1
0
mirror of https://github.com/sqlite/sqlite.git synced 2025-07-30 19:03:16 +03:00

Fix the way parenthesis in MATCH expressions are handled by FTS if the tokenizer considers them to be token characters.

FossilOrigin-Name: e21bf7a2ade6373e94ea403c665f78e1ad22143f
This commit is contained in:
dan
2014-05-07 19:59:36 +00:00
parent faecf50504
commit d2d8ca600e
6 changed files with 178 additions and 112 deletions

View File

@ -185,40 +185,23 @@ static int getNextToken(
int rc;
sqlite3_tokenizer_cursor *pCursor;
Fts3Expr *pRet = 0;
int nConsumed = 0;
int i = 0;
rc = sqlite3Fts3OpenTokenizer(pTokenizer, pParse->iLangid, z, n, &pCursor);
/* Set variable i to the maximum number of bytes of input to tokenize. */
for(i=0; i<n; i++){
if( sqlite3_fts3_enable_parentheses && (z[i]=='(' || z[i]==')') ) break;
if( z[i]=='*' || z[i]=='"' ) break;
}
*pnConsumed = i;
rc = sqlite3Fts3OpenTokenizer(pTokenizer, pParse->iLangid, z, i, &pCursor);
if( rc==SQLITE_OK ){
const char *zToken;
int nToken = 0, iStart = 0, iEnd = 0, iPosition = 0;
int nByte; /* total space to allocate */
rc = pModule->xNext(pCursor, &zToken, &nToken, &iStart, &iEnd, &iPosition);
if( (rc==SQLITE_OK || rc==SQLITE_DONE) && sqlite3_fts3_enable_parentheses ){
int i;
if( rc==SQLITE_DONE ) iStart = n;
for(i=0; i<iStart; i++){
if( z[i]=='(' ){
pParse->nNest++;
rc = fts3ExprParse(pParse, &z[i+1], n-i-1, &pRet, &nConsumed);
if( rc==SQLITE_OK && !pRet ){
rc = SQLITE_DONE;
}
nConsumed = (int)(i + 1 + nConsumed);
break;
}
if( z[i]==')' ){
rc = SQLITE_DONE;
pParse->nNest--;
nConsumed = i+1;
break;
}
}
}
if( nConsumed==0 && rc==SQLITE_OK ){
if( rc==SQLITE_OK ){
nByte = sizeof(Fts3Expr) + sizeof(Fts3Phrase) + nToken;
pRet = (Fts3Expr *)fts3MallocZero(nByte);
if( !pRet ){
@ -252,13 +235,14 @@ static int getNextToken(
}
}
nConsumed = iEnd;
*pnConsumed = iEnd;
}else if( i && rc==SQLITE_DONE ){
rc = SQLITE_OK;
}
pModule->xClose(pCursor);
}
*pnConsumed = nConsumed;
*ppExpr = pRet;
return rc;
}
@ -508,6 +492,21 @@ static int getNextNode(
return getNextString(pParse, &zInput[1], ii-1, ppExpr);
}
if( sqlite3_fts3_enable_parentheses ){
if( *zInput=='(' ){
int nConsumed = 0;
pParse->nNest++;
rc = fts3ExprParse(pParse, zInput+1, nInput-1, ppExpr, &nConsumed);
if( rc==SQLITE_OK && !*ppExpr ){ rc = SQLITE_DONE; }
*pnConsumed = (int)(zInput - z) + 1 + nConsumed;
return rc;
}else if( *zInput==')' ){
pParse->nNest--;
*pnConsumed = (zInput - z) + 1;
*ppExpr = 0;
return SQLITE_DONE;
}
}
/* If control flows to this point, this must be a regular token, or
** the end of the input. Read a regular token using the sqlite3_tokenizer
@ -626,96 +625,100 @@ static int fts3ExprParse(
while( rc==SQLITE_OK ){
Fts3Expr *p = 0;
int nByte = 0;
rc = getNextNode(pParse, zIn, nIn, &p, &nByte);
assert( nByte>0 || (rc!=SQLITE_OK && p==0) );
if( rc==SQLITE_OK ){
int isPhrase;
if( p ){
int isPhrase;
if( !sqlite3_fts3_enable_parentheses
&& p->eType==FTSQUERY_PHRASE && pParse->isNot
){
/* Create an implicit NOT operator. */
Fts3Expr *pNot = fts3MallocZero(sizeof(Fts3Expr));
if( !pNot ){
sqlite3Fts3ExprFree(p);
rc = SQLITE_NOMEM;
goto exprparse_out;
}
pNot->eType = FTSQUERY_NOT;
pNot->pRight = p;
p->pParent = pNot;
if( pNotBranch ){
pNot->pLeft = pNotBranch;
pNotBranch->pParent = pNot;
}
pNotBranch = pNot;
p = pPrev;
}else{
int eType = p->eType;
isPhrase = (eType==FTSQUERY_PHRASE || p->pLeft);
/* The isRequirePhrase variable is set to true if a phrase or
** an expression contained in parenthesis is required. If a
** binary operator (AND, OR, NOT or NEAR) is encounted when
** isRequirePhrase is set, this is a syntax error.
*/
if( !isPhrase && isRequirePhrase ){
sqlite3Fts3ExprFree(p);
rc = SQLITE_ERROR;
goto exprparse_out;
}
if( isPhrase && !isRequirePhrase ){
/* Insert an implicit AND operator. */
Fts3Expr *pAnd;
assert( pRet && pPrev );
pAnd = fts3MallocZero(sizeof(Fts3Expr));
if( !pAnd ){
if( !sqlite3_fts3_enable_parentheses
&& p->eType==FTSQUERY_PHRASE && pParse->isNot
){
/* Create an implicit NOT operator. */
Fts3Expr *pNot = fts3MallocZero(sizeof(Fts3Expr));
if( !pNot ){
sqlite3Fts3ExprFree(p);
rc = SQLITE_NOMEM;
goto exprparse_out;
}
pAnd->eType = FTSQUERY_AND;
insertBinaryOperator(&pRet, pPrev, pAnd);
pPrev = pAnd;
}
pNot->eType = FTSQUERY_NOT;
pNot->pRight = p;
p->pParent = pNot;
if( pNotBranch ){
pNot->pLeft = pNotBranch;
pNotBranch->pParent = pNot;
}
pNotBranch = pNot;
p = pPrev;
}else{
int eType = p->eType;
isPhrase = (eType==FTSQUERY_PHRASE || p->pLeft);
/* This test catches attempts to make either operand of a NEAR
** operator something other than a phrase. For example, either of
** the following:
**
** (bracketed expression) NEAR phrase
** phrase NEAR (bracketed expression)
**
** Return an error in either case.
*/
if( pPrev && (
/* The isRequirePhrase variable is set to true if a phrase or
** an expression contained in parenthesis is required. If a
** binary operator (AND, OR, NOT or NEAR) is encounted when
** isRequirePhrase is set, this is a syntax error.
*/
if( !isPhrase && isRequirePhrase ){
sqlite3Fts3ExprFree(p);
rc = SQLITE_ERROR;
goto exprparse_out;
}
if( isPhrase && !isRequirePhrase ){
/* Insert an implicit AND operator. */
Fts3Expr *pAnd;
assert( pRet && pPrev );
pAnd = fts3MallocZero(sizeof(Fts3Expr));
if( !pAnd ){
sqlite3Fts3ExprFree(p);
rc = SQLITE_NOMEM;
goto exprparse_out;
}
pAnd->eType = FTSQUERY_AND;
insertBinaryOperator(&pRet, pPrev, pAnd);
pPrev = pAnd;
}
/* This test catches attempts to make either operand of a NEAR
** operator something other than a phrase. For example, either of
** the following:
**
** (bracketed expression) NEAR phrase
** phrase NEAR (bracketed expression)
**
** Return an error in either case.
*/
if( pPrev && (
(eType==FTSQUERY_NEAR && !isPhrase && pPrev->eType!=FTSQUERY_PHRASE)
|| (eType!=FTSQUERY_PHRASE && isPhrase && pPrev->eType==FTSQUERY_NEAR)
)){
sqlite3Fts3ExprFree(p);
rc = SQLITE_ERROR;
goto exprparse_out;
}
if( isPhrase ){
if( pRet ){
assert( pPrev && pPrev->pLeft && pPrev->pRight==0 );
pPrev->pRight = p;
p->pParent = pPrev;
}else{
pRet = p;
)){
sqlite3Fts3ExprFree(p);
rc = SQLITE_ERROR;
goto exprparse_out;
}
}else{
insertBinaryOperator(&pRet, pPrev, p);
if( isPhrase ){
if( pRet ){
assert( pPrev && pPrev->pLeft && pPrev->pRight==0 );
pPrev->pRight = p;
p->pParent = pPrev;
}else{
pRet = p;
}
}else{
insertBinaryOperator(&pRet, pPrev, p);
}
isRequirePhrase = !isPhrase;
}
isRequirePhrase = !isPhrase;
pPrev = p;
}
assert( nByte>0 );
}
assert( rc!=SQLITE_OK || (nByte>0 && nByte<=nIn) );
nIn -= nByte;
zIn += nByte;
pPrev = p;
}
if( rc==SQLITE_DONE && pRet && isRequirePhrase ){