Another change to the fts5 tokenizer API.

FossilOrigin-Name: fc71868496f45f9c7a79ed2bf2d164a7c4718ce1

diff --git a/ext/fts5/fts5.h b/ext/fts5/fts5.h
index 2e145d1068447a0416d82bec34be0c6b6e2a8241..c9eb91d4cca0450e4547388eed39e2ba604dcb7a 100644 (file)
--- a/ext/fts5/fts5.h
+++ b/ext/fts5/fts5.h
@@ -217,7 +217,7 @@ struct Fts5ExtensionApi {
   int (*xTokenize)(Fts5Context*, 
     const char *pText, int nText, /* Text to tokenize */
     void *pCtx,                   /* Context passed to xToken() */
-    int (*xToken)(void*, const char*, int, int, int, int)       /* Callback */
+    int (*xToken)(void*, int, const char*, int, int, int)       /* Callback */
   );
 
   int (*xPhraseCount)(Fts5Context*);
@@ -313,20 +313,25 @@ struct fts5_tokenizer {
       const char *pText, int nText, 
       int (*xToken)(
         void *pCtx,         /* Copy of 2nd argument to xTokenize() */
+        int tflags,         /* Mask of FTS5_TOKEN_* flags */
         const char *pToken, /* Pointer to buffer containing token */
         int nToken,         /* Size of token in bytes */
         int iStart,         /* Byte offset of token within input text */
-        int iEnd,           /* Byte offset of end of token within input text */
-        int iPos            /* Number of tokens before this one in input text */
+        int iEnd            /* Byte offset of end of token within input text */
       )
   );
 };
 
+/* Flags that may be passed as the third argument to xTokenize() */
 #define FTS5_TOKENIZE_QUERY     0x0001
 #define FTS5_TOKENIZE_PREFIX    0x0002
 #define FTS5_TOKENIZE_DOCUMENT  0x0004
 #define FTS5_TOKENIZE_AUX       0x0008
 
+/* Flags that may be passed by the tokenizer implementation back to FTS5
+** as the third argument to the supplied xToken callback. */
+#define FTS5_TOKEN_COLOCATED    0x0001      /* Same position as prev. token */
+
 /*
 ** END OF CUSTOM TOKENIZERS
 *************************************************************************/

diff --git a/ext/fts5/fts5Int.h b/ext/fts5/fts5Int.h
index 0357903c7b9f235398bccef40a8c3ae1a0f22644..c61f8c405236522a10c3d7a12a683046baa48e31 100644 (file)
--- a/ext/fts5/fts5Int.h
+++ b/ext/fts5/fts5Int.h
@@ -169,7 +169,7 @@ int sqlite3Fts5Tokenize(
   int flags,                      /* FTS5_TOKENIZE_* flags */
   const char *pText, int nText,   /* Text to tokenize */
   void *pCtx,                     /* Context passed to xToken() */
-  int (*xToken)(void*, const char*, int, int, int, int)    /* Callback */
+  int (*xToken)(void*, int, const char*, int, int, int)    /* Callback */
 );
 
 void sqlite3Fts5Dequote(char *z);

diff --git a/ext/fts5/fts5_aux.c b/ext/fts5/fts5_aux.c
index 0e608721f753714b3e25f076006ca0af8aa31a8e..2e33d5132f65dbcb3f9cac6bbb5dbd8f64a1fcb1 100644 (file)
--- a/ext/fts5/fts5_aux.c
+++ b/ext/fts5/fts5_aux.c
@@ -148,17 +148,18 @@ static void fts5HighlightAppend(
 */
 static int fts5HighlightCb(
   void *pContext,                 /* Pointer to HighlightContext object */
+  int tflags,                     /* Mask of FTS5_TOKEN_* flags */
   const char *pToken,             /* Buffer containing token */
   int nToken,                     /* Size of token in bytes */
   int iStartOff,                  /* Start offset of token */
-  int iEndOff,                    /* End offset of token */
-  int iPos
+  int iEndOff                     /* End offset of token */
 ){
   HighlightContext *p = (HighlightContext*)pContext;
   int rc = SQLITE_OK;
+  int iPos;
 
-  if( iPos<p->iPos ) return SQLITE_OK;
-  p->iPos = iPos+1;
+  if( tflags & FTS5_TOKEN_COLOCATED ) return SQLITE_OK;
+  iPos = p->iPos++;
 
   if( p->iRangeEnd>0 ){
     if( iPos<p->iRangeStart || iPos>p->iRangeEnd ) return SQLITE_OK;

diff --git a/ext/fts5/fts5_config.c b/ext/fts5/fts5_config.c
index 7987b8af76c0656a2d01a2c310890ade6d7193ba..74faf6dd30c6c0cf7efeeed345b3db05d0cf4927 100644 (file)
--- a/ext/fts5/fts5_config.c
+++ b/ext/fts5/fts5_config.c
@@ -648,7 +648,7 @@ int sqlite3Fts5Tokenize(
   int flags,                      /* FTS5_TOKENIZE_* flags */
   const char *pText, int nText,   /* Text to tokenize */
   void *pCtx,                     /* Context passed to xToken() */
-  int (*xToken)(void*, const char*, int, int, int, int)    /* Callback */
+  int (*xToken)(void*, int, const char*, int, int, int)    /* Callback */
 ){
   if( pText==0 ) return SQLITE_OK;
   return pConfig->pTokApi->xTokenize(

diff --git a/ext/fts5/fts5_expr.c b/ext/fts5/fts5_expr.c
index 46b4091e3797f07c0ae6cb968dff4361818c1604..9b52bb08c6c41b0539dff35f9a80192bc27a652f 100644 (file)
--- a/ext/fts5/fts5_expr.c
+++ b/ext/fts5/fts5_expr.c
@@ -1338,11 +1338,11 @@ struct TokenCtx {
 */
 static int fts5ParseTokenize(
   void *pContext,                 /* Pointer to Fts5InsertCtx object */
+  int tflags,                     /* Mask of FTS5_TOKEN_* flags */
   const char *pToken,             /* Buffer containing token */
   int nToken,                     /* Size of token in bytes */
   int iStart,                     /* Start offset of token */
-  int iEnd,                       /* End offset of token */
-  int iPos
+  int iEnd                        /* End offset of token */
 ){
   int rc = SQLITE_OK;
   const int SZALLOC = 8;
@@ -1350,6 +1350,8 @@ static int fts5ParseTokenize(
   Fts5ExprPhrase *pPhrase = pCtx->pPhrase;
   Fts5ExprTerm *pTerm;
 
+  if( tflags & FTS5_TOKEN_COLOCATED ) return rc;
+
   if( pPhrase==0 || (pPhrase->nTerm % SZALLOC)==0 ){
     Fts5ExprPhrase *pNew;
     int nNew = SZALLOC + (pPhrase ? pPhrase->nTerm : 0);

diff --git a/ext/fts5/fts5_main.c b/ext/fts5/fts5_main.c
index 5f0a90eadf6cb81535c92fd6354cc55cb74c58b9..4704fb890644a67f263b50ade0114d8166ea424c 100644 (file)
--- a/ext/fts5/fts5_main.c
+++ b/ext/fts5/fts5_main.c
@@ -1498,7 +1498,7 @@ static int fts5ApiTokenize(
   Fts5Context *pCtx, 
   const char *pText, int nText, 
   void *pUserData,
-  int (*xToken)(void*, const char*, int, int, int, int)
+  int (*xToken)(void*, int, const char*, int, int, int)
 ){
   Fts5Cursor *pCsr = (Fts5Cursor*)pCtx;
   Fts5Table *pTab = (Fts5Table*)(pCsr->base.pVtab);
@@ -1657,14 +1657,16 @@ static int fts5ApiColumnText(
 
 static int fts5ColumnSizeCb(
   void *pContext,                 /* Pointer to int */
+  int tflags,
   const char *pToken,             /* Buffer containing token */
   int nToken,                     /* Size of token in bytes */
   int iStart,                     /* Start offset of token */
-  int iEnd,                       /* End offset of token */
-  int iPos
+  int iEnd                        /* End offset of token */
 ){
   int *pCnt = (int*)pContext;
-  *pCnt = iPos+1;
+  if( (tflags & FTS5_TOKEN_COLOCATED)==0 ){
+    (*pCnt)++;
+  }
   return SQLITE_OK;
 }
 

diff --git a/ext/fts5/fts5_storage.c b/ext/fts5/fts5_storage.c
index 6b1c69c0082fbec3bd9c4f98cd24b4b5c88d55f2..d0572e60e65e4494494631ba9adce8bd6e4e1741 100644 (file)
--- a/ext/fts5/fts5_storage.c
+++ b/ext/fts5/fts5_storage.c
@@ -359,17 +359,18 @@ struct Fts5InsertCtx {
 */
 static int fts5StorageInsertCallback(
   void *pContext,                 /* Pointer to Fts5InsertCtx object */
+  int tflags,
   const char *pToken,             /* Buffer containing token */
   int nToken,                     /* Size of token in bytes */
   int iStart,                     /* Start offset of token */
-  int iEnd,                       /* End offset of token */
-  int iPos
+  int iEnd                        /* End offset of token */
 ){
   Fts5InsertCtx *pCtx = (Fts5InsertCtx*)pContext;
   Fts5Index *pIdx = pCtx->pStorage->pIndex;
-  assert( iPos+1>=pCtx->szCol );
-  pCtx->szCol = iPos+1;
-  return sqlite3Fts5IndexWrite(pIdx, pCtx->iCol, iPos, pToken, nToken);
+  if( (tflags & FTS5_TOKEN_COLOCATED)==0 ){
+    pCtx->szCol++;
+  }
+  return sqlite3Fts5IndexWrite(pIdx, pCtx->iCol, pCtx->szCol-1, pToken, nToken);
 }
 
 /*
@@ -844,17 +845,18 @@ struct Fts5IntegrityCtx {
 */
 static int fts5StorageIntegrityCallback(
   void *pContext,                 /* Pointer to Fts5InsertCtx object */
+  int tflags,
   const char *pToken,             /* Buffer containing token */
   int nToken,                     /* Size of token in bytes */
   int iStart,                     /* Start offset of token */
-  int iEnd,                       /* End offset of token */
-  int iPos
+  int iEnd                        /* End offset of token */
 ){
   Fts5IntegrityCtx *pCtx = (Fts5IntegrityCtx*)pContext;
-  assert( iPos+1>=pCtx->szCol );
-  pCtx->szCol = iPos+1;
+  if( (tflags & FTS5_TOKEN_COLOCATED)==0 ){
+    pCtx->szCol++;
+  }
   pCtx->cksum ^= sqlite3Fts5IndexCksum(
-      pCtx->pConfig, pCtx->iRowid, pCtx->iCol, iPos, pToken, nToken
+      pCtx->pConfig, pCtx->iRowid, pCtx->iCol, pCtx->szCol-1, pToken, nToken
   );
   return SQLITE_OK;
 }

diff --git a/ext/fts5/fts5_tcl.c b/ext/fts5/fts5_tcl.c
index fecbd41d32504ce7f49dd58c18bfbbeff82e7ab4..528ca958f25e359402baf8e4e6fc2178f595833e 100644 (file)
--- a/ext/fts5/fts5_tcl.c
+++ b/ext/fts5/fts5_tcl.c
@@ -141,8 +141,9 @@ struct F5tAuxData {
 
 static int xTokenizeCb(
   void *pCtx, 
+  int tflags,
   const char *zToken, int nToken, 
-  int iStart, int iEnd, int iPos
+  int iStart, int iEnd
 ){
   F5tFunction *p = (F5tFunction*)pCtx;
   Tcl_Obj *pEval = Tcl_DuplicateObj(p->pScript);
@@ -584,8 +585,9 @@ struct F5tTokenizeCtx {
 
 static int xTokenizeCb2(
   void *pCtx, 
+  int tflags,
   const char *zToken, int nToken, 
-  int iStart, int iEnd, int iPos
+  int iStart, int iEnd
 ){
   F5tTokenizeCtx *p = (F5tTokenizeCtx*)pCtx;
   if( p->bSubst ){
@@ -694,7 +696,7 @@ typedef struct F5tTokenizerModule F5tTokenizerInstance;
 
 struct F5tTokenizerContext {
   void *pCtx;
-  int (*xToken)(void*, const char*, int, int, int);
+  int (*xToken)(void*, int, const char*, int, int, int);
 };
 
 struct F5tTokenizerModule {
@@ -752,11 +754,11 @@ static int f5tTokenizerTokenize(
   void *pCtx,
   int flags,
   const char *pText, int nText, 
-  int (*xToken)(void*, const char*, int, int, int, int)
+  int (*xToken)(void*, int, const char*, int, int, int)
 ){
   F5tTokenizerInstance *pInst = (F5tTokenizerInstance*)p;
   void *pOldCtx;
-  int (*xOldToken)(void*, const char*, int, int, int);
+  int (*xOldToken)(void*, int, const char*, int, int, int);
   Tcl_Obj *pEval;
   int rc;
 
@@ -813,7 +815,7 @@ static int f5tTokenizerReturn(
     return TCL_ERROR;
   }
 
-  rc = p->xToken(p->pCtx, zToken, nToken, iStart, iEnd);
+  rc = p->xToken(p->pCtx, 0, zToken, nToken, iStart, iEnd);
   Tcl_SetResult(interp, (char*)sqlite3ErrName(rc), TCL_VOLATILE);
   return TCL_OK;
 }

diff --git a/ext/fts5/fts5_tokenize.c b/ext/fts5/fts5_tokenize.c
index d2477a050b6d245a372e0cbe8b7d64a32f7cb1ef..6b1129c24d775472586ecdc3b146d4c750304ef1 100644 (file)
--- a/ext/fts5/fts5_tokenize.c
+++ b/ext/fts5/fts5_tokenize.c
@@ -118,13 +118,12 @@ static int fts5AsciiTokenize(
   void *pCtx,
   int flags,
   const char *pText, int nText,
-  int (*xToken)(void*, const char*, int nToken, int iStart, int iEnd, int iPos)
+  int (*xToken)(void*, int, const char*, int nToken, int iStart, int iEnd)
 ){
   AsciiTokenizer *p = (AsciiTokenizer*)pTokenizer;
   int rc = SQLITE_OK;
   int ie;
   int is = 0;
-  int iPos = 0;
 
   char aFold[64];
   int nFold = sizeof(aFold);
@@ -160,7 +159,7 @@ static int fts5AsciiTokenize(
     asciiFold(pFold, &pText[is], nByte);
 
     /* Invoke the token callback */
-    rc = xToken(pCtx, pFold, nByte, is, ie, iPos++);
+    rc = xToken(pCtx, 0, pFold, nByte, is, ie);
     is = ie+1;
   }
   
@@ -389,12 +388,11 @@ static int fts5UnicodeTokenize(
   void *pCtx,
   int flags,
   const char *pText, int nText,
-  int (*xToken)(void*, const char*, int nToken, int iStart, int iEnd, int iPos)
+  int (*xToken)(void*, int, const char*, int nToken, int iStart, int iEnd)
 ){
   Unicode61Tokenizer *p = (Unicode61Tokenizer*)pTokenizer;
   int rc = SQLITE_OK;
   unsigned char *a = p->aTokenChar;
-  int iPos = 0;
 
   unsigned char *zTerm = (unsigned char*)&pText[nText];
   unsigned char *zCsr = (unsigned char *)pText;
@@ -479,7 +477,7 @@ static int fts5UnicodeTokenize(
     }
 
     /* Invoke the token callback */
-    rc = xToken(pCtx, aFold, zOut-aFold, is, ie, iPos++);
+    rc = xToken(pCtx, 0, aFold, zOut-aFold, is, ie); 
   }
   
  tokenize_done:
@@ -557,7 +555,7 @@ static int fts5PorterCreate(
 typedef struct PorterContext PorterContext;
 struct PorterContext {
   void *pCtx;
-  int (*xToken)(void*, const char*, int, int, int, int);
+  int (*xToken)(void*, int, const char*, int, int, int);
   char *aBuf;
 };
 
@@ -1122,11 +1120,11 @@ static void fts5PorterStep1A(char *aBuf, int *pnBuf){
 
 static int fts5PorterCb(
   void *pCtx, 
+  int tflags,
   const char *pToken, 
   int nToken, 
   int iStart, 
-  int iEnd,
-  int iPos
+  int iEnd
 ){
   PorterContext *p = (PorterContext*)pCtx;
 
@@ -1180,10 +1178,10 @@ static int fts5PorterCb(
     nBuf--;
   }
 
-  return p->xToken(p->pCtx, aBuf, nBuf, iStart, iEnd, iPos);
+  return p->xToken(p->pCtx, tflags, aBuf, nBuf, iStart, iEnd);
 
  pass_through:
-  return p->xToken(p->pCtx, pToken, nToken, iStart, iEnd, iPos);
+  return p->xToken(p->pCtx, tflags, pToken, nToken, iStart, iEnd);
 }
 
 /*
@@ -1194,7 +1192,7 @@ static int fts5PorterTokenize(
   void *pCtx,
   int flags,
   const char *pText, int nText,
-  int (*xToken)(void*, const char*, int nToken, int iStart, int iEnd, int iPos)
+  int (*xToken)(void*, int, const char*, int nToken, int iStart, int iEnd)
 ){
   PorterTokenizer *p = (PorterTokenizer*)pTokenizer;
   PorterContext sCtx;

diff --git a/manifest b/manifest
index 824f32d4c09c237b4ee8308135075833f1862932..964f295f100749ecfb0068c878123293910b7338 100644 (file)
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Change\sthe\sfts5\stokenizer\sAPI\sto\sallow\smore\sthan\sone\stoken\sto\soccupy\sa\ssingle\sposition\swithin\sa\sdocument.
-D 2015-08-28T19:56:47.300
+C Another\schange\sto\sthe\sfts5\stokenizer\sAPI.
+D 2015-08-29T15:44:27.938
 F Makefile.arm-wince-mingw32ce-gcc d6df77f1f48d690bd73162294bbba7f59507c72f
 F Makefile.in e2218eb228374422969de7b1680eda6864affcef
 F Makefile.linux-gcc 91d710bdc4998cb015f39edf3cb314ec4f4d7e23
@@ -105,19 +105,19 @@ F ext/fts3/unicode/UnicodeData.txt cd07314edb62d49fde34debdaf92fa2aa69011e7
 F ext/fts3/unicode/mkunicode.tcl 95cf7ec186e48d4985e433ff8a1c89090a774252
 F ext/fts3/unicode/parseunicode.tcl da577d1384810fb4e2b209bf3313074353193e95
 F ext/fts5/extract_api_docs.tcl 06583c935f89075ea0b32f85efa5dd7619fcbd03
-F ext/fts5/fts5.h b9dfb487ada3caab4400210609b8309b71a4fb4d
-F ext/fts5/fts5Int.h b0cfe44ec9451f766b77c4e5f771e7919c6dc8d5
-F ext/fts5/fts5_aux.c 7d0e275ee94ad7afdd4208d6b071b4319e8f9ca0
+F ext/fts5/fts5.h 0784692f406588e6c90e13a78e1f36e7e3236e42
+F ext/fts5/fts5Int.h 9fd31e682acae32806f77e7c3b543c4294274c92
+F ext/fts5/fts5_aux.c 7a307760a9c57c750d043188ec0bad59f5b5ec7e
 F ext/fts5/fts5_buffer.c 80f9ba4431848cb857e3d2158f5280093dcd8015
-F ext/fts5/fts5_config.c ab81c8ccff6c0fb79f21c369e18e8e0dec365ec5
-F ext/fts5/fts5_expr.c f53917b6e68dee62e4c525466edacacf82eb7cbc
+F ext/fts5/fts5_config.c 80b61fd2c6844b64a3e72a64572d50a812da9384
+F ext/fts5/fts5_expr.c 7ea46f676491989069d31ae1f75c9439b0858711
 F ext/fts5/fts5_hash.c 4bf4b99708848357b8a2b5819e509eb6d3df9246
 F ext/fts5/fts5_index.c 076c4995bf06a6d1559a6e31f9a86b90f2105374
-F ext/fts5/fts5_main.c 7afdb84ac40b0e5bbb920a07a5cd5e062963816c
-F ext/fts5/fts5_storage.c 9c263323479a4aa554738e421813cd05615d379c
-F ext/fts5/fts5_tcl.c 41e2d6b455547a157085fd35fd59d4fd890dc7d3
+F ext/fts5/fts5_main.c b00834ac543431dc35edbe18018b4befe0c7fd42
+F ext/fts5/fts5_storage.c 9820e7b53ea12baf3c818485efd66346b73030c3
+F ext/fts5/fts5_tcl.c 058f8da51964458e9859edfc1ee13b1863edaeae
 F ext/fts5/fts5_test_mi.c 80a9e86fb4c5b6b58f8fefac05e9b96d1a6574e1
-F ext/fts5/fts5_tokenize.c 07a894410bc074685ddc0a9d89b5e7bf57ea4482
+F ext/fts5/fts5_tokenize.c 710541513ecf3fe6d9365326fc85aee6efe97229
 F ext/fts5/fts5_unicode2.c 78273fbd588d1d9bd0a7e4e0ccc9207348bae33c
 F ext/fts5/fts5_varint.c 3f86ce09cab152e3d45490d7586b7ed2e40c13f1
 F ext/fts5/fts5_vocab.c 4622e0b7d84a488a1585aaa56eb214ee67a988bc
@@ -1380,10 +1380,7 @@ F tool/vdbe_profile.tcl 67746953071a9f8f2f668b73fe899074e2c6d8c1
 F tool/warnings-clang.sh f6aa929dc20ef1f856af04a730772f59283631d4
 F tool/warnings.sh 48bd54594752d5be3337f12c72f28d2080cb630b
 F tool/win/sqlite.vsix deb315d026cc8400325c5863eef847784a219a2f
-P 0fdc36fe35ae2fc8e9688fe6c53437f4d47502d9
-R 694c0e23ba08ed9bcc32d2c502ed8f13
-T *branch * fts5-incompatible
-T *sym-fts5-incompatible *
-T -sym-trunk *
+P 90b85b42f2b2dd3e939b129b7df2b822a05e243d
+R f343432805e01f14633e088d58d566cf
 U dan
-Z 745a50831400d199b74f44c2476ec260
+Z e4288542e4294b868813263b0597051d

diff --git a/manifest.uuid b/manifest.uuid
index 91a79bb98f727318e27f3955b2e422aa91f62650..c7283c3a07e5edec1ba0764fd38fbc3c3743c7b0 100644 (file)
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-90b85b42f2b2dd3e939b129b7df2b822a05e243d
\ No newline at end of file
+fc71868496f45f9c7a79ed2bf2d164a7c4718ce1
\ No newline at end of file