sqlite3Isspace(z[i]); i++){}
122       *tokenType = TK_SPACE;
123       return i;
124     }
125     case '-': {
126       if( z[1]=='-' ){
127         for(i=2; (c=z[i])!=0 && c!='\n'; i++){}
128         *tokenType = TK_SPACE;   /* IMP: R-22934-25134 */
129         return i;
130       }
131       *tokenType = TK_MINUS;
132       return 1;
133     }
134     case '(': {
135       *tokenType = TK_LP;
136       return 1;
137     }
138     case ')': {
139       *tokenType = TK_RP;
140       return 1;
141     }
142     case ';': {
143       *tokenType = TK_SEMI;
144       return 1;
145     }
146     case '+': {
147       *tokenType = TK_PLUS;
148       return 1;
149     }
150     case '*': {
151       *tokenType = TK_STAR;
152       return 1;
153     }
154     case '/': {
155       if( z[1]!='*' || z[2]==0 ){
156         *tokenType = TK_SLASH;
157         return 1;
158       }
159       for(i=3, c=z[2]; (c!='*' || z[i]!='/') && (c=z[i])!=0; i++){}
160       if( c ) i++;
161       *tokenType = TK_SPACE;   /* IMP: R-22934-25134 */
162       return i;
163     }
164     case '%': {
165       *tokenType = TK_REM;
166       return 1;
167     }
168     case '=': {
169       *tokenType = TK_EQ;
170       return 1 + (z[1]=='=');
171     }
172     case '<': {
173       if( (c=z[1])=='=' ){
174         *tokenType = TK_LE;
175         return 2;
176       }else if( c=='>' ){
177         *tokenType = TK_NE;
178         return 2;
179       }else if( c=='<' ){
180         *tokenType = TK_LSHIFT;
181         return 2;
182       }else{
183         *tokenType = TK_LT;
184         return 1;
185       }
186     }
187     case '>': {
188       if( (c=z[1])=='=' ){
189         *tokenType = TK_GE;
190         return 2;
191       }else if( c=='>' ){
192         *tokenType = TK_RSHIFT;
193         return 2;
194       }else{
195         *tokenType = TK_GT;
196         return 1;
197       }
198     }
199     case '!': {
200       if( z[1]!='=' ){
201         *tokenType = TK_ILLEGAL;
202         return 2;
203       }else{
204         *tokenType = TK_NE;
205         return 2;
206       }
207     }
208     case '|': {
209       if( z[1]!='|' ){
210         *tokenType = TK_BITOR;
211         return 1;
212       }else{
213         *tokenType = TK_CONCAT;
214         return 2;
215       }
216     }
217     case ',': {
218       *tokenType = TK_COMMA;
219       return 1;
220     }
221     case '&': {
222       *tokenType = TK_BITAND;
223       return 1;
224     }
225     case '~': {
226       *tokenType = TK_BITNOT;
227       return 1;
228     }
229     case '`':
230     case '\'':
231     case '"': {
232       int delim = z[0];
233       testcase( delim=='`' );
234       testcase( delim=='\'' );
235       testcase( delim=='"' );
236       for(i=1; (c=z[i])!=0; i++){
237         if( c==delim ){
238           if( z[i+1]==delim ){
239             i++;
240           }else{
241             break;
242           }
243         }
244       }
245       if( c=='\'' ){
246         *tokenType = TK_STRING;
247         return i+1;
248       }else if( c!=0 ){
249         *tokenType = TK_ID;
250         return i+1;
251       }else{
252         *tokenType = TK_ILLEGAL;
253         return i;
254       }
255     }
256     case '.': {
257 #ifndef SQLITE_OMIT_FLOATING_POINT
258       if( !sqlite3Isdigit(z[1]) )
259 #endif
260       {
261         *tokenType = TK_DOT;
262         return 1;
263       }
264       /* If the next character is a digit, this is a floating point
265       ** number that begins with ".".  Fall thru into the next case */
266     }
267     case '0': case '1': case '2': case '3': case '4':
268     case '5': case '6': case '7': case '8': case '9': {
269       testcase( z[0]=='0' );  testcase( z[0]=='1' );  testcase( z[0]=='2' );
270       testcase( z[0]=='3' );  testcase( z[0]=='4' );  testcase( z[0]=='5' );
271       testcase( z[0]=='6' );  testcase( z[0]=='7' );  testcase( z[0]=='8' );
272       testcase( z[0]=='9' );
273       *tokenType = TK_INTEGER;
274 #ifndef SQLITE_OMIT_HEX_INTEGER
275       if( z[0]=='0' && (z[1]=='x' || z[1]=='X') && sqlite3Isxdigit(z[2]) ){
276         for(i=3; sqlite3Isxdigit(z[i]); i++){}
277         return i;
278       }
279 #endif
280       for(i=0; sqlite3Isdigit(z[i]); i++){}
281 #ifndef SQLITE_OMIT_FLOATING_POINT
282       if( z[i]=='.' ){
283         i++;
284         while( sqlite3Isdigit(z[i]) ){ i++; }
285         *tokenType = TK_FLOAT;
286       }
287       if( (z[i]=='e' || z[i]=='E') &&
288            ( sqlite3Isdigit(z[i+1])
289             || ((z[i+1]=='+' || z[i+1]=='-') && sqlite3Isdigit(z[i+2]))
290            )
291       ){
292         i += 2;
293         while( sqlite3Isdigit(z[i]) ){ i++; }
294         *tokenType = TK_FLOAT;
295       }
296 #endif
297       while( IdChar(z[i]) ){
298         *tokenType = TK_ILLEGAL;
299         i++;
300       }
301       return i;
302     }
303     case '[': {
304       for(i=1, c=z[0]; c!=']' && (c=z[i])!=0; i++){}
305       *tokenType = c==']' ? TK_ID : TK_ILLEGAL;
306       return i;
307     }
308     case '?': {
309       *tokenType = TK_VARIABLE;
310       for(i=1; sqlite3Isdigit(z[i]); i++){}
311       return i;
312     }
313 #ifndef SQLITE_OMIT_TCL_VARIABLE
314     case '$':
315 #endif
316     case '@':  /* For compatibility with MS SQL Server */
317     case '#':
318     case ':': {
319       int n = 0;
320       testcase( z[0]=='$' );  testcase( z[0]=='@' );
321       testcase( z[0]==':' );  testcase( z[0]=='#' );
322       *tokenType = TK_VARIABLE;
323       for(i=1; (c=z[i])!=0; i++){
324         if( IdChar(c) ){
325           n++;
326 #ifndef SQLITE_OMIT_TCL_VARIABLE
327         }else if( c=='(' && n>0 ){
328           do{
329             i++;
330           }while( (c=z[i])!=0 && !sqlite3Isspace(c) && c!=')' );
331           if( c==')' ){
332             i++;
333           }else{
334             *tokenType = TK_ILLEGAL;
335           }
336           break;
337         }else if( c==':' && z[i+1]==':' ){
338           i++;
339 #endif
340         }else{
341           break;
342         }
343       }
344       if( n==0 ) *tokenType = TK_ILLEGAL;
345       return i;
346     }
347 #ifndef SQLITE_OMIT_BLOB_LITERAL
348     case 'x': case 'X': {
349       testcase( z[0]=='x' ); testcase( z[0]=='X' );
350       if( z[1]=='\'' ){
351         *tokenType = TK_BLOB;
352         for(i=2; sqlite3Isxdigit(z[i]); i++){}
353         if( z[i]!='\'' || i%2 ){
354           *tokenType = TK_ILLEGAL;
355           while( z[i] && z[i]!='\'' ){ i++; }
356         }
357         if( z[i] ) i++;
358         return i;
359       }
360       /* Otherwise fall through to the next case */
361     }
362 #endif
363     default: {
364       if( !IdChar(*z) ){
365         break;
366       }
367       for(i=1; IdChar(z[i]); i++){}
368       *tokenType = keywordCode((char*)z, i);
369       return i;
370     }
371   }
372   *tokenType = TK_ILLEGAL;
373   return 1;
374 }
375 
376 /*
377 ** Run the parser on the given SQL string.  The parser structure is
378 ** passed in.  An SQLITE_ status code is returned.  If an error occurs
379 ** then an and attempt is made to write an error message into
380 ** memory obtained from sqlite3_malloc() and to make *pzErrMsg point to that
381 ** error message.
382 */
383 int sqlite3RunParser(Parse *pParse, const char *zSql, char **pzErrMsg){
384   int nErr = 0;                   /* Number of errors encountered */
385   int i;                          /* Loop counter */
386   void *pEngine;                  /* The LEMON-generated LALR(1) parser */
387   int tokenType;                  /* type of the next token */
388   int lastTokenParsed = -1;       /* type of the previous token */
389   u8 enableLookaside;             /* Saved value of db->lookaside.bEnabled */
390   sqlite3 *db = pParse->db;       /* The database connection */
391   int mxSqlLen;                   /* Max length of an SQL string */
392 
393   assert( zSql!=0 );
394   mxSqlLen = db->aLimit[SQLITE_LIMIT_SQL_LENGTH];
395   if( db->nVdbeActive==0 ){
396     db->u1.isInterrupted = 0;
397   }
398   pParse->rc = SQLITE_OK;
399   pParse->zTail = zSql;
400   i = 0;
401   assert( pzErrMsg!=0 );
402   pEngine = sqlite3ParserAlloc(sqlite3Malloc);
403   if( pEngine==0 ){
404     db->mallocFailed = 1;
405     return SQLITE_NOMEM;
406   }
407   assert( pParse->pNewTable==0 );
408   assert( pParse->pNewTrigger==0 );
409   assert( pParse->nVar==0 );
410   assert( pParse->nzVar==0 );
411   assert( pParse->azVar==0 );
412   enableLookaside = db->lookaside.bEnabled;
413   if( db->lookaside.pStart ) db->lookaside.bEnabled = 1;
414   while( !db->mallocFailed && zSql[i]!=0 ){
415     assert( i>=0 );
416     pParse->sLastToken.z = &zSql[i];
417     pParse->sLastToken.n = sqlite3GetToken((unsigned char*)&zSql[i],&tokenType);
418     i += pParse->sLastToken.n;
419     if( i>mxSqlLen ){
420       pParse->rc = SQLITE_TOOBIG;
421       break;
422     }
423     switch( tokenType ){
424       case TK_SPACE: {
425         if( db->u1.isInterrupted ){
426           sqlite3ErrorMsg(pParse, "interrupt");
427           pParse->rc = SQLITE_INTERRUPT;
428           goto abort_parse;
429         }
430         break;
431       }
432       case TK_ILLEGAL: {
433         sqlite3DbFree(db, *pzErrMsg);
434         *pzErrMsg = sqlite3MPrintf(db, "unrecognized token: \"%T\"",
435                         &pParse->sLastToken);
436         nErr++;
437         goto abort_parse;
438       }
439       case TK_SEMI: {
440         pParse->zTail = &zSql[i];
441         /* Fall thru into the default case */
442       }
443       default: {
444         sqlite3Parser(pEngine, tokenType, pParse->sLastToken, pParse);
445         lastTokenParsed = tokenType;
446         if( pParse->rc!=SQLITE_OK ){
447           goto abort_parse;
448         }
449         break;
450       }
451     }
452   }
453 abort_parse:
454   if( zSql[i]==0 && nErr==0 && pParse->rc==SQLITE_OK ){
455     if( lastTokenParsed!=TK_SEMI ){
456       sqlite3Parser(pEngine, TK_SEMI, pParse->sLastToken, pParse);
457       pParse->zTail = &zSql[i];
458     }
459     sqlite3Parser(pEngine, 0, pParse->sLastToken, pParse);
460   }
461 #ifdef YYTRACKMAXSTACKDEPTH
462   sqlite3_mutex_enter(sqlite3MallocMutex());
463   sqlite3StatusSet(SQLITE_STATUS_PARSER_STACK,
464       sqlite3ParserStackPeak(pEngine)
465   );
466   sqlite3_mutex_leave(sqlite3MallocMutex());
467 #endif /* YYDEBUG */
468   sqlite3ParserFree(pEngine, sqlite3_free);
469   db->lookaside.bEnabled = enableLookaside;
470   if( db->mallocFailed ){
471     pParse->rc = SQLITE_NOMEM;
472   }
473   if( pParse->rc!=SQLITE_OK && pParse->rc!=SQLITE_DONE && pParse->zErrMsg==0 ){
474     sqlite3SetString(&pParse->zErrMsg, db, "%s", sqlite3ErrStr(pParse->rc));
475   }
476   assert( pzErrMsg!=0 );
477   if( pParse->zErrMsg ){
478     *pzErrMsg = pParse->zErrMsg;
479     sqlite3_log(pParse->rc, "%s", *pzErrMsg);
480     pParse->zErrMsg = 0;
481     nErr++;
482   }
483   if( pParse->pVdbe && pParse->nErr>0 && pParse->nested==0 ){
484     sqlite3VdbeDelete(pParse->pVdbe);
485     pParse->pVdbe = 0;
486   }
487 #ifndef SQLITE_OMIT_SHARED_CACHE
488   if( pParse->nested==0 ){
489     sqlite3DbFree(db, pParse->aTableLock);
490     pParse->aTableLock = 0;
491     pParse->nTableLock = 0;
492   }
493 #endif
494 #ifndef SQLITE_OMIT_VIRTUALTABLE
495   sqlite3_free(pParse->apVtabLock);
496 #endif
497 
498   if( !IN_DECLARE_VTAB ){
499     /* If the pParse->declareVtab flag is set, do not delete any table
500     ** structure built up in pParse->pNewTable. The calling code (see vtab.c)
501     ** will take responsibility for freeing the Table structure.
502     */
503     sqlite3DeleteTable(db, pParse->pNewTable);
504   }
505 
506   if( pParse->bFreeWith ) sqlite3WithDelete(db, pParse->pWith);
507   sqlite3DeleteTrigger(db, pParse->pNewTrigger);
508   for(i=pParse->nzVar-1; i>=0; i--) sqlite3DbFree(db, pParse->azVar[i]);
509   sqlite3DbFree(db, pParse->azVar);
510   while( pParse->pAinc ){
511     AutoincInfo *p = pParse->pAinc;
512     pParse->pAinc = p->pNext;
513     sqlite3DbFree(db, p);
514   }
515   while( pParse->pZombieTab ){
516     Table *p = pParse->pZombieTab;
517     pParse->pZombieTab = p->pNextZombie;
518     sqlite3DeleteTable(db, p);
519   }
520   if( nErr>0 && pParse->rc==SQLITE_OK ){
521     pParse->rc = SQLITE_ERROR;
522   }
523   return nErr;
524 }
525