xref: /sqlite-3.40.0/src/tokenize.c (revision c56fac74)
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 
106 /* Make the IdChar function accessible from ctime.c */
107 #ifndef SQLITE_OMIT_COMPILEOPTION_DIAGS
108 int sqlite3IsIdChar(u8 c){ return IdChar(c); }
109 #endif
110 
111 
112 /*
113 ** Return the length of the token that begins at z[0].
114 ** Store the token type in *tokenType before returning.
115 */
116 int sqlite3GetToken(const unsigned char *z, int *tokenType){
117   int i, c;
118   switch( *z ){
119     case ' ': case '\t': case '\n': case '\f': case '\r': {
120       testcase( z[0]==' ' );
121       testcase( z[0]=='\t' );
122       testcase( z[0]=='\n' );
123       testcase( z[0]=='\f' );
124       testcase( z[0]=='\r' );
125       for(i=1; sqlite3Isspace(z[i]); i++){}
126       *tokenType = TK_SPACE;
127       return i;
128     }
129     case '-': {
130       if( z[1]=='-' ){
131         for(i=2; (c=z[i])!=0 && c!='\n'; i++){}
132         *tokenType = TK_SPACE;   /* IMP: R-22934-25134 */
133         return i;
134       }
135       *tokenType = TK_MINUS;
136       return 1;
137     }
138     case '(': {
139       *tokenType = TK_LP;
140       return 1;
141     }
142     case ')': {
143       *tokenType = TK_RP;
144       return 1;
145     }
146     case ';': {
147       *tokenType = TK_SEMI;
148       return 1;
149     }
150     case '+': {
151       *tokenType = TK_PLUS;
152       return 1;
153     }
154     case '*': {
155       *tokenType = TK_STAR;
156       return 1;
157     }
158     case '/': {
159       if( z[1]!='*' || z[2]==0 ){
160         *tokenType = TK_SLASH;
161         return 1;
162       }
163       for(i=3, c=z[2]; (c!='*' || z[i]!='/') && (c=z[i])!=0; i++){}
164       if( c ) i++;
165       *tokenType = TK_SPACE;   /* IMP: R-22934-25134 */
166       return i;
167     }
168     case '%': {
169       *tokenType = TK_REM;
170       return 1;
171     }
172     case '=': {
173       *tokenType = TK_EQ;
174       return 1 + (z[1]=='=');
175     }
176     case '<': {
177       if( (c=z[1])=='=' ){
178         *tokenType = TK_LE;
179         return 2;
180       }else if( c=='>' ){
181         *tokenType = TK_NE;
182         return 2;
183       }else if( c=='<' ){
184         *tokenType = TK_LSHIFT;
185         return 2;
186       }else{
187         *tokenType = TK_LT;
188         return 1;
189       }
190     }
191     case '>': {
192       if( (c=z[1])=='=' ){
193         *tokenType = TK_GE;
194         return 2;
195       }else if( c=='>' ){
196         *tokenType = TK_RSHIFT;
197         return 2;
198       }else{
199         *tokenType = TK_GT;
200         return 1;
201       }
202     }
203     case '!': {
204       if( z[1]!='=' ){
205         *tokenType = TK_ILLEGAL;
206         return 2;
207       }else{
208         *tokenType = TK_NE;
209         return 2;
210       }
211     }
212     case '|': {
213       if( z[1]!='|' ){
214         *tokenType = TK_BITOR;
215         return 1;
216       }else{
217         *tokenType = TK_CONCAT;
218         return 2;
219       }
220     }
221     case ',': {
222       *tokenType = TK_COMMA;
223       return 1;
224     }
225     case '&': {
226       *tokenType = TK_BITAND;
227       return 1;
228     }
229     case '~': {
230       *tokenType = TK_BITNOT;
231       return 1;
232     }
233     case '`':
234     case '\'':
235     case '"': {
236       int delim = z[0];
237       testcase( delim=='`' );
238       testcase( delim=='\'' );
239       testcase( delim=='"' );
240       for(i=1; (c=z[i])!=0; i++){
241         if( c==delim ){
242           if( z[i+1]==delim ){
243             i++;
244           }else{
245             break;
246           }
247         }
248       }
249       if( c=='\'' ){
250         *tokenType = TK_STRING;
251         return i+1;
252       }else if( c!=0 ){
253         *tokenType = TK_ID;
254         return i+1;
255       }else{
256         *tokenType = TK_ILLEGAL;
257         return i;
258       }
259     }
260     case '.': {
261 #ifndef SQLITE_OMIT_FLOATING_POINT
262       if( !sqlite3Isdigit(z[1]) )
263 #endif
264       {
265         *tokenType = TK_DOT;
266         return 1;
267       }
268       /* If the next character is a digit, this is a floating point
269       ** number that begins with ".".  Fall thru into the next case */
270     }
271     case '0': case '1': case '2': case '3': case '4':
272     case '5': case '6': case '7': case '8': case '9': {
273       testcase( z[0]=='0' );  testcase( z[0]=='1' );  testcase( z[0]=='2' );
274       testcase( z[0]=='3' );  testcase( z[0]=='4' );  testcase( z[0]=='5' );
275       testcase( z[0]=='6' );  testcase( z[0]=='7' );  testcase( z[0]=='8' );
276       testcase( z[0]=='9' );
277       *tokenType = TK_INTEGER;
278 #ifndef SQLITE_OMIT_HEX_INTEGER
279       if( z[0]=='0' && (z[1]=='x' || z[1]=='X') && sqlite3Isxdigit(z[2]) ){
280         for(i=3; sqlite3Isxdigit(z[i]); i++){}
281         return i;
282       }
283 #endif
284       for(i=0; sqlite3Isdigit(z[i]); i++){}
285 #ifndef SQLITE_OMIT_FLOATING_POINT
286       if( z[i]=='.' ){
287         i++;
288         while( sqlite3Isdigit(z[i]) ){ i++; }
289         *tokenType = TK_FLOAT;
290       }
291       if( (z[i]=='e' || z[i]=='E') &&
292            ( sqlite3Isdigit(z[i+1])
293             || ((z[i+1]=='+' || z[i+1]=='-') && sqlite3Isdigit(z[i+2]))
294            )
295       ){
296         i += 2;
297         while( sqlite3Isdigit(z[i]) ){ i++; }
298         *tokenType = TK_FLOAT;
299       }
300 #endif
301       while( IdChar(z[i]) ){
302         *tokenType = TK_ILLEGAL;
303         i++;
304       }
305       return i;
306     }
307     case '[': {
308       for(i=1, c=z[0]; c!=']' && (c=z[i])!=0; i++){}
309       *tokenType = c==']' ? TK_ID : TK_ILLEGAL;
310       return i;
311     }
312     case '?': {
313       *tokenType = TK_VARIABLE;
314       for(i=1; sqlite3Isdigit(z[i]); i++){}
315       return i;
316     }
317 #ifndef SQLITE_OMIT_TCL_VARIABLE
318     case '$':
319 #endif
320     case '@':  /* For compatibility with MS SQL Server */
321     case '#':
322     case ':': {
323       int n = 0;
324       testcase( z[0]=='$' );  testcase( z[0]=='@' );
325       testcase( z[0]==':' );  testcase( z[0]=='#' );
326       *tokenType = TK_VARIABLE;
327       for(i=1; (c=z[i])!=0; i++){
328         if( IdChar(c) ){
329           n++;
330 #ifndef SQLITE_OMIT_TCL_VARIABLE
331         }else if( c=='(' && n>0 ){
332           do{
333             i++;
334           }while( (c=z[i])!=0 && !sqlite3Isspace(c) && c!=')' );
335           if( c==')' ){
336             i++;
337           }else{
338             *tokenType = TK_ILLEGAL;
339           }
340           break;
341         }else if( c==':' && z[i+1]==':' ){
342           i++;
343 #endif
344         }else{
345           break;
346         }
347       }
348       if( n==0 ) *tokenType = TK_ILLEGAL;
349       return i;
350     }
351 #ifndef SQLITE_OMIT_BLOB_LITERAL
352     case 'x': case 'X': {
353       testcase( z[0]=='x' ); testcase( z[0]=='X' );
354       if( z[1]=='\'' ){
355         *tokenType = TK_BLOB;
356         for(i=2; sqlite3Isxdigit(z[i]); i++){}
357         if( z[i]!='\'' || i%2 ){
358           *tokenType = TK_ILLEGAL;
359           while( z[i] && z[i]!='\'' ){ i++; }
360         }
361         if( z[i] ) i++;
362         return i;
363       }
364       /* Otherwise fall through to the next case */
365     }
366 #endif
367     default: {
368       if( !IdChar(*z) ){
369         break;
370       }
371       for(i=1; IdChar(z[i]); i++){}
372       *tokenType = keywordCode((char*)z, i);
373       return i;
374     }
375   }
376   *tokenType = TK_ILLEGAL;
377   return 1;
378 }
379 
380 /*
381 ** Run the parser on the given SQL string.  The parser structure is
382 ** passed in.  An SQLITE_ status code is returned.  If an error occurs
383 ** then an and attempt is made to write an error message into
384 ** memory obtained from sqlite3_malloc() and to make *pzErrMsg point to that
385 ** error message.
386 */
387 int sqlite3RunParser(Parse *pParse, const char *zSql, char **pzErrMsg){
388   int nErr = 0;                   /* Number of errors encountered */
389   int i;                          /* Loop counter */
390   void *pEngine;                  /* The LEMON-generated LALR(1) parser */
391   int tokenType;                  /* type of the next token */
392   int lastTokenParsed = -1;       /* type of the previous token */
393   u8 enableLookaside;             /* Saved value of db->lookaside.bEnabled */
394   sqlite3 *db = pParse->db;       /* The database connection */
395   int mxSqlLen;                   /* Max length of an SQL string */
396 
397   assert( zSql!=0 );
398   mxSqlLen = db->aLimit[SQLITE_LIMIT_SQL_LENGTH];
399   if( db->nVdbeActive==0 ){
400     db->u1.isInterrupted = 0;
401   }
402   pParse->rc = SQLITE_OK;
403   pParse->zTail = zSql;
404   i = 0;
405   assert( pzErrMsg!=0 );
406   /* sqlite3ParserTrace(stdout, "parser: "); */
407   pEngine = sqlite3ParserAlloc(sqlite3Malloc);
408   if( pEngine==0 ){
409     db->mallocFailed = 1;
410     return SQLITE_NOMEM;
411   }
412   assert( pParse->pNewTable==0 );
413   assert( pParse->pNewTrigger==0 );
414   assert( pParse->nVar==0 );
415   assert( pParse->nzVar==0 );
416   assert( pParse->azVar==0 );
417   enableLookaside = db->lookaside.bEnabled;
418   if( db->lookaside.pStart ) db->lookaside.bEnabled = 1;
419   while( !db->mallocFailed && zSql[i]!=0 ){
420     assert( i>=0 );
421     pParse->sLastToken.z = &zSql[i];
422     pParse->sLastToken.n = sqlite3GetToken((unsigned char*)&zSql[i],&tokenType);
423     i += pParse->sLastToken.n;
424     if( i>mxSqlLen ){
425       pParse->rc = SQLITE_TOOBIG;
426       break;
427     }
428     switch( tokenType ){
429       case TK_SPACE: {
430         if( db->u1.isInterrupted ){
431           sqlite3ErrorMsg(pParse, "interrupt");
432           pParse->rc = SQLITE_INTERRUPT;
433           goto abort_parse;
434         }
435         break;
436       }
437       case TK_ILLEGAL: {
438         sqlite3ErrorMsg(pParse, "unrecognized token: \"%T\"",
439                         &pParse->sLastToken);
440         goto abort_parse;
441       }
442       case TK_SEMI: {
443         pParse->zTail = &zSql[i];
444         /* Fall thru into the default case */
445       }
446       default: {
447         sqlite3Parser(pEngine, tokenType, pParse->sLastToken, pParse);
448         lastTokenParsed = tokenType;
449         if( pParse->rc!=SQLITE_OK ){
450           goto abort_parse;
451         }
452         break;
453       }
454     }
455   }
456 abort_parse:
457   assert( nErr==0 );
458   if( pParse->rc==SQLITE_OK && db->mallocFailed==0 ){
459     assert( zSql[i]==0 );
460     if( lastTokenParsed!=TK_SEMI ){
461       sqlite3Parser(pEngine, TK_SEMI, pParse->sLastToken, pParse);
462       pParse->zTail = &zSql[i];
463     }
464     if( pParse->rc==SQLITE_OK && db->mallocFailed==0 ){
465       sqlite3Parser(pEngine, 0, pParse->sLastToken, pParse);
466     }
467   }
468 #ifdef YYTRACKMAXSTACKDEPTH
469   sqlite3_mutex_enter(sqlite3MallocMutex());
470   sqlite3StatusHighwater(SQLITE_STATUS_PARSER_STACK,
471       sqlite3ParserStackPeak(pEngine)
472   );
473   sqlite3_mutex_leave(sqlite3MallocMutex());
474 #endif /* YYDEBUG */
475   sqlite3ParserFree(pEngine, sqlite3_free);
476   db->lookaside.bEnabled = enableLookaside;
477   if( db->mallocFailed ){
478     pParse->rc = SQLITE_NOMEM;
479   }
480   if( pParse->rc!=SQLITE_OK && pParse->rc!=SQLITE_DONE && pParse->zErrMsg==0 ){
481     pParse->zErrMsg = sqlite3MPrintf(db, "%s", sqlite3ErrStr(pParse->rc));
482   }
483   assert( pzErrMsg!=0 );
484   if( pParse->zErrMsg ){
485     *pzErrMsg = pParse->zErrMsg;
486     sqlite3_log(pParse->rc, "%s", *pzErrMsg);
487     pParse->zErrMsg = 0;
488     nErr++;
489   }
490   if( pParse->pVdbe && pParse->nErr>0 && pParse->nested==0 ){
491     sqlite3VdbeDelete(pParse->pVdbe);
492     pParse->pVdbe = 0;
493   }
494 #ifndef SQLITE_OMIT_SHARED_CACHE
495   if( pParse->nested==0 ){
496     sqlite3DbFree(db, pParse->aTableLock);
497     pParse->aTableLock = 0;
498     pParse->nTableLock = 0;
499   }
500 #endif
501 #ifndef SQLITE_OMIT_VIRTUALTABLE
502   sqlite3_free(pParse->apVtabLock);
503 #endif
504 
505   if( !IN_DECLARE_VTAB ){
506     /* If the pParse->declareVtab flag is set, do not delete any table
507     ** structure built up in pParse->pNewTable. The calling code (see vtab.c)
508     ** will take responsibility for freeing the Table structure.
509     */
510     sqlite3DeleteTable(db, pParse->pNewTable);
511   }
512 
513   if( pParse->bFreeWith ) sqlite3WithDelete(db, pParse->pWith);
514   sqlite3DeleteTrigger(db, pParse->pNewTrigger);
515   for(i=pParse->nzVar-1; i>=0; i--) sqlite3DbFree(db, pParse->azVar[i]);
516   sqlite3DbFree(db, pParse->azVar);
517   while( pParse->pAinc ){
518     AutoincInfo *p = pParse->pAinc;
519     pParse->pAinc = p->pNext;
520     sqlite3DbFree(db, p);
521   }
522   while( pParse->pZombieTab ){
523     Table *p = pParse->pZombieTab;
524     pParse->pZombieTab = p->pNextZombie;
525     sqlite3DeleteTable(db, p);
526   }
527   assert( nErr==0 || pParse->rc!=SQLITE_OK );
528   return nErr;
529 }
530