MDEV-23162 Improve Protocol performance for numeric data

Avoid character set conversion for numeric data in this scenario:

SET NAMES utf8;
CREATE OR REPLACE TABLE t1 (a INT);
INSERT INTO t1 VALUES (1);
SELECT a FROM t1;

sql/protocol.cc

@@ -853,7 +853,7 @@ bool Protocol_text::store_field_metadata(const THD * thd,
   if (thd->client_capabilities & CLIENT_PROTOCOL_41)
   {
     const LEX_CSTRING def= {STRING_WITH_LEN("def")};
-    if (store_ident(def) ||
+    if (store_ident(def, MY_REPERTOIRE_ASCII) ||
         store_ident(field.db_name) ||
         store_ident(field.table_name) ||
         store_ident(field.org_table_name) ||
@@ -868,7 +868,8 @@ bool Protocol_text::store_field_metadata(const THD * thd,
         Don't apply character set conversion:
         extended metadata is a binary encoded data.
       */
-      if (store_lex_cstring(metadata.lex_cstring(), cs, &my_charset_bin))
+      if (store_lex_cstring(metadata.lex_cstring(), cs,
+                            MY_REPERTOIRE_UNICODE30, &my_charset_bin))
         return true;
     }
     if (packet->realloc(packet->length() + 12))
@@ -1182,10 +1183,12 @@ bool Protocol_text::store_null()
 */
 
 bool Protocol::store_string_aux(const char *from, size_t length,
-                                CHARSET_INFO *fromcs, CHARSET_INFO *tocs)
+                                CHARSET_INFO *fromcs,
+                                my_repertoire_t from_repertoire,
+                                CHARSET_INFO *tocs)
 {
   /* 'tocs' is set 0 when client issues SET character_set_results=NULL */
-  if (needs_conversion(fromcs, tocs))
+  if (needs_conversion(fromcs, from_repertoire, tocs))
   {
     /* Store with conversion */
     return net_store_data_cs((uchar*) from, length, fromcs, tocs);
@@ -1208,7 +1211,9 @@ bool Protocol::store_warning(const char *from, size_t length)
 
 
 bool Protocol_text::store_str(const char *from, size_t length,
-                              CHARSET_INFO *fromcs, CHARSET_INFO *tocs)
+                              CHARSET_INFO *fromcs,
+                              my_repertoire_t from_repertoire,
+                              CHARSET_INFO *tocs)
 {
 #ifndef DBUG_OFF
   DBUG_PRINT("info", ("Protocol_text::store field %u (%u): %.*b", field_pos,
@@ -1217,7 +1222,7 @@ bool Protocol_text::store_str(const char *from, size_t length,
   DBUG_ASSERT(valid_handler(field_pos, PROTOCOL_SEND_STRING));
   field_pos++;
 #endif
-  return store_string_aux(from, length, fromcs, tocs);
+  return store_string_aux(from, length, fromcs, from_repertoire, tocs);
 }
 
 
@@ -1330,7 +1335,8 @@ bool Protocol_text::store(Field *field)
     dbug_tmp_restore_column_map(table->read_set, old_map);
 #endif
 
-  return store_string_aux(str.ptr(), str.length(), str.charset(), tocs);
+  return store_string_aux(str.ptr(), str.length(), str.charset(),
+                          field->dtcollation().repertoire, tocs);
 }
 
 
@@ -1449,10 +1455,12 @@ void Protocol_binary::prepare_for_resend()
 
 
 bool Protocol_binary::store_str(const char *from, size_t length,
-                                CHARSET_INFO *fromcs, CHARSET_INFO *tocs)
+                                CHARSET_INFO *fromcs,
+                                my_repertoire_t from_repertoire,
+                                CHARSET_INFO *tocs)
 {
   field_pos++;
-  return store_string_aux(from, length, fromcs, tocs);
+  return store_string_aux(from, length, fromcs, from_repertoire, tocs);
 }
 
 bool Protocol_binary::store_null()
@@ -1515,6 +1523,7 @@ bool Protocol_binary::store_decimal(const my_decimal *d)
   StringBuffer<DECIMAL_MAX_STR_LENGTH> str;
   (void) d->to_string(&str);
   return store_str(str.ptr(), str.length(), str.charset(),
+                   MY_REPERTOIRE_ASCII,
                    thd->variables.character_set_results);
 }
 

sql/protocol.h

@@ -62,19 +62,25 @@ class Protocol
   MEM_ROOT *alloc;
 #endif
   bool needs_conversion(CHARSET_INFO *fromcs,
+                        my_repertoire_t from_repertoire,
                         CHARSET_INFO *tocs) const
   {
     // 'tocs' is set 0 when client issues SET character_set_results=NULL
     return tocs && !my_charset_same(fromcs, tocs) &&
            fromcs != &my_charset_bin &&
-           tocs != &my_charset_bin;
+           tocs != &my_charset_bin &&
+           (from_repertoire != MY_REPERTOIRE_ASCII ||
+           (fromcs->state & MY_CS_NONASCII) ||
+           (tocs->state & MY_CS_NONASCII));
   }
   /* 
     The following two are low-level functions that are invoked from
     higher-level store_xxx() funcs.  The data is stored into this->packet.
   */
   bool store_string_aux(const char *from, size_t length,
-                        CHARSET_INFO *fromcs, CHARSET_INFO *tocs);
+                        CHARSET_INFO *fromcs,
+                        my_repertoire_t from_repertoire,
+                        CHARSET_INFO *tocs);
 
   virtual bool send_ok(uint server_status, uint statement_warn_count,
                        ulonglong affected_rows, ulonglong last_insert_id,
@@ -131,7 +137,9 @@ class Protocol
   virtual bool store_longlong(longlong from, bool unsigned_flag)=0;
   virtual bool store_decimal(const my_decimal *)=0;
   virtual bool store_str(const char *from, size_t length,
-                         CHARSET_INFO *fromcs, CHARSET_INFO *tocs)=0;
+                         CHARSET_INFO *fromcs,
+                         my_repertoire_t from_repertoire,
+                         CHARSET_INFO *tocs)=0;
   virtual bool store(float from, uint32 decimals, String *buffer)=0;
   virtual bool store(double from, uint32 decimals, String *buffer)=0;
   virtual bool store(MYSQL_TIME *time, int decimals)=0;
@@ -141,19 +149,23 @@ class Protocol
 
   // Various useful wrappers for the virtual store*() methods.
   // Backward wrapper for store_str()
-  bool store(const char *from, size_t length, CHARSET_INFO *cs)
+  bool store(const char *from, size_t length, CHARSET_INFO *cs,
+             my_repertoire_t repertoire= MY_REPERTOIRE_UNICODE30)
   {
-    return store_str(from, length, cs, character_set_results());
+    return store_str(from, length, cs, repertoire, character_set_results());
   }
   bool store_lex_cstring(const LEX_CSTRING &s,
                          CHARSET_INFO *fromcs,
+                         my_repertoire_t from_repertoire,
                          CHARSET_INFO *tocs)
   {
-    return store_str(s.str, (uint) s.length, fromcs, tocs);
+    return store_str(s.str, (uint) s.length, fromcs, from_repertoire, tocs);
   }
-  bool store_ident(const LEX_CSTRING &s)
+  bool store_ident(const LEX_CSTRING &s,
+                   my_repertoire_t repertoire= MY_REPERTOIRE_UNICODE30)
   {
-    return store_lex_cstring(s, system_charset_info, character_set_results());
+    return store_lex_cstring(s, system_charset_info, repertoire,
+                             character_set_results());
   }
   // End of wrappers
 
@@ -202,7 +214,9 @@ class Protocol_text :public Protocol
   virtual bool store_longlong(longlong from, bool unsigned_flag);
   virtual bool store_decimal(const my_decimal *);
   virtual bool store_str(const char *from, size_t length,
-                         CHARSET_INFO *fromcs, CHARSET_INFO *tocs);
+                         CHARSET_INFO *fromcs,
+                         my_repertoire_t from_repertoire,
+                         CHARSET_INFO *tocs);
   virtual bool store(MYSQL_TIME *time, int decimals);
   virtual bool store_date(MYSQL_TIME *time);
   virtual bool store_time(MYSQL_TIME *time, int decimals);
@@ -246,7 +260,9 @@ class Protocol_binary :public Protocol
   virtual bool store_longlong(longlong from, bool unsigned_flag);
   virtual bool store_decimal(const my_decimal *);
   virtual bool store_str(const char *from, size_t length,
-                         CHARSET_INFO *fromcs, CHARSET_INFO *tocs);
+                         CHARSET_INFO *fromcs,
+                         my_repertoire_t from_repertoire,
+                         CHARSET_INFO *tocs);
   virtual bool store(MYSQL_TIME *time, int decimals);
   virtual bool store_date(MYSQL_TIME *time);
   virtual bool store_time(MYSQL_TIME *time, int decimals);
@@ -293,7 +309,8 @@ class Protocol_discard : public Protocol_text
   bool store_long(longlong) { return false; }
   bool store_longlong(longlong, bool) { return false; }
   bool store_decimal(const my_decimal *) { return false; }
-  bool store_str(const char *, size_t, CHARSET_INFO *, CHARSET_INFO *)
+  bool store_str(const char *, size_t, CHARSET_INFO *, my_repertoire_t,
+                 CHARSET_INFO *)
   {
     return false;
   }

sql/sql_prepare.cc

@@ -263,7 +263,9 @@ class Protocol_local :public Protocol
   virtual bool store_longlong(longlong from, bool unsigned_flag);
   virtual bool store_decimal(const my_decimal *);
   virtual bool store_str(const char *from, size_t length,
-                         CHARSET_INFO *fromcs, CHARSET_INFO *tocs);
+                         CHARSET_INFO *fromcs,
+                         my_repertoire_t from_repertoire,
+                         CHARSET_INFO *tocs);
   virtual bool store(MYSQL_TIME *time, int decimals);
   virtual bool store_date(MYSQL_TIME *time);
   virtual bool store_time(MYSQL_TIME *time, int decimals);
@@ -286,7 +288,9 @@ class Protocol_local :public Protocol
   virtual bool send_error(uint sql_errno, const char *err_msg, const char* sqlstate);
 private:
   bool store_string(const char *str, size_t length,
-                    CHARSET_INFO *src_cs, CHARSET_INFO *dst_cs);
+                    CHARSET_INFO *src_cs,
+                    my_repertoire_t src_repertoire,
+                    CHARSET_INFO *dst_cs);
 
   bool store_column(const void *data, size_t length);
   void opt_add_row_to_rset();
@@ -5269,12 +5273,14 @@ bool Protocol_local::store_column(const void *data, size_t length)
 
 bool
 Protocol_local::store_string(const char *str, size_t length,
-                             CHARSET_INFO *src_cs, CHARSET_INFO *dst_cs)
+                             CHARSET_INFO *src_cs,
+                             my_repertoire_t src_repertoire,
+                             CHARSET_INFO *dst_cs)
 {
   /* Store with conversion */
   uint error_unused;
 
-  if (needs_conversion(src_cs, dst_cs))
+  if (needs_conversion(src_cs, src_repertoire, dst_cs))
   {
     if (unlikely(convert->copy(str, length, src_cs, dst_cs, &error_unused)))
       return TRUE;
@@ -5334,9 +5340,11 @@ bool Protocol_local::store_decimal(const my_decimal *value)
 /** Store a string. */
 
 bool Protocol_local::store_str(const char *str, size_t length,
-                               CHARSET_INFO *src_cs, CHARSET_INFO *dst_cs)
+                               CHARSET_INFO *src_cs,
+                               my_repertoire_t from_repertoire,
+                               CHARSET_INFO *dst_cs)
 {
-  return store_string(str, length, src_cs, dst_cs);
+  return store_string(str, length, src_cs, from_repertoire, dst_cs);
 }