1 files changed, 61 insertions, 135 deletions

diff --git a/mysys/charset.c b/mysys/charset.c
index d801fcdbd76..1388fc40c6d 100644
--- a/mysys/charset.c
+++ b/mysys/charset.c
@@ -21,16 +21,6 @@
 #include <my_dir.h>
 #include <my_xml.h>
 
-typedef struct
-{
-  int		nchars;
-  MY_UNI_IDX	uidx;
-} uni_idx;
-
-#define PLANE_SIZE	0x100
-#define PLANE_NUM	0x100
-#define PLANE_NUMBER(x)	(((x)>>8) % PLANE_NUM)
-
 
 /*
   The code below implements this functionality:
@@ -48,32 +38,21 @@ my_bool my_charset_same(CHARSET_INFO *cs1, CHARSET_INFO *cs2)
 }
 
 
-static void set_max_sort_char(CHARSET_INFO *cs)
-{
-  uchar max_char;
-  uint  i;
-  
-  if (!cs->sort_order)
-    return;
-  
-  max_char=cs->sort_order[(uchar) cs->max_sort_char];
-  for (i= 0; i < 256; i++)
-  {
-    if ((uchar) cs->sort_order[i] > max_char)
-    {
-      max_char=(uchar) cs->sort_order[i];
-      cs->max_sort_char= i;
-    }
-  }
-}
-
-
-static void init_state_maps(CHARSET_INFO *cs)
+static my_bool init_state_maps(CHARSET_INFO *cs)
 {
   uint i;
-  uchar *state_map= cs->state_map;
-  uchar *ident_map= cs->ident_map;
+  uchar *state_map;
+  uchar *ident_map;
 
+  if (!(cs->state_map= (uchar*) my_once_alloc(256, MYF(MY_WME))))
+    return 1;
+    
+  if (!(cs->ident_map= (uchar*) my_once_alloc(256, MYF(MY_WME))))
+    return 1;
+
+  state_map= cs->state_map;
+  ident_map= cs->ident_map;
+  
   /* Fill state_map with states to get a faster parser */
   for (i=0; i < 256 ; i++)
   {
@@ -120,6 +99,7 @@ static void init_state_maps(CHARSET_INFO *cs)
   state_map[(uchar)'x']= state_map[(uchar)'X']= (uchar) MY_LEX_IDENT_OR_HEX;
   state_map[(uchar)'b']= state_map[(uchar)'b']= (uchar) MY_LEX_IDENT_OR_BIN;
   state_map[(uchar)'n']= state_map[(uchar)'N']= (uchar) MY_LEX_IDENT_OR_NCHAR;
+  return 0;
 }
 
 
@@ -131,98 +111,11 @@ static void simple_cs_init_functions(CHARSET_INFO *cs)
     cs->coll= &my_collation_8bit_simple_ci_handler;
   
   cs->cset= &my_charset_8bit_handler;
-  cs->mbminlen= 1;
-  cs->mbmaxlen= 1;
 }
 
 
-static int pcmp(const void * f, const void * s)
-{
-  const uni_idx *F= (const uni_idx*) f;
-  const uni_idx *S= (const uni_idx*) s;
-  int res;
-
-  if (!(res=((S->nchars)-(F->nchars))))
-    res=((F->uidx.from)-(S->uidx.to));
-  return res;
-}
 
-
-static my_bool create_fromuni(CHARSET_INFO *cs)
-{
-  uni_idx	idx[PLANE_NUM];
-  int		i,n;
-  
-  /* Clear plane statistics */
-  bzero(idx,sizeof(idx));
-  
-  /* Count number of characters in each plane */
-  for (i=0; i< 0x100; i++)
-  {
-    uint16 wc=cs->tab_to_uni[i];
-    int pl= PLANE_NUMBER(wc);
-    
-    if (wc || !i)
-    {
-      if (!idx[pl].nchars)
-      {
-        idx[pl].uidx.from=wc;
-        idx[pl].uidx.to=wc;
-      }else
-      {
-        idx[pl].uidx.from=wc<idx[pl].uidx.from?wc:idx[pl].uidx.from;
-        idx[pl].uidx.to=wc>idx[pl].uidx.to?wc:idx[pl].uidx.to;
-      }
-      idx[pl].nchars++;
-    }
-  }
-  
-  /* Sort planes in descending order */
-  qsort(&idx,PLANE_NUM,sizeof(uni_idx),&pcmp);
-  
-  for (i=0; i < PLANE_NUM; i++)
-  {
-    int ch,numchars;
-    
-    /* Skip empty plane */
-    if (!idx[i].nchars)
-      break;
-    
-    numchars=idx[i].uidx.to-idx[i].uidx.from+1;
-    if (!(idx[i].uidx.tab=(uchar*) my_once_alloc(numchars *
-						 sizeof(*idx[i].uidx.tab),
-						 MYF(MY_WME))))
-      return TRUE;
-
-    bzero(idx[i].uidx.tab,numchars*sizeof(*idx[i].uidx.tab));
-    
-    for (ch=1; ch < PLANE_SIZE; ch++)
-    {
-      uint16 wc=cs->tab_to_uni[ch];
-      if (wc >= idx[i].uidx.from && wc <= idx[i].uidx.to && wc)
-      {
-        int ofs= wc - idx[i].uidx.from;
-        idx[i].uidx.tab[ofs]= ch;
-      }
-    }
-  }
-  
-  /* Allocate and fill reverse table for each plane */
-  n=i;
-  if (!(cs->tab_from_uni= (MY_UNI_IDX*) my_once_alloc(sizeof(MY_UNI_IDX)*(n+1),
-						      MYF(MY_WME))))
-    return TRUE;
-
-  for (i=0; i< n; i++)
-    cs->tab_from_uni[i]= idx[i].uidx;
-  
-  /* Set end-of-list marker */
-  bzero(&cs->tab_from_uni[i],sizeof(MY_UNI_IDX));
-  return FALSE;
-}
-
-
-static int simple_cs_copy_data(CHARSET_INFO *to, CHARSET_INFO *from)
+static int cs_copy_data(CHARSET_INFO *to, CHARSET_INFO *from)
 {
   to->number= from->number ? from->number : to->number;
 
@@ -244,7 +137,8 @@ static int simple_cs_copy_data(CHARSET_INFO *to, CHARSET_INFO *from)
 					     MY_CS_CTYPE_TABLE_SIZE,
 					     MYF(MY_WME))))
       goto err;
-    init_state_maps(to);
+    if (init_state_maps(to))
+      goto err;
   }
   if (from->to_lower)
     if (!(to->to_lower= (uchar*) my_once_memdup((char*) from->to_lower,
@@ -263,7 +157,7 @@ static int simple_cs_copy_data(CHARSET_INFO *to, CHARSET_INFO *from)
 						  MY_CS_SORT_ORDER_TABLE_SIZE,
 						  MYF(MY_WME))))
       goto err;
-    set_max_sort_char(to);
+
   }
   if (from->tab_to_uni)
   {
@@ -271,11 +165,10 @@ static int simple_cs_copy_data(CHARSET_INFO *to, CHARSET_INFO *from)
     if (!(to->tab_to_uni= (uint16*)  my_once_memdup((char*)from->tab_to_uni,
 						    sz, MYF(MY_WME))))
       goto err;
-    if (create_fromuni(to))
-      goto err;
   }
-  to->mbminlen= 1;
-  to->mbmaxlen= 1;
+  if (from->tailoring)
+    if (!(to->tailoring= my_once_strdup(from->tailoring,MYF(MY_WME))))
+      goto err;
 
   return 0;
 
@@ -284,6 +177,7 @@ err:
 }
 
 
+
 static my_bool simple_cs_is_full(CHARSET_INFO *cs)
 {
   return ((cs->csname && cs->tab_to_uni && cs->ctype && cs->to_upper &&
@@ -315,14 +209,34 @@ static int add_collation(CHARSET_INFO *cs)
     
     if (!(all_charsets[cs->number]->state & MY_CS_COMPILED))
     {
-      simple_cs_init_functions(all_charsets[cs->number]);
-      if (simple_cs_copy_data(all_charsets[cs->number],cs))
-	return MY_XML_ERROR;
-      if (simple_cs_is_full(all_charsets[cs->number]))
+      CHARSET_INFO *new= all_charsets[cs->number];
+      if (cs_copy_data(all_charsets[cs->number],cs))
+        return MY_XML_ERROR;
+
+      if (!strcmp(cs->csname,"ucs2") )
+      {
+#ifdef HAVE_CHARSET_ucs2
+        new->cset= my_charset_ucs2_general_uca.cset;
+        new->coll= my_charset_ucs2_general_uca.coll;
+        new->strxfrm_multiply= my_charset_ucs2_general_uca.strxfrm_multiply;
+        new->min_sort_char= my_charset_ucs2_general_uca.min_sort_char;
+        new->max_sort_char= my_charset_ucs2_general_uca.max_sort_char;
+        new->mbminlen= 2;
+        new->mbmaxlen= 2;
+        new->state |= MY_CS_AVAILABLE | MY_CS_LOADED;
+#endif        
+      }
+      else
       {
-        all_charsets[cs->number]->state |= MY_CS_LOADED;
+        simple_cs_init_functions(all_charsets[cs->number]);
+        new->mbminlen= 1;
+        new->mbmaxlen= 1;
+        if (simple_cs_is_full(all_charsets[cs->number]))
+        {
+          all_charsets[cs->number]->state |= MY_CS_LOADED;
+        }
+        all_charsets[cs->number]->state|= MY_CS_AVAILABLE;
       }
-      all_charsets[cs->number]->state|= MY_CS_AVAILABLE;
     }
     else
     {
@@ -432,6 +346,10 @@ void add_compiled_collation(CHARSET_INFO *cs)
   cs->state|= MY_CS_AVAILABLE;
 }
 
+static void *cs_alloc(uint size)
+{
+  return my_once_alloc(size, MYF(MY_WME));
+}
 
 
 #ifdef __NETWARE__
@@ -465,9 +383,9 @@ static my_bool init_available_charsets(myf myflags)
     {
       if (*cs)
       {
-        set_max_sort_char(*cs);
         if (cs[0]->ctype)
-          init_state_maps(*cs);
+          if (init_state_maps(*cs))
+            *cs= NULL;
       }
     }
     
@@ -551,6 +469,14 @@ static CHARSET_INFO *get_internal_charset(uint cs_number, myf flags)
     }
     cs= (cs->state & MY_CS_AVAILABLE) ? cs : NULL;
   }
+  if (cs && !(cs->state & MY_CS_READY))
+  {
+    if ((cs->cset->init && cs->cset->init(cs, cs_alloc)) ||
+        (cs->coll->init && cs->coll->init(cs, cs_alloc)))
+      cs= NULL;
+    else
+      cs->state|= MY_CS_READY;
+  }
   pthread_mutex_unlock(&THR_LOCK_charset);
   return cs;
 }