*/
typedef struct Fts5Config Fts5Config;
+typedef struct Fts5TokenizerConfig Fts5TokenizerConfig;
+
+struct Fts5TokenizerConfig {
+ Fts5Tokenizer *pTok;
+ fts5_tokenizer *pTokApi;
+ const char **azArg;
+ int nArg;
+ int ePattern; /* FTS_PATTERN_XXX constant */
+};
/*
** An instance of the following structure encodes all information that can
*/
struct Fts5Config {
sqlite3 *db; /* Database handle */
+ Fts5Global *pGlobal; /* Global fts5 object for handle db */
char *zDb; /* Database holding FTS index (e.g. "main") */
char *zName; /* Name of FTS index */
int nCol; /* Number of columns */
int bTokendata; /* "tokendata=" option value (dflt==0) */
int eDetail; /* FTS5_DETAIL_XXX value */
char *zContentExprlist;
- Fts5Tokenizer *pTok;
- fts5_tokenizer *pTokApi;
+ Fts5TokenizerConfig t;
int bLock; /* True when table is preparing statement */
- int ePattern; /* FTS_PATTERN_XXX constant */
/* Values loaded from the %_config table */
int iVersion; /* fts5 file format 'version' */
Fts5Index *pIndex; /* Full-text index */
};
-int sqlite3Fts5GetTokenizer(
- Fts5Global*,
- const char **azArg,
- int nArg,
- Fts5Config*,
- char **pzErr
-);
+int sqlite3Fts5LoadTokenizer(Fts5Config *pConfig);
Fts5Table *sqlite3Fts5TableFromCsrid(Fts5Global*, i64);
if( sqlite3_strnicmp("tokenize", zCmd, nCmd)==0 ){
const char *p = (const char*)zArg;
sqlite3_int64 nArg = strlen(zArg) + 1;
- char **azArg = sqlite3Fts5MallocZero(&rc, sizeof(char*) * nArg);
- char *pDel = sqlite3Fts5MallocZero(&rc, nArg * 2);
- char *pSpace = pDel;
+ char **azArg = sqlite3Fts5MallocZero(&rc, (sizeof(char*) + 2) * nArg);
- if( azArg && pSpace ){
- if( pConfig->pTok ){
+ if( azArg ){
+ char *pSpace = (char*)&azArg[nArg];
+ if( pConfig->t.azArg ){
*pzErr = sqlite3_mprintf("multiple tokenize=... directives");
rc = SQLITE_ERROR;
}else{
*pzErr = sqlite3_mprintf("parse error in tokenize directive");
rc = SQLITE_ERROR;
}else{
- rc = sqlite3Fts5GetTokenizer(pGlobal,
- (const char**)azArg, (int)nArg, pConfig,
- pzErr
- );
+ pConfig->t.azArg = (const char**)azArg;
+ pConfig->t.nArg = nArg;
+ azArg = 0;
}
}
}
-
sqlite3_free(azArg);
- sqlite3_free(pDel);
+
return rc;
}
return SQLITE_ERROR;
}
-/*
-** Allocate an instance of the default tokenizer ("simple") at
-** Fts5Config.pTokenizer. Return SQLITE_OK if successful, or an SQLite error
-** code if an error occurs.
-*/
-static int fts5ConfigDefaultTokenizer(Fts5Global *pGlobal, Fts5Config *pConfig){
- assert( pConfig->pTok==0 && pConfig->pTokApi==0 );
- return sqlite3Fts5GetTokenizer(pGlobal, 0, 0, pConfig, 0);
-}
-
/*
** Gobble up the first bareword or quoted word from the input buffer zIn.
** Return a pointer to the character immediately following the last in
*ppOut = pRet = (Fts5Config*)sqlite3_malloc(sizeof(Fts5Config));
if( pRet==0 ) return SQLITE_NOMEM;
memset(pRet, 0, sizeof(Fts5Config));
+ pRet->pGlobal = pGlobal;
pRet->db = db;
pRet->iCookie = -1;
rc = SQLITE_ERROR;
}
- /* If a tokenizer= option was successfully parsed, the tokenizer has
- ** already been allocated. Otherwise, allocate an instance of the default
- ** tokenizer (unicode61) now. */
- if( rc==SQLITE_OK && pRet->pTok==0 ){
- rc = fts5ConfigDefaultTokenizer(pGlobal, pRet);
- }
-
/* If no zContent option was specified, fill in the default values. */
if( rc==SQLITE_OK && pRet->zContent==0 ){
const char *zTail = 0;
void sqlite3Fts5ConfigFree(Fts5Config *pConfig){
if( pConfig ){
int i;
- if( pConfig->pTok ){
- pConfig->pTokApi->xDelete(pConfig->pTok);
+ if( pConfig->t.pTok ){
+ pConfig->t.pTokApi->xDelete(pConfig->t.pTok);
}
+ sqlite3_free(pConfig->t.azArg);
sqlite3_free(pConfig->zDb);
sqlite3_free(pConfig->zName);
for(i=0; i<pConfig->nCol; i++){
void *pCtx, /* Context passed to xToken() */
int (*xToken)(void*, int, const char*, int, int, int) /* Callback */
){
- if( pText==0 ) return SQLITE_OK;
- return pConfig->pTokApi->xTokenize(
- pConfig->pTok, pCtx, flags, pText, nText, xToken
- );
+ int rc = SQLITE_OK;
+ if( pText ){
+ if( pConfig->t.pTok==0 ){
+ rc = sqlite3Fts5LoadTokenizer(pConfig);
+ }
+ if( rc==SQLITE_OK ){
+ rc = pConfig->t.pTokApi->xTokenize(
+ pConfig->t.pTok, pCtx, flags, pText, nText, xToken
+ );
+ }
+ }
+ return rc;
}
/*
assert( (rc==SQLITE_OK && *pzErr==0) || pConfig==0 );
}
if( rc==SQLITE_OK ){
+ pConfig->pzErrmsg = pzErr;
pTab->p.pConfig = pConfig;
pTab->pGlobal = pGlobal;
+ if( bCreate ){
+ rc = sqlite3Fts5LoadTokenizer(pConfig);
+ }
}
/* Open the index sub-system */
/* Load the initial configuration */
if( rc==SQLITE_OK ){
- assert( pConfig->pzErrmsg==0 );
- pConfig->pzErrmsg = pzErr;
rc = sqlite3Fts5IndexLoadConfig(pTab->p.pIndex);
sqlite3Fts5IndexRollback(pTab->p.pIndex);
- pConfig->pzErrmsg = 0;
}
if( rc==SQLITE_OK && pConfig->eContent==FTS5_CONTENT_NORMAL ){
rc = sqlite3_vtab_config(db, SQLITE_VTAB_INNOCUOUS);
}
+ if( pConfig ) pConfig->pzErrmsg = 0;
if( rc!=SQLITE_OK ){
fts5FreeVtab(pTab);
pTab = 0;
){
assert( FTS5_PATTERN_GLOB==SQLITE_INDEX_CONSTRAINT_GLOB );
assert( FTS5_PATTERN_LIKE==SQLITE_INDEX_CONSTRAINT_LIKE );
- if( pConfig->ePattern==FTS5_PATTERN_GLOB && p->op==FTS5_PATTERN_GLOB ){
+ if( pConfig->t.ePattern==FTS5_PATTERN_GLOB && p->op==FTS5_PATTERN_GLOB ){
return 1;
}
- if( pConfig->ePattern==FTS5_PATTERN_LIKE
+ if( pConfig->t.ePattern==FTS5_PATTERN_LIKE
&& (p->op==FTS5_PATTERN_LIKE || p->op==FTS5_PATTERN_GLOB)
){
return 1;
return SQLITE_ERROR;
}
+ if( pConfig->t.pTok==0 ){
+ int rc;
+ pConfig->pzErrmsg = &pVTab->zErrMsg;
+ rc = sqlite3Fts5LoadTokenizer(pConfig);
+ pConfig->pzErrmsg = 0;
+ if( rc!=SQLITE_OK ) return rc;
+ }
+
idxStr = (char*)sqlite3_malloc(pInfo->nConstraint * 8 + 1);
if( idxStr==0 ) return SQLITE_NOMEM;
pInfo->idxStr = idxStr;
idxStr += strlen(&idxStr[iIdxStr]);
pInfo->aConstraintUsage[i].argvIndex = ++iCons;
assert( idxStr[iIdxStr]=='\0' );
+ bSeenMatch = 1;
}else if( bSeenEq==0 && p->op==SQLITE_INDEX_CONSTRAINT_EQ && iCol<0 ){
idxStr[iIdxStr++] = '=';
bSeenEq = 1;
return rc;
}
-int sqlite3Fts5GetTokenizer(
+int fts5GetTokenizer(
Fts5Global *pGlobal,
const char **azArg,
int nArg,
*pzErr = sqlite3_mprintf("no such tokenizer: %s", azArg[0]);
}else{
rc = pMod->x.xCreate(
- pMod->pUserData, (azArg?&azArg[1]:0), (nArg?nArg-1:0), &pConfig->pTok
+ pMod->pUserData, (azArg?&azArg[1]:0), (nArg?nArg-1:0), &pConfig->t.pTok
);
- pConfig->pTokApi = &pMod->x;
+ pConfig->t.pTokApi = &pMod->x;
if( rc!=SQLITE_OK ){
if( pzErr ) *pzErr = sqlite3_mprintf("error in tokenizer constructor");
}else{
- pConfig->ePattern = sqlite3Fts5TokenizerPattern(
- pMod->x.xCreate, pConfig->pTok
+ pConfig->t.ePattern = sqlite3Fts5TokenizerPattern(
+ pMod->x.xCreate, pConfig->t.pTok
);
}
}
if( rc!=SQLITE_OK ){
- pConfig->pTokApi = 0;
- pConfig->pTok = 0;
+ pConfig->t.pTokApi = 0;
+ pConfig->t.pTok = 0;
}
return rc;
}
+/*
+** Attempt to instantiate the tokenizer.
+*/
+int sqlite3Fts5LoadTokenizer(Fts5Config *pConfig){
+ return fts5GetTokenizer(
+ pConfig->pGlobal, pConfig->t.azArg, pConfig->t.nArg,
+ pConfig, pConfig->pzErrmsg
+ );
+}
+
+
static void fts5ModuleDestroy(void *pCtx){
Fts5TokenizerModule *pTok, *pNextTok;
Fts5Auxiliary *pAux, *pNextAux;
do_execsql_test 9.5.1 { SELECT * FROM t1('"abc xyz*"'); } {}
do_test 9.5.2 { set ::flags } {query}
+#-------------------------------------------------------------------------
+#
+reset_db
+do_execsql_test 10.1 {
+ CREATE VIRTUAL TABLE x1 USING fts5(x, tokenize=unicode61);
+ PRAGMA writable_schema = 1;
+ UPDATE sqlite_schema
+ SET sql = 'CREATE VIRTUAL TABLE x1 USING fts5(x, tokenize="unicode61 error");'
+ WHERE name = 'x1';
+}
+
+db close
+sqlite3 db test.db
+
+do_catchsql_test 10.2 {
+ SELECT * FROM x1
+} {1 {error in tokenizer constructor}}
+
+do_catchsql_test 10.3 {
+ INSERT INTO x1 VALUES('abc');
+} {1 {error in tokenizer constructor}}
+
+do_execsql_test 10.4 {
+ PRAGMA writable_schema = 1;
+ UPDATE sqlite_schema
+ SET sql = 'CREATE VIRTUAL TABLE x1 USING fts5(x, tokenize="nosuch error");'
+ WHERE name = 'x1';
+}
+
+db close
+sqlite3 db test.db
+
+do_catchsql_test 10.5 {
+ SELECT * FROM x1
+} {1 {no such tokenizer: nosuch}}
+do_catchsql_test 10.6 {
+ INSERT INTO x1 VALUES('abc');
+} {1 {no such tokenizer: nosuch}}
+
+do_execsql_test 10.7 {
+ DROP TABLE x1;
+ SELECT * FROM sqlite_schema;
+}
finish_test
CREATE VIRTUAL TABLE t4 USING fts5(z, tokenize=trigram);
} {}
-breakpoint
do_execsql_test 4.1 {
INSERT INTO t4 VALUES('ABCD');
+ INSERT INTO t4 VALUES('DEFG');
} {}
+db close
+sqlite3 db test.db
+
+do_eqp_test 4.1 {
+ SELECT rowid FROM t4 WHERE z LIKE '%abc%'
+} {VIRTUAL TABLE INDEX 0:L0}
+
+do_execsql_test 4.2 {
+ SELECT rowid FROM t4 WHERE z LIKE '%abc%'
+} {1}
+
finish_test
-C Add\stests\sto\sbestindexC.test.\sNo\schanges\sto\scode.
-D 2024-05-11T16:44:56.528
+C Allow\sexisting\sfts5\stables\sto\sbe\sdropped\seven\sif\sthe\sassociated\stokenizer\sis\snot\savailable.
+D 2024-05-13T20:06:08.273
F .fossil-settings/empty-dirs dbb81e8fc0401ac46a1491ab34a7f2c7c0452f2f06b54ebb845d024ca8283ef1
F .fossil-settings/ignore-glob 35175cdfcf539b2318cb04a9901442804be81cd677d8b889fcc9149c21f239ea
F LICENSE.md df5091916dbb40e6e9686186587125e1b2ff51f022cc334e886c19a0e9982724
F ext/fts3/unicode/parseunicode.tcl a981bd6466d12dd17967515801c3ff23f74a281be1a03cf1e6f52a6959fc77eb
F ext/fts5/extract_api_docs.tcl bc3a0ca78be7d3df08e7602c00ca48021ebae40682d75eb001bfdf6e54ffb44e
F ext/fts5/fts5.h 8856e11a5f0269cd346754cea0765efe8089635b80cad3222e8bfdb08cd5348a
-F ext/fts5/fts5Int.h defa43c0932265138ee910ca416e6baccf8b774e0f3d610e74be1ab2880e9834
+F ext/fts5/fts5Int.h 965ded9eedfa4bbe6b39048d4b0cce804a13ec954f11a088d1205a48703ce03f
F ext/fts5/fts5_aux.c 4584e88878e54828bf7d4d0d83deedd232ec60628b7731be02bad6adb62304b1
F ext/fts5/fts5_buffer.c 0eec58bff585f1a44ea9147eae5da2447292080ea435957f7488c70673cb6f09
-F ext/fts5/fts5_config.c 8072a207034b51ae9b7694121d1b5715c794e94b275e088f70ae532378ca5cdf
+F ext/fts5/fts5_config.c 74093394dc26750becc302922e5fe308a8597a70e6e83e66c892242952468163
F ext/fts5/fts5_expr.c e91156ebdcc08d837f4f324168f69f3c0d7fdef0e521fd561efb48ef3297b696
F ext/fts5/fts5_hash.c adda4272be401566a6e0ba1acbe70ee5cb97fce944bc2e04dc707152a0ec91b1
F ext/fts5/fts5_index.c ee0f4d50bc0c58a7c5ef7d645e7e38e1e59315b8ea9d722ae00c5f949ee65379
-F ext/fts5/fts5_main.c d68bd9533d5a638b7f6fae61c3cb0a15257dcdcccedaf3d0b3c9f55940c85048
+F ext/fts5/fts5_main.c 942fea921d67f75f745d02bf383e5b949bf96a8b78bcfffca40f06439d252fae
F ext/fts5/fts5_storage.c f9e31b0d155e9b2c92d5d3a09ad7a56b937fbf1c7f962e10f4ca6281349f3934
F ext/fts5/fts5_tcl.c fdf7e2bb9a9186cfcaf2d2ce11d338309342b7a7593c2812bc54455db53da5d2
F ext/fts5/fts5_test_mi.c 08c11ec968148d4cb4119d96d819f8c1f329812c568bac3684f5464be177d3ee
F ext/fts5/test/fts5synonym2.test e2f6ff68c4fbe12a866a3a87510f553d9dac99bcb74c10b56487c4c0a562fcf5
F ext/fts5/test/fts5tok1.test 1f7817499f5971450d8c4a652114b3d833393c8134e32422d0af27884ffe9cef
F ext/fts5/test/fts5tok2.test dcacb32d4a2a3f0dd3215d4a3987f78ae4be21a2
-F ext/fts5/test/fts5tokenizer.test ac3c9112b263a639fb0508ae73a3ee886bf4866d2153771a8e8a20c721305a43
+F ext/fts5/test/fts5tokenizer.test e5ff1fa93ae267592ad2280ae545133ac24d19cf44116c736e7ece9c02285b5b
F ext/fts5/test/fts5tokenizer2.test cb5428c7cfb3b6a74b7adfcde65506e329112003e8dffa7501d01c2d18d02569
F ext/fts5/test/fts5trigram.test 6c4e37864f3e7d90673db5563d9736d7e40080ab94d10ebdffa94c1b77941da0
-F ext/fts5/test/fts5trigram2.test 9fe4207f8a4241747aff1005258b564958588d21bfd240d6cd4c2e955d31c156
+F ext/fts5/test/fts5trigram2.test c91f0a94f7e1ff859682228646abeab4c0eba2effc46af2cbc8f0f48b05a0566
F ext/fts5/test/fts5ubsan.test 783d5a8d13ebfa169e634940228db54540780e3ba7a87ad1e4510e61440bf64b
F ext/fts5/test/fts5umlaut.test a42fe2fe6387c40c49ab27ccbd070e1ae38e07f38d05926482cc0bccac9ad602
F ext/fts5/test/fts5unicode.test 17056f4efe6b0a5d4f41fdf7a7dc9af2873004562eaa899d40633b93dc95f5a9
F vsixtest/vsixtest.vcxproj.data 2ed517e100c66dc455b492e1a33350c1b20fbcdc
F vsixtest/vsixtest.vcxproj.filters 37e51ffedcdb064aad6ff33b6148725226cd608e
F vsixtest/vsixtest_TemporaryKey.pfx e5b1b036facdb453873e7084e1cae9102ccc67a0
-P c6fd70b3c23fa00eaac9286d4a67e5c8ac76f926c11c220250c34032647bedc1
-R da518a0d510ea768080aa36a2357bedd
+P 58d938c0e03c3c8d8796c537f89e69734ba6263d60ba37e345259cb6fdffbea5
+R 8cd6e7131b6ecbfb841a3c6de5597a39
+T *branch * fts5-delay-tokenizer
+T *sym-fts5-delay-tokenizer *
+T -sym-trunk *
U dan
-Z 00211d5eb53202476ca10f2c5f60b56e
+Z f97ec1a719d45c2730606c10145cb9ed
# Remove this line to create a well-formed Fossil manifest.
-58d938c0e03c3c8d8796c537f89e69734ba6263d60ba37e345259cb6fdffbea5
\ No newline at end of file
+69ef47eeee8b53684c321393be34f03600694fbc86377f8720ff80307846aff6
\ No newline at end of file