Merge 4.1 -> 5.0

8 files changed, 61 insertions, 25 deletions

diff --git a/myisam/ft_boolean_search.c b/myisam/ft_boolean_search.c
index 1d4bfee86a4..5f88122e730 100644
--- a/myisam/ft_boolean_search.c
+++ b/myisam/ft_boolean_search.c
@@ -204,7 +204,8 @@ static int _ft2_search(FTB *ftb, FTB_WORD *ftbw, my_bool init_search)
 {
   int r;
   uint off;
-  int subkeys;
+  int subkeys=1;
+  my_bool can_go_down;
   MI_INFO *info=ftb->info;
 
   if (init_search)
@@ -221,6 +222,24 @@ static int _ft2_search(FTB *ftb, FTB_WORD *ftbw, my_bool init_search)
     r=_mi_search(info, ftbw->keyinfo, (uchar*) ftbw->word+ftbw->off,
                    USE_WHOLE_KEY, SEARCH_BIGGER, ftbw->key_root);
   }
+
+  can_go_down=(!ftbw->off && (init_search || (ftbw->flags & FTB_FLAG_TRUNC)));
+  /* Skip rows inserted by concurrent insert */
+  while (!r)
+  {
+    if (can_go_down)
+    {
+      /* going down ? */
+      off=info->lastkey_length-HA_FT_WLEN-info->s->base.rec_reflength;
+      subkeys=ft_sintXkorr(info->lastkey+off);
+    }
+    if (subkeys<0 || info->lastpos < info->state->data_file_length)
+      break;
+    r= _mi_search_next(info, ftbw->keyinfo, info->lastkey,
+                       info->lastkey_length,
+		       SEARCH_BIGGER, ftbw->key_root);
+  }
+
   if (!r && !ftbw->off)
   {
     r= mi_compare_text(ftb->charset,
@@ -260,22 +279,19 @@ static int _ft2_search(FTB *ftb, FTB_WORD *ftbw, my_bool init_search)
 
   /* matching key found */
   memcpy(ftbw->word+ftbw->off, info->lastkey, info->lastkey_length);
-  if (!ftbw->off && (init_search || (ftbw->flags & FTB_FLAG_TRUNC)))
+  /* going down ? */
+  if (subkeys<0)
   {
-    /* going down ? */
-    get_key_full_length_rdonly(off, info->lastkey);
-    subkeys=ft_sintXkorr(info->lastkey+off);
-    if (subkeys<0)
-    {
-      /* yep, going down, to the second-level tree */
-      /* TODO here: subkey-based optimization */
-      ftbw->off=off;
-      ftbw->key_root=info->lastpos;
-      ftbw->keyinfo=& info->s->ft2_keyinfo;
-      r=_mi_search_first(info, ftbw->keyinfo, ftbw->key_root);
-      DBUG_ASSERT(r==0);  /* found something */
-      memcpy(ftbw->word+off, info->lastkey, info->lastkey_length);
-    }
+    /*
+      yep, going down, to the second-level tree
+      TODO here: subkey-based optimization
+    */
+    ftbw->off=off;
+    ftbw->key_root=info->lastpos;
+    ftbw->keyinfo=& info->s->ft2_keyinfo;
+    r=_mi_search_first(info, ftbw->keyinfo, ftbw->key_root);
+    DBUG_ASSERT(r==0);  /* found something */
+    memcpy(ftbw->word+off, info->lastkey, info->lastkey_length);
   }
   ftbw->docid[0]=info->lastpos;
   return 0;
diff --git a/myisam/ft_nlq_search.c b/myisam/ft_nlq_search.c
index 03875abe7b0..9b42d51ab55 100644
--- a/myisam/ft_nlq_search.c
+++ b/myisam/ft_nlq_search.c
@@ -88,7 +88,14 @@ static int walk_and_match(FT_WORD *word, uint32 count, ALL_IN_ONE *aio)
   keylen-=HA_FT_WLEN;
   doc_cnt=0;
 
-  r=_mi_search(info, keyinfo, keybuff, keylen, SEARCH_FIND, key_root);
+  /* Skip rows inserted by current inserted */
+  for (r=_mi_search(info, keyinfo, keybuff, keylen, SEARCH_FIND, key_root) ;
+       (subkeys=ft_sintXkorr(info->lastkey+info->lastkey_length-extra)) > 0 &&
+       !r && info->lastpos >= info->state->data_file_length ;
+       r= _mi_search_next(info, keyinfo, info->lastkey,
+                          info->lastkey_length, SEARCH_BIGGER, key_root))
+    ;
+
   info->update|= HA_STATE_AKTIV;              /* for _mi_test_if_changed() */
 
   while (!r && gweight)
@@ -99,7 +106,6 @@ static int walk_and_match(FT_WORD *word, uint32 count, ALL_IN_ONE *aio)
                         info->lastkey_length-extra-1, keybuff+1,keylen-1,0,0))
      break;
 
-    subkeys=ft_sintXkorr(info->lastkey+info->lastkey_length-extra);
     if (subkeys<0)
     {
       if (doc_cnt)
@@ -113,7 +119,7 @@ static int walk_and_match(FT_WORD *word, uint32 count, ALL_IN_ONE *aio)
       key_root=info->lastpos;
       keylen=0;
       r=_mi_search_first(info, keyinfo, key_root);
-      continue;
+      goto do_skip;
     }
 #if HA_FT_WTYPE == HA_KEYTYPE_FLOAT
     tmp_weight=*(float*)&subkeys;
@@ -151,6 +157,12 @@ static int walk_and_match(FT_WORD *word, uint32 count, ALL_IN_ONE *aio)
     else
 	r=_mi_search(info, keyinfo, info->lastkey, info->lastkey_length,
                      SEARCH_BIGGER, key_root);
+do_skip:
+    while ((subkeys=ft_sintXkorr(info->lastkey+info->lastkey_length-extra)) > 0 &&
+           !r && info->lastpos >= info->state->data_file_length)
+      r= _mi_search_next(info, keyinfo, info->lastkey, info->lastkey_length,
+                         SEARCH_BIGGER, key_root);
+
   }
   word->weight=gweight;
 
diff --git a/myisam/mi_create.c b/myisam/mi_create.c
index 5682862c39a..838d647ea56 100644
--- a/myisam/mi_create.c
+++ b/myisam/mi_create.c
@@ -206,9 +206,9 @@ int mi_create(const char *name,uint keys,MI_KEYDEF *keydefs,
 					  3 : 0)));
 
   if (options & (HA_OPTION_COMPRESS_RECORD | HA_OPTION_PACK_RECORD))
-    pointer=mi_get_pointer_length(ci->data_file_length,4);
+    pointer=mi_get_pointer_length(ci->data_file_length,myisam_data_pointer_size);
   else
-    pointer=mi_get_pointer_length(ci->max_rows,4);
+    pointer=mi_get_pointer_length(ci->max_rows,myisam_data_pointer_size);
   if (!(max_rows=(ulonglong) ci->max_rows))
     max_rows= ((((ulonglong) 1 << (pointer*8)) -1) / min_pack_length);
 
diff --git a/myisam/mi_info.c b/myisam/mi_info.c
index f4eace198f9..cf63ef63618 100644
--- a/myisam/mi_info.c
+++ b/myisam/mi_info.c
@@ -75,7 +75,8 @@ int mi_status(MI_INFO *info, register MI_ISAMINFO *x, uint flag)
     x->filenr	 = info->dfile;
     x->options	 = share->options;
     x->create_time=share->state.create_time;
-    x->reflength= mi_get_pointer_length(share->base.max_data_file_length,4);
+    x->reflength= mi_get_pointer_length(share->base.max_data_file_length,
+                                        myisam_data_pointer_size);
     x->record_offset= ((share->options &
 			(HA_OPTION_PACK_RECORD | HA_OPTION_COMPRESS_RECORD)) ?
 		       0L : share->base.pack_reclength);
diff --git a/myisam/mi_static.c b/myisam/mi_static.c
index a36c38e3c01..f7d008ffbb7 100644
--- a/myisam/mi_static.c
+++ b/myisam/mi_static.c
@@ -41,6 +41,7 @@ my_bool myisam_concurrent_insert=0;
 my_off_t myisam_max_extra_temp_length= MI_MAX_TEMP_LENGTH;
 my_off_t myisam_max_temp_length= MAX_FILE_SIZE;
 ulong    myisam_bulk_insert_tree_size=8192*1024;
+ulong    myisam_data_pointer_size=4;
 
 /*
   read_vec[] is used for converting between P_READ_KEY.. and SEARCH_
diff --git a/myisam/myisam_ftdump.c b/myisam/myisam_ftdump.c
index 8ab6a7600b2..075a0fd19e8 100644
--- a/myisam/myisam_ftdump.c
+++ b/myisam/myisam_ftdump.c
@@ -83,7 +83,7 @@ int main(int argc,char *argv[])
 
   {
     char *end;
-    inx= strtoll(argv[1], &end, 10);
+    inx= (uint) strtoll(argv[1], &end, 10);
     if (*end)
       usage();
   }
diff --git a/myisam/myisamchk.c b/myisam/myisamchk.c
index a96632b4f3c..cb7c8ae8288 100644
--- a/myisam/myisamchk.c
+++ b/myisam/myisamchk.c
@@ -154,7 +154,8 @@ enum options_mc {
   OPT_KEY_CACHE_BLOCK_SIZE, OPT_MYISAM_BLOCK_SIZE,
   OPT_READ_BUFFER_SIZE, OPT_WRITE_BUFFER_SIZE, OPT_SORT_BUFFER_SIZE,
   OPT_SORT_KEY_BLOCKS, OPT_DECODE_BITS, OPT_FT_MIN_WORD_LEN,
-  OPT_FT_MAX_WORD_LEN, OPT_FT_MAX_WORD_LEN_FOR_SORT, OPT_MAX_RECORD_LENGTH
+  OPT_FT_MAX_WORD_LEN, OPT_FT_MAX_WORD_LEN_FOR_SORT, OPT_FT_STOPWORD_FILE,
+  OPT_MAX_RECORD_LENGTH
 };
 
 static struct my_option my_long_options[] =
@@ -327,7 +328,11 @@ static struct my_option my_long_options[] =
   { "ft_max_word_len", OPT_FT_MAX_WORD_LEN, "", (gptr*) &ft_max_word_len,
     (gptr*) &ft_max_word_len, 0, GET_ULONG, REQUIRED_ARG, HA_FT_MAXCHARLEN, 10,
     HA_FT_MAXCHARLEN, 0, 1, 0},
-  { 0, 0, 0, 0, 0, 0, GET_NO_ARG, NO_ARG, 0, 0, 0, 0, 0, 0}
+  { 0, 0, 0, 0, 0, 0, GET_NO_ARG, NO_ARG, 0, 0, 0, 0, 0, 0},
+  { "ft_stopword_file", OPT_FT_STOPWORD_FILE,
+    "Use stopwords from this file instead of built-in list.",
+    (gptr*) &ft_stopword_file, (gptr*) &ft_stopword_file, 0, GET_STR,
+    REQUIRED_ARG, 0, 0, 0, 0, 0, 0}
 };
 
 
diff --git a/myisam/myisamdef.h b/myisam/myisamdef.h
index 987a9dfc39e..72260699d1e 100644
--- a/myisam/myisamdef.h
+++ b/myisam/myisamdef.h
@@ -417,6 +417,7 @@ typedef struct st_mi_sort_param
 #define MI_MIN_SIZE_BULK_INSERT_TREE 16384             /* this is per key */
 #define MI_MIN_ROWS_TO_USE_BULK_INSERT 100
 #define MI_MIN_ROWS_TO_DISABLE_INDEXES 100
+#define MI_MIN_ROWS_TO_USE_WRITE_CACHE 10
 
 /* The UNIQUE check is done with a hashed long key */