sl@0: /* sl@0: ** 2001 September 15 sl@0: ** sl@0: ** The author disclaims copyright to this source code. In place of sl@0: ** a legal notice, here is a blessing: sl@0: ** sl@0: ** May you do good and not evil. sl@0: ** May you find forgiveness for yourself and forgive others. sl@0: ** May you share freely, never taking more than you give. sl@0: ** sl@0: ************************************************************************* sl@0: ** An tokenizer for SQL sl@0: ** sl@0: ** This file contains C code that implements the sqlite3_complete() API. sl@0: ** This code used to be part of the tokenizer.c source file. But by sl@0: ** separating it out, the code will be automatically omitted from sl@0: ** static links that do not use it. sl@0: ** sl@0: ** $Id: complete.c,v 1.7 2008/06/13 18:24:27 drh Exp $ sl@0: */ sl@0: #include "sqliteInt.h" sl@0: #ifndef SQLITE_OMIT_COMPLETE sl@0: sl@0: /* sl@0: ** This is defined in tokenize.c. We just have to import the definition. sl@0: */ sl@0: #ifndef SQLITE_AMALGAMATION sl@0: #ifdef SQLITE_ASCII sl@0: extern const char sqlite3IsAsciiIdChar[]; sl@0: #define IdChar(C) (((c=C)&0x80)!=0 || (c>0x1f && sqlite3IsAsciiIdChar[c-0x20])) sl@0: #endif sl@0: #ifdef SQLITE_EBCDIC sl@0: extern const char sqlite3IsEbcdicIdChar[]; sl@0: #define IdChar(C) (((c=C)>=0x42 && sqlite3IsEbcdicIdChar[c-0x40])) sl@0: #endif sl@0: #endif /* SQLITE_AMALGAMATION */ sl@0: sl@0: sl@0: /* sl@0: ** Token types used by the sqlite3_complete() routine. See the header sl@0: ** comments on that procedure for additional information. sl@0: */ sl@0: #define tkSEMI 0 sl@0: #define tkWS 1 sl@0: #define tkOTHER 2 sl@0: #define tkEXPLAIN 3 sl@0: #define tkCREATE 4 sl@0: #define tkTEMP 5 sl@0: #define tkTRIGGER 6 sl@0: #define tkEND 7 sl@0: sl@0: /* sl@0: ** Return TRUE if the given SQL string ends in a semicolon. sl@0: ** sl@0: ** Special handling is require for CREATE TRIGGER statements. sl@0: ** Whenever the CREATE TRIGGER keywords are seen, the statement sl@0: ** must end with ";END;". sl@0: ** sl@0: ** This implementation uses a state machine with 7 states: sl@0: ** sl@0: ** (0) START At the beginning or end of an SQL statement. This routine sl@0: ** returns 1 if it ends in the START state and 0 if it ends sl@0: ** in any other state. sl@0: ** sl@0: ** (1) NORMAL We are in the middle of statement which ends with a single sl@0: ** semicolon. sl@0: ** sl@0: ** (2) EXPLAIN The keyword EXPLAIN has been seen at the beginning of sl@0: ** a statement. sl@0: ** sl@0: ** (3) CREATE The keyword CREATE has been seen at the beginning of a sl@0: ** statement, possibly preceeded by EXPLAIN and/or followed by sl@0: ** TEMP or TEMPORARY sl@0: ** sl@0: ** (4) TRIGGER We are in the middle of a trigger definition that must be sl@0: ** ended by a semicolon, the keyword END, and another semicolon. sl@0: ** sl@0: ** (5) SEMI We've seen the first semicolon in the ";END;" that occurs at sl@0: ** the end of a trigger definition. sl@0: ** sl@0: ** (6) END We've seen the ";END" of the ";END;" that occurs at the end sl@0: ** of a trigger difinition. sl@0: ** sl@0: ** Transitions between states above are determined by tokens extracted sl@0: ** from the input. The following tokens are significant: sl@0: ** sl@0: ** (0) tkSEMI A semicolon. sl@0: ** (1) tkWS Whitespace sl@0: ** (2) tkOTHER Any other SQL token. sl@0: ** (3) tkEXPLAIN The "explain" keyword. sl@0: ** (4) tkCREATE The "create" keyword. sl@0: ** (5) tkTEMP The "temp" or "temporary" keyword. sl@0: ** (6) tkTRIGGER The "trigger" keyword. sl@0: ** (7) tkEND The "end" keyword. sl@0: ** sl@0: ** Whitespace never causes a state transition and is always ignored. sl@0: ** sl@0: ** If we compile with SQLITE_OMIT_TRIGGER, all of the computation needed sl@0: ** to recognize the end of a trigger can be omitted. All we have to do sl@0: ** is look for a semicolon that is not part of an string or comment. sl@0: */ sl@0: int sqlite3_complete(const char *zSql){ sl@0: u8 state = 0; /* Current state, using numbers defined in header comment */ sl@0: u8 token; /* Value of the next token */ sl@0: sl@0: #ifndef SQLITE_OMIT_TRIGGER sl@0: /* A complex statement machine used to detect the end of a CREATE TRIGGER sl@0: ** statement. This is the normal case. sl@0: */ sl@0: static const u8 trans[7][8] = { sl@0: /* Token: */ sl@0: /* State: ** SEMI WS OTHER EXPLAIN CREATE TEMP TRIGGER END */ sl@0: /* 0 START: */ { 0, 0, 1, 2, 3, 1, 1, 1, }, sl@0: /* 1 NORMAL: */ { 0, 1, 1, 1, 1, 1, 1, 1, }, sl@0: /* 2 EXPLAIN: */ { 0, 2, 1, 1, 3, 1, 1, 1, }, sl@0: /* 3 CREATE: */ { 0, 3, 1, 1, 1, 3, 4, 1, }, sl@0: /* 4 TRIGGER: */ { 5, 4, 4, 4, 4, 4, 4, 4, }, sl@0: /* 5 SEMI: */ { 5, 5, 4, 4, 4, 4, 4, 6, }, sl@0: /* 6 END: */ { 0, 6, 4, 4, 4, 4, 4, 4, }, sl@0: }; sl@0: #else sl@0: /* If triggers are not suppored by this compile then the statement machine sl@0: ** used to detect the end of a statement is much simplier sl@0: */ sl@0: static const u8 trans[2][3] = { sl@0: /* Token: */ sl@0: /* State: ** SEMI WS OTHER */ sl@0: /* 0 START: */ { 0, 0, 1, }, sl@0: /* 1 NORMAL: */ { 0, 1, 1, }, sl@0: }; sl@0: #endif /* SQLITE_OMIT_TRIGGER */ sl@0: sl@0: while( *zSql ){ sl@0: switch( *zSql ){ sl@0: case ';': { /* A semicolon */ sl@0: token = tkSEMI; sl@0: break; sl@0: } sl@0: case ' ': sl@0: case '\r': sl@0: case '\t': sl@0: case '\n': sl@0: case '\f': { /* White space is ignored */ sl@0: token = tkWS; sl@0: break; sl@0: } sl@0: case '/': { /* C-style comments */ sl@0: if( zSql[1]!='*' ){ sl@0: token = tkOTHER; sl@0: break; sl@0: } sl@0: zSql += 2; sl@0: while( zSql[0] && (zSql[0]!='*' || zSql[1]!='/') ){ zSql++; } sl@0: if( zSql[0]==0 ) return 0; sl@0: zSql++; sl@0: token = tkWS; sl@0: break; sl@0: } sl@0: case '-': { /* SQL-style comments from "--" to end of line */ sl@0: if( zSql[1]!='-' ){ sl@0: token = tkOTHER; sl@0: break; sl@0: } sl@0: while( *zSql && *zSql!='\n' ){ zSql++; } sl@0: if( *zSql==0 ) return state==0; sl@0: token = tkWS; sl@0: break; sl@0: } sl@0: case '[': { /* Microsoft-style identifiers in [...] */ sl@0: zSql++; sl@0: while( *zSql && *zSql!=']' ){ zSql++; } sl@0: if( *zSql==0 ) return 0; sl@0: token = tkOTHER; sl@0: break; sl@0: } sl@0: case '`': /* Grave-accent quoted symbols used by MySQL */ sl@0: case '"': /* single- and double-quoted strings */ sl@0: case '\'': { sl@0: int c = *zSql; sl@0: zSql++; sl@0: while( *zSql && *zSql!=c ){ zSql++; } sl@0: if( *zSql==0 ) return 0; sl@0: token = tkOTHER; sl@0: break; sl@0: } sl@0: default: { sl@0: int c; sl@0: if( IdChar((u8)*zSql) ){ sl@0: /* Keywords and unquoted identifiers */ sl@0: int nId; sl@0: for(nId=1; IdChar(zSql[nId]); nId++){} sl@0: #ifdef SQLITE_OMIT_TRIGGER sl@0: token = tkOTHER; sl@0: #else sl@0: switch( *zSql ){ sl@0: case 'c': case 'C': { sl@0: if( nId==6 && sqlite3StrNICmp(zSql, "create", 6)==0 ){ sl@0: token = tkCREATE; sl@0: }else{ sl@0: token = tkOTHER; sl@0: } sl@0: break; sl@0: } sl@0: case 't': case 'T': { sl@0: if( nId==7 && sqlite3StrNICmp(zSql, "trigger", 7)==0 ){ sl@0: token = tkTRIGGER; sl@0: }else if( nId==4 && sqlite3StrNICmp(zSql, "temp", 4)==0 ){ sl@0: token = tkTEMP; sl@0: }else if( nId==9 && sqlite3StrNICmp(zSql, "temporary", 9)==0 ){ sl@0: token = tkTEMP; sl@0: }else{ sl@0: token = tkOTHER; sl@0: } sl@0: break; sl@0: } sl@0: case 'e': case 'E': { sl@0: if( nId==3 && sqlite3StrNICmp(zSql, "end", 3)==0 ){ sl@0: token = tkEND; sl@0: }else sl@0: #ifndef SQLITE_OMIT_EXPLAIN sl@0: if( nId==7 && sqlite3StrNICmp(zSql, "explain", 7)==0 ){ sl@0: token = tkEXPLAIN; sl@0: }else sl@0: #endif sl@0: { sl@0: token = tkOTHER; sl@0: } sl@0: break; sl@0: } sl@0: default: { sl@0: token = tkOTHER; sl@0: break; sl@0: } sl@0: } sl@0: #endif /* SQLITE_OMIT_TRIGGER */ sl@0: zSql += nId-1; sl@0: }else{ sl@0: /* Operators and special symbols */ sl@0: token = tkOTHER; sl@0: } sl@0: break; sl@0: } sl@0: } sl@0: state = trans[state][token]; sl@0: zSql++; sl@0: } sl@0: return state==0; sl@0: } sl@0: sl@0: #ifndef SQLITE_OMIT_UTF16 sl@0: /* sl@0: ** This routine is the same as the sqlite3_complete() routine described sl@0: ** above, except that the parameter is required to be UTF-16 encoded, not sl@0: ** UTF-8. sl@0: */ sl@0: int sqlite3_complete16(const void *zSql){ sl@0: sqlite3_value *pVal; sl@0: char const *zSql8; sl@0: int rc = SQLITE_NOMEM; sl@0: sl@0: #ifndef SQLITE_OMIT_AUTOINIT sl@0: rc = sqlite3_initialize(); sl@0: if( rc ) return rc; sl@0: #endif sl@0: pVal = sqlite3ValueNew(0); sl@0: sqlite3ValueSetStr(pVal, -1, zSql, SQLITE_UTF16NATIVE, SQLITE_STATIC); sl@0: zSql8 = sqlite3ValueText(pVal, SQLITE_UTF8); sl@0: if( zSql8 ){ sl@0: rc = sqlite3_complete(zSql8); sl@0: }else{ sl@0: rc = SQLITE_NOMEM; sl@0: } sl@0: sqlite3ValueFree(pVal); sl@0: return sqlite3ApiExit(0, rc); sl@0: } sl@0: #endif /* SQLITE_OMIT_UTF16 */ sl@0: #endif /* SQLITE_OMIT_COMPLETE */