Enhance the tokenizer so that it is able to deal with individual tokens

larger than 2GiB.

FossilOrigin-Name: 4e713558f6c6ea3dfbc97d2eb0a328955ef6ad6f2ad05517efe9ed4a60181440

diff --git a/ext/misc/normalize.c b/ext/misc/normalize.c
index 800e1291121448a0a1a5843d70df9391c7222551..44ddcd3882885cc786ead4ad15069f04db689366 100644 (file)
--- a/ext/misc/normalize.c
+++ b/ext/misc/normalize.c
@@ -297,8 +297,9 @@ static const unsigned char sqlite3CtypeMap[256] = {
 ** Return the length (in bytes) of the token that begins at z[0]. 
 ** Store the token type in *tokenType before returning.
 */
-static int sqlite3GetToken(const unsigned char *z, int *tokenType){
-  int i, c;
+static sqlite3_int64 sqlite3GetToken(const unsigned char *z, int *tokenType){
+  sqlite3_int64 i;
+  int c;
   switch( aiClass[*z] ){  /* Switch on the character-class of the first byte
                           ** of the token. See the comment on the CC_ defines
                           ** above. */
@@ -559,7 +560,7 @@ char *sqlite3_normalize(const char *zSql){
   int i;                /* Next character to read from zSql[] */
   int j;                /* Next slot to fill in on z[] */
   int tokenType;        /* Type of the next token */
-  int n;                /* Size of the next token */
+  sqlite3_int64 n;      /* Size of the next token */
   int k;                /* Loop counter */
 
   nSql = strlen(zSql);

diff --git a/ext/rtree/rtree.c b/ext/rtree/rtree.c
index fb35bc10e630e95a4b894bd78ec15981cec53703..cf125a705d81110d074372a8dfe1e82f1e04fcfe 100644 (file)
--- a/ext/rtree/rtree.c
+++ b/ext/rtree/rtree.c
@@ -62,7 +62,7 @@
 #else
   #include "sqlite3.h"
 #endif
-int sqlite3GetToken(const unsigned char*,int*); /* In the SQLite core */
+i64 sqlite3GetToken(const unsigned char*,int*); /* In the SQLite core */
 
 #include <stddef.h>
 

diff --git a/manifest b/manifest
index 2af899cdc404eaad2be14127aba84d7eaf2db861..311eba21bb0b19e437ff11b732379c72e1a7d98f 100644 (file)
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C In\sthe\sCLI,\simproved\serror\smessages\son\sinput\serrors.\s\sLimit\sthe\ssize\sof\sa\nsingle\sSQL\sstatement\sto\sa\slittle\smore\sthan\s2GB.
-D 2025-10-24T12:32:32.322
+C Enhance\sthe\stokenizer\sso\sthat\sit\sis\sable\sto\sdeal\swith\sindividual\stokens\nlarger\sthan\s2GiB.
+D 2025-10-24T13:18:31.816
 F .fossil-settings/binary-glob 61195414528fb3ea9693577e1980230d78a1f8b0a54c78cf1b9b24d0a409ed6a x
 F .fossil-settings/empty-dirs dbb81e8fc0401ac46a1491ab34a7f2c7c0452f2f06b54ebb845d024ca8283ef1
 F .fossil-settings/ignore-glob 35175cdfcf539b2318cb04a9901442804be81cd677d8b889fcc9149c21f239ea
@@ -381,7 +381,7 @@ F ext/misc/memtrace.c 7c0d115d2ef716ad0ba632c91e05bd119cb16c1aedf3bec9f06196ead2
 F ext/misc/mmapwarm.c a81af4aaec00f24f308e2f4c19bf1d88f3ac3ce848c36daa7a4cd38145c4080d
 F ext/misc/nextchar.c 7877914c2a80c2f181dd04c3dbef550dfb54c93495dc03da2403b5dd58f34edd
 F ext/misc/noop.c f1a21cc9b7a4e667e5c8458d80ba680b8bd4315a003f256006046879f679c5a0
-F ext/misc/normalize.c 4782be3b74b9bd9f67281036ff1f41e5edcad20ad486171a2d671c4bb2586011
+F ext/misc/normalize.c fbb144a861809686ff2b5b6eee8bb2e1207f9bf13ce7376e5273c700a1eafbd5
 F ext/misc/pcachetrace.c f4227ce03fb16aa8d6f321b72dd051097419d7a028a9853af048bee7645cb405
 F ext/misc/percentile.c 72e05a21db20a2fa85264b99515941f00ae698824c9db82d7edfbb16cea8ec80
 F ext/misc/prefixes.c 82645f79229877afab08c8b08ca1e7fa31921280906b90a61c294e4f540cd2a6
@@ -493,7 +493,7 @@ F ext/repair/test/checkindex01.test b530f141413b587c9eb78ff734de6bb79bc3515c3350
 F ext/repair/test/test.tcl 686d76d888dffd021f64260abf29a55c57b2cedfa7fc69150b42b1d6119aac3c
 F ext/rtree/README 734aa36238bcd2dee91db5dba107d5fcbdb02396612811377a8ad50f1272b1c1
 F ext/rtree/geopoly.c f0573d5109fdc658a180db0db6eec86ab2a1cf5ce58ec66cbf3356167ea757eb
-F ext/rtree/rtree.c 86967c5a501f895b9705366b8cd9c37f15a9ebdff770ceb719c7deeeb2c22b72
+F ext/rtree/rtree.c a884309b7d7f81ec2845a4a0f93885cd3a3de3278584d0cbee74c92711190869
 F ext/rtree/rtree.h 4a690463901cb5e6127cf05eb8e642f127012fd5003830dbc974eca5802d9412
 F ext/rtree/rtree1.test e0608db762b2aadca0ecb6f97396cf66244490adc3ba88f2a292b27be3e1da3e
 F ext/rtree/rtree2.test 9d9deddbb16fd0c30c36e6b4fdc3ee3132d765567f0f9432ee71e1303d32603d
@@ -740,7 +740,7 @@ F src/shell.c.in 8d218cb62f8badb58cb6d91302ca8b6ccdc93232dd2ba39a9ace2e41fd8495c
 F src/sqlite.h.in 10faecc456d3962c7cedae70d69305f7c80129f28dd8524bd8a06b3eac955e54
 F src/sqlite3.rc 015537e6ac1eec6c7050e17b616c2ffe6f70fca241835a84a4f0d5937383c479
 F src/sqlite3ext.h 7f236ca1b175ffe03316d974ef57df79b3938466c28d2f95caef5e08c57f3a52
-F src/sqliteInt.h e16df0914718acba6fb7e5f03055f7cf519d0d61314c78ec2f5c9f6a45608d86
+F src/sqliteInt.h 88f7fc9ce1630d9a5f7e0a8e1f3287cdc63882fba985c18e7eee1b9f457f59aa
 F src/sqliteLimit.h fe70bd8983e5d317a264f2ea97473b359faf3ebb0827877a76813f5cf0cdc364
 F src/status.c 7565d63a79aa2f326339a24a0461a60096d0bd2bce711fefb50b5c89335f3592
 F src/table.c 0f141b58a16de7e2fbe81c308379e7279f4c6b50eb08efeec5892794a0ba30d1
@@ -794,7 +794,7 @@ F src/test_vfs.c b4135c1308516adf0dfd494e6d6c33114e03732be899eace0502919b674586b
 F src/test_window.c 6d80e11fba89a1796525e6f0048ff0c7789aa2c6b0b11c80827dc1437bd8ea72
 F src/test_wsd.c 41cadfd9d97fe8e3e4e44f61a4a8ccd6f7ca8fe9
 F src/threads.c 4ae07fa022a3dc7c5beb373cf744a85d3c5c6c3c
-F src/tokenize.c 8400646d2830afc2f2dc465a75e3a92e4bedeea623f19dbd79c0c12d0dd6dda2
+F src/tokenize.c cb3294cf23c11106b50d9af6998a6c1bf389b52e15b17698c9fab97bbaa9b37f
 F src/treeview.c 3ce7ac9835d2d70cc1c868b01b747ae8a062322e155701e58e3d62ca79aada7a
 F src/trigger.c d5cf2541ff048f30b6a0507eb3d1ec4e695c53584e3b2298a5bf248714fe185e
 F src/update.c 3e5e7ff66fa19ebe4d1b113d480639a24cc1175adbefabbd1a948a07f28e37cf
@@ -810,7 +810,7 @@ F src/vdbeaux.c cb907297d03b374544a5a4f81ca981497ff4d09d13838f06de507118367f0b73
 F src/vdbeblob.c b3f0640db9642fbdc88bd6ebcc83d6009514cafc98f062f675f2c8d505d82692
 F src/vdbemem.c e67d9c6484d868c879d20c70d00bf4a9058082f1d4058607ca15d50eb3aebc21
 F src/vdbesort.c b69220f4ea9ffea5fdef34d968c60305444eea909252a81933b54c296d9cca70
-F src/vdbetrace.c fe0bc29ebd4e02c8bc5c1945f1d2e6be5927ec12c06d89b03ef2a4def34bf823
+F src/vdbetrace.c 49e689f751505839742f4a243a1a566e57d5c9eaf0d33bbaa26e2de3febf7b41
 F src/vdbevtab.c fc46b9cbd759dc013f0b3724549cc0d71379183c667df3a5988f7e2f1bd485f3
 F src/vtab.c 828221bdbeaaa6d62126ee6d07fd4ec0d09dcaea846f87ad01944d8b7e548859
 F src/vxworks.h 9d18819c5235b49c2340a8a4d48195ec5d5afb637b152406de95a9436beeaeab
@@ -2171,8 +2171,8 @@ F tool/version-info.c 33d0390ef484b3b1cb685d59362be891ea162123cea181cb8e6d2cf6dd
 F tool/warnings-clang.sh bbf6a1e685e534c92ec2bfba5b1745f34fb6f0bc2a362850723a9ee87c1b31a7
 F tool/warnings.sh 1ad0169b022b280bcaaf94a7fa231591be96b514230ab5c98fbf15cd7df842dd
 F tool/win/sqlite.vsix deb315d026cc8400325c5863eef847784a219a2f
-P 2adfd0f47b028b8378e6cc08dc22abf1606036bbd285a7bc3a0de0eaf6feeb8f
-R 1cb5603af7ab9bea0b6181cf5c07b1e1
+P 0b50e7e10aae3d57f1cbd38d02be50cfbd2d5148ab4a9be712afd8678787b7bd
+R 12ad4680ceff58939b23031232c526dd
 U drh
-Z 8148667aee325651b978014f5189e836
+Z ef25866760ed6d5dd12ab7c0cb6034f8
 # Remove this line to create a well-formed Fossil manifest.

diff --git a/manifest.uuid b/manifest.uuid
index 2b4c46a6828f5691f8b89708eb42157013862c96..14630e131a93fbd2f760e5dcb29d6a43bc216e46 100644 (file)
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-0b50e7e10aae3d57f1cbd38d02be50cfbd2d5148ab4a9be712afd8678787b7bd
+4e713558f6c6ea3dfbc97d2eb0a328955ef6ad6f2ad05517efe9ed4a60181440

diff --git a/src/sqliteInt.h b/src/sqliteInt.h
index 25bb07a7c8283161aec6712973f2b4e664883023..523bcfb3bd92c7ecff95e5f0b22384a0efc6252d 100644 (file)
--- a/src/sqliteInt.h
+++ b/src/sqliteInt.h
@@ -5452,7 +5452,7 @@ void sqlite3Reindex(Parse*, Token*, Token*);
 void sqlite3AlterFunctions(void);
 void sqlite3AlterRenameTable(Parse*, SrcList*, Token*);
 void sqlite3AlterRenameColumn(Parse*, SrcList*, Token*, Token*);
-int sqlite3GetToken(const unsigned char *, int *);
+i64 sqlite3GetToken(const unsigned char *, int *);
 void sqlite3NestedParse(Parse*, const char*, ...);
 void sqlite3ExpirePreparedStatements(sqlite3*, int);
 void sqlite3CodeRhsOfIN(Parse*, Expr*, int);

diff --git a/src/tokenize.c b/src/tokenize.c
index 6f7bab35bc1a06b40b0bd33e250a5fbfc146f57f..152ada64f6fb41648a9b8352f1cc73dfdc217b8b 100644 (file)
--- a/src/tokenize.c
+++ b/src/tokenize.c
@@ -270,8 +270,9 @@ static int analyzeFilterKeyword(const unsigned char *z, int lastToken){
 ** Return the length (in bytes) of the token that begins at z[0]. 
 ** Store the token type in *tokenType before returning.
 */
-int sqlite3GetToken(const unsigned char *z, int *tokenType){
-  int i, c;
+i64 sqlite3GetToken(const unsigned char *z, int *tokenType){
+  i64 i;
+  int c;
   switch( aiClass[*z] ){  /* Switch on the character-class of the first byte
                           ** of the token. See the comment on the CC_ defines
                           ** above. */
@@ -599,7 +600,7 @@ int sqlite3GetToken(const unsigned char *z, int *tokenType){
 int sqlite3RunParser(Parse *pParse, const char *zSql){
   int nErr = 0;                   /* Number of errors encountered */
   void *pEngine;                  /* The LEMON-generated LALR(1) parser */
-  int n = 0;                      /* Length of the next token token */
+  i64 n = 0;                      /* Length of the next token token */
   int tokenType;                  /* type of the next token */
   int lastTokenParsed = -1;       /* type of the previous token */
   sqlite3 *db = pParse->db;       /* The database connection */
@@ -702,13 +703,13 @@ int sqlite3RunParser(Parse *pParse, const char *zSql){
       }else if( tokenType!=TK_QNUMBER ){
         Token x;
         x.z = zSql;
-        x.n = n;
+        x.n = (u32)n;
         sqlite3ErrorMsg(pParse, "unrecognized token: \"%T\"", &x);
         break;
       }
     }
     pParse->sLastToken.z = zSql;
-    pParse->sLastToken.n = n;
+    pParse->sLastToken.n = (u32)n;
     sqlite3Parser(pEngine, tokenType, pParse->sLastToken);
     lastTokenParsed = tokenType;
     zSql += n;
@@ -784,7 +785,7 @@ char *sqlite3Normalize(
 ){
   sqlite3 *db;       /* The database connection */
   int i;             /* Next unread byte of zSql[] */
-  int n;             /* length of current token */
+  i64 n;             /* length of current token */
   int tokenType;     /* type of current token */
   int prevType = 0;  /* Previous non-whitespace token */
   int nParen;        /* Number of nested levels of parentheses */

diff --git a/src/vdbetrace.c b/src/vdbetrace.c
index ae8ad3115fe60dfdf2b812f45a87d4789e20d7d5..1a59f0e4d5580d863f7513152a2e45266ae188bc 100644 (file)
--- a/src/vdbetrace.c
+++ b/src/vdbetrace.c
@@ -26,10 +26,10 @@
 ** a host parameter.  If the text contains no host parameters, return
 ** the total number of bytes in the text.
 */
-static int findNextHostParameter(const char *zSql, int *pnToken){
+static i64 findNextHostParameter(const char *zSql, i64 *pnToken){
   int tokenType;
-  int nTotal = 0;
-  int n;
+  i64 nTotal = 0;
+  i64 n;
 
   *pnToken = 0;
   while( zSql[0] ){
@@ -76,8 +76,8 @@ char *sqlite3VdbeExpandSql(
   sqlite3 *db;             /* The database connection */
   int idx = 0;             /* Index of a host parameter */
   int nextIndex = 1;       /* Index of next ? host parameter */
-  int n;                   /* Length of a token prefix */
-  int nToken;              /* Length of the parameter token */
+  i64 n;                   /* Length of a token prefix */
+  i64 nToken;              /* Length of the parameter token */
   int i;                   /* Loop counter */
   Mem *pVar;               /* Value of a host parameter */
   StrAccum out;            /* Accumulate the output here */