Avoid signed integer overflow when finding snippets in fts3 by using 64-bit integer offsets.

FossilOrigin-Name: 4cc09a872f627f4a2b94345bef07cd49c3ec3627f8d78c1eb091741cdb4ec0b3

ext/fts3/fts3_snippet.c

@@ -17,6 +17,10 @@
 #include <string.h>
 #include <assert.h>
 
+#ifndef SQLITE_AMALGAMATION
+typedef sqlite3_int64 i64;
+#endif
+
 /*
 ** Characters that may appear in the second argument to matchinfo().
 */
@@ -67,9 +71,9 @@ struct SnippetIter {
 struct SnippetPhrase {
   int nToken;                     /* Number of tokens in phrase */
   char *pList;                    /* Pointer to start of phrase position list */
-  int iHead;                      /* Next value in position list */
+  i64 iHead;                      /* Next value in position list */
   char *pHead;                    /* Position list data following iHead */
-  int iTail;                      /* Next value in trailing position list */
+  i64 iTail;                      /* Next value in trailing position list */
   char *pTail;                    /* Position list data following iTail */
 };
 
@@ -234,7 +238,7 @@ void sqlite3Fts3MIBufferFree(MatchinfoBuffer *p){
 ** After it returns, *piPos contains the value of the next element of the
 ** list and *pp is advanced to the following varint.
 */
-static void fts3GetDeltaPosition(char **pp, int *piPos){
+static void fts3GetDeltaPosition(char **pp, i64 *piPos){
   int iVal;
   *pp += fts3GetVarint32(*pp, &iVal);
   *piPos += (iVal-2);
@@ -343,10 +347,10 @@ static int fts3ExprPhraseCount(Fts3Expr *pExpr){
 ** arguments so that it points to the first element with a value greater
 ** than or equal to parameter iNext.
 */
-static void fts3SnippetAdvance(char **ppIter, int *piIter, int iNext){
+static void fts3SnippetAdvance(char **ppIter, i64 *piIter, int iNext){
   char *pIter = *ppIter;
   if( pIter ){
-    int iIter = *piIter;
+    i64 iIter = *piIter;
 
     while( iIter<iNext ){
       if( 0==(*pIter & 0xFE) ){
@@ -429,7 +433,7 @@ static void fts3SnippetDetails(
     SnippetPhrase *pPhrase = &pIter->aPhrase[i];
     if( pPhrase->pTail ){
       char *pCsr = pPhrase->pTail;
-      int iCsr = pPhrase->iTail;
+      i64 iCsr = pPhrase->iTail;
 
       while( iCsr<(iStart+pIter->nSnippet) && iCsr>=iStart ){
         int j;
@@ -475,7 +479,7 @@ static int fts3SnippetFindPositions(Fts3Expr *pExpr, int iPhrase, void *ctx){
   rc = sqlite3Fts3EvalPhrasePoslist(p->pCsr, pExpr, p->iCol, &pCsr);
   assert( rc==SQLITE_OK || pCsr==0 );
   if( pCsr ){
-    int iFirst = 0;
+    i64 iFirst = 0;
     pPhrase->pList = pCsr;
     fts3GetDeltaPosition(&pCsr, &iFirst);
     if( iFirst<0 ){
@@ -1539,8 +1543,8 @@ typedef struct TermOffsetCtx TermOffsetCtx;
 
 struct TermOffset {
   char *pList;                    /* Position-list */
-  int iPos;                       /* Position just read from pList */
-  int iOff;                       /* Offset of this term from read positions */
+  i64 iPos;                       /* Position just read from pList */
+  i64 iOff;                       /* Offset of this term from read positions */
 };
 
 struct TermOffsetCtx {
@@ -1559,7 +1563,7 @@ static int fts3ExprTermOffsetInit(Fts3Expr *pExpr, int iPhrase, void *ctx){
   int nTerm;                      /* Number of tokens in phrase */
   int iTerm;                      /* For looping through nTerm phrase terms */
   char *pList;                    /* Pointer to position list for phrase */
-  int iPos = 0;                   /* First position in position-list */
+  i64 iPos = 0;                   /* First position in position-list */
   int rc;
 
   UNUSED_PARAMETER(iPhrase);

manifest

@@ -1,5 +1,5 @@
-C Remove\san\sALWAYS()\sfrom\sa\sbranch\sin\ssqlite3ExprAddCollateToken()\swhich\scan\nbe\sfalse\sfollowing\san\sOOM.\ndbsqlfuzz\s9e8516bf1e786c84e520ae43141b75b7399f8618.
-D 2021-04-16T12:33:52.638
+C Avoid\ssigned\sinteger\soverflow\swhen\sfinding\ssnippets\sin\sfts3\sby\susing\s64-bit\sinteger\soffsets.
+D 2021-04-16T16:55:28.621
 F .fossil-settings/empty-dirs dbb81e8fc0401ac46a1491ab34a7f2c7c0452f2f06b54ebb845d024ca8283ef1
 F .fossil-settings/ignore-glob 35175cdfcf539b2318cb04a9901442804be81cd677d8b889fcc9149c21f239ea
 F LICENSE.md df5091916dbb40e6e9686186587125e1b2ff51f022cc334e886c19a0e9982724
@@ -93,7 +93,7 @@ F ext/fts3/fts3_hash.c 8b6e31bfb0844c27dc6092c2620bdb1fca17ed613072db057d96952c6
 F ext/fts3/fts3_hash.h 39cf6874dc239d6b4e30479b1975fe5b22a3caaf
 F ext/fts3/fts3_icu.c 305ce7fb6036484085b5556a9c8e62acdc7763f0f4cdf5fd538212a9f3720116
 F ext/fts3/fts3_porter.c 3565faf04b626cddf85f03825e86056a4562c009
-F ext/fts3/fts3_snippet.c 86e7e947a176f0f005720b3ca17631aca2fd2f9daa6729d4adbf2d16ab1b9613
+F ext/fts3/fts3_snippet.c 57e1965906f3ecfcb5251a14aea98ac2d2dd1868a123bcd5a569601719ab9ead
 F ext/fts3/fts3_term.c f45a1e7c6ef464abb1231245d123dae12266b69e05cc56e14045b76591ae92d1
 F ext/fts3/fts3_test.c d8d7b2734f894e8a489987447658e374cdd3a3bc8575c401decf1911cb7c6454
 F ext/fts3/fts3_tokenize_vtab.c 8d15b148e7d88a4280389a200b26e8d52abda4c4ec2e9a35e9d7a1fa50e5aa03
@@ -1912,7 +1912,7 @@ F vsixtest/vsixtest.tcl 6a9a6ab600c25a91a7acc6293828957a386a8a93
 F vsixtest/vsixtest.vcxproj.data 2ed517e100c66dc455b492e1a33350c1b20fbcdc
 F vsixtest/vsixtest.vcxproj.filters 37e51ffedcdb064aad6ff33b6148725226cd608e
 F vsixtest/vsixtest_TemporaryKey.pfx e5b1b036facdb453873e7084e1cae9102ccc67a0
-P 5cb40e53edcb1e54cde0e54c38a647144bc3158435df0161f95c89732de086e6
-R ddfb443f324f92c0f7d01464d567a778
-U drh
-Z 4128704ba35132d25cf1033fb5faf21a
+P 6af4e6d054efd8445e1010aabd584f36e70dfad0de13bd0e2f1761cad4d9a7d6
+R bb2fc97da9924027621d054c5f109d3c
+U dan
+Z d63dd194e7fabc0d0616c3cb1802e30f

manifest.uuid

@@ -1 +1 @@
-6af4e6d054efd8445e1010aabd584f36e70dfad0de13bd0e2f1761cad4d9a7d6
+4cc09a872f627f4a2b94345bef07cd49c3ec3627f8d78c1eb091741cdb4ec0b3