Bug#37575 UCASE fails on monthname

The MONTHNAME/DAYNAME functions
returns binary string, so the LOWER/UPPER functions
are not effective on the result of MONTHNAME/DAYNAME call.  
Character set of the MONTHNAME/DAYNAME function
result has been changed to connection character set.


include/m_ctype.h:
  added my_charset_repertoire function
mysql-test/r/ctype_ucs.result:
  test result
mysql-test/r/func_time.result:
  test result
mysql-test/t/ctype_ucs.test:
  test case
mysql-test/t/func_time.test:
  test case
sql/item_timefunc.cc:
  Item_func_monthname::fix_length_and_dec and
  Item_func_dayname::fix_length_and_dec methods have been
  modified to use connection character set
sql/item_timefunc.h:
  Item_func_monthname::fix_length_and_dec and
  Item_func_dayname::fix_length_and_dec methods have been
  modified to use connection character set
sql/mysql_priv.h:
  added max_month_name_length, max_day_name_length fields into MY_LOCALE struct
sql/mysqld.cc:
  The test_lc_time_sz function controls modifications
  of the locale database in debugging mode.
sql/sql_locale.cc:
  initialization of max_month_name_length, max_day_name_length fields
strings/ctype.c:
  added my_charset_repertoire function

include/m_ctype.h

@@ -445,6 +445,7 @@ my_bool my_propagate_complex(CHARSET_INFO *cs, const uchar *str, uint len);
 uint my_string_repertoire(CHARSET_INFO *cs, const char *str, ulong len);
 my_bool my_charset_is_ascii_based(CHARSET_INFO *cs);
 my_bool my_charset_is_8bit_pure_ascii(CHARSET_INFO *cs);
+uint my_charset_repertoire(CHARSET_INFO *cs);
 
 
 #define	_MY_U	01	/* Upper case */

mysql-test/r/ctype_ucs.result

@@ -1111,4 +1111,57 @@ set names latin1;
 select hex(char(0x41 using ucs2));
 hex(char(0x41 using ucs2))
 0041
+SET character_set_connection=ucs2;
+SELECT CHARSET(DAYNAME(19700101));
+CHARSET(DAYNAME(19700101))
+ucs2
+SELECT CHARSET(MONTHNAME(19700101));
+CHARSET(MONTHNAME(19700101))
+ucs2
+SELECT LOWER(DAYNAME(19700101));
+LOWER(DAYNAME(19700101))
+thursday
+SELECT LOWER(MONTHNAME(19700101));
+LOWER(MONTHNAME(19700101))
+january
+SELECT UPPER(DAYNAME(19700101));
+UPPER(DAYNAME(19700101))
+THURSDAY
+SELECT UPPER(MONTHNAME(19700101));
+UPPER(MONTHNAME(19700101))
+JANUARY
+SELECT HEX(MONTHNAME(19700101));
+HEX(MONTHNAME(19700101))
+004A0061006E0075006100720079
+SELECT HEX(DAYNAME(19700101));
+HEX(DAYNAME(19700101))
+00540068007500720073006400610079
+SET LC_TIME_NAMES=ru_RU;
+SET NAMES utf8;
+SET character_set_connection=ucs2;
+SELECT CHARSET(DAYNAME(19700101));
+CHARSET(DAYNAME(19700101))
+ucs2
+SELECT CHARSET(MONTHNAME(19700101));
+CHARSET(MONTHNAME(19700101))
+ucs2
+SELECT LOWER(DAYNAME(19700101));
+LOWER(DAYNAME(19700101))
+четверг
+SELECT LOWER(MONTHNAME(19700101));
+LOWER(MONTHNAME(19700101))
+января
+SELECT UPPER(DAYNAME(19700101));
+UPPER(DAYNAME(19700101))
+ЧЕТВЕРГ
+SELECT UPPER(MONTHNAME(19700101));
+UPPER(MONTHNAME(19700101))
+ЯНВАРЯ
+SELECT HEX(MONTHNAME(19700101));
+HEX(MONTHNAME(19700101))
+042F043D043204300440044F
+SELECT HEX(DAYNAME(19700101));
+HEX(DAYNAME(19700101))
+0427043504420432043504400433
+SET character_set_connection=latin1;
 End of 5.0 tests

mysql-test/r/func_time.result

@@ -592,6 +592,21 @@ unix_timestamp('1970-01-01 03:00:01')
 select unix_timestamp('2038-01-19 07:14:07');
 unix_timestamp('2038-01-19 07:14:07')
 0
+SELECT CHARSET(DAYNAME(19700101));
+CHARSET(DAYNAME(19700101))
+latin1
+SELECT CHARSET(MONTHNAME(19700101));
+CHARSET(MONTHNAME(19700101))
+latin1
+SELECT LOWER(DAYNAME(19700101));
+LOWER(DAYNAME(19700101))
+thursday
+SELECT LOWER(MONTHNAME(19700101));
+LOWER(MONTHNAME(19700101))
+january
+SELECT COERCIBILITY(MONTHNAME('1970-01-01')),COERCIBILITY(DAYNAME('1970-01-01'));
+COERCIBILITY(MONTHNAME('1970-01-01'))	COERCIBILITY(DAYNAME('1970-01-01'))
+4	4
 CREATE TABLE t1 (datetime datetime, timestamp timestamp, date date, time time);
 INSERT INTO t1 values ("2001-01-02 03:04:05", "2002-01-02 03:04:05", "2003-01-02", "06:07:08");
 SELECT * from t1;

mysql-test/t/ctype_ucs.test

@@ -685,4 +685,29 @@ set names latin1;
 #
 select hex(char(0x41 using ucs2));
 
+#
+# Bug#37575: UCASE fails on monthname
+#
+SET character_set_connection=ucs2;
+SELECT CHARSET(DAYNAME(19700101));
+SELECT CHARSET(MONTHNAME(19700101));
+SELECT LOWER(DAYNAME(19700101));
+SELECT LOWER(MONTHNAME(19700101));
+SELECT UPPER(DAYNAME(19700101));
+SELECT UPPER(MONTHNAME(19700101));
+SELECT HEX(MONTHNAME(19700101));
+SELECT HEX(DAYNAME(19700101));
+SET LC_TIME_NAMES=ru_RU;
+SET NAMES utf8;
+SET character_set_connection=ucs2;
+SELECT CHARSET(DAYNAME(19700101));
+SELECT CHARSET(MONTHNAME(19700101));
+SELECT LOWER(DAYNAME(19700101));
+SELECT LOWER(MONTHNAME(19700101));
+SELECT UPPER(DAYNAME(19700101));
+SELECT UPPER(MONTHNAME(19700101));
+SELECT HEX(MONTHNAME(19700101));
+SELECT HEX(DAYNAME(19700101));
+SET character_set_connection=latin1;
+
 --echo End of 5.0 tests

mysql-test/t/func_time.test

@@ -304,6 +304,15 @@ select unix_timestamp('1970-01-01 03:00:01');
 # check bad date, close to the boundary (we cut them off in the very end)
 select unix_timestamp('2038-01-19 07:14:07');
 
+#
+# Bug #28759: DAYNAME() and MONTHNAME() return binary string
+#
+
+SELECT CHARSET(DAYNAME(19700101));
+SELECT CHARSET(MONTHNAME(19700101));
+SELECT LOWER(DAYNAME(19700101));
+SELECT LOWER(MONTHNAME(19700101));
+SELECT COERCIBILITY(MONTHNAME('1970-01-01')),COERCIBILITY(DAYNAME('1970-01-01'));
 
 #
 # Test types from + INTERVAL

sql/item_timefunc.cc

@@ -1036,12 +1036,25 @@ longlong Item_func_month::val_int()
 }
 
 
+void Item_func_monthname::fix_length_and_dec()
+{
+  THD* thd= current_thd;
+  CHARSET_INFO *cs= thd->variables.collation_connection;
+  uint32 repertoire= my_charset_repertoire(cs);
+  locale= thd->variables.lc_time_names;  
+  collation.set(cs, DERIVATION_COERCIBLE, repertoire);
+  decimals=0;
+  max_length= locale->max_month_name_length * collation.collation->mbmaxlen;
+  maybe_null=1; 
+}
+
+
 String* Item_func_monthname::val_str(String* str)
 {
   DBUG_ASSERT(fixed == 1);
   const char *month_name;
-  uint   month= (uint) val_int();
-  THD *thd= current_thd;
+  uint month= (uint) val_int();
+  uint err;
 
   if (null_value || !month)
   {
@@ -1049,8 +1062,9 @@ String* Item_func_monthname::val_str(String* str)
     return (String*) 0;
   }
   null_value=0;
-  month_name= thd->variables.lc_time_names->month_names->type_names[month-1];
-  str->set(month_name, strlen(month_name), system_charset_info);
+  month_name= locale->month_names->type_names[month-1];
+  str->copy(month_name, strlen(month_name), &my_charset_utf8_bin,
+	    collation.collation, &err);
   return str;
 }
 
@@ -1169,19 +1183,32 @@ longlong Item_func_weekday::val_int()
                                  odbc_type) + test(odbc_type);
 }
 
+void Item_func_dayname::fix_length_and_dec()
+{
+  THD* thd= current_thd;
+  CHARSET_INFO *cs= thd->variables.collation_connection;
+  uint32 repertoire= my_charset_repertoire(cs);
+  locale= thd->variables.lc_time_names;  
+  collation.set(cs, DERIVATION_COERCIBLE, repertoire);
+  decimals=0;
+  max_length= locale->max_day_name_length * collation.collation->mbmaxlen;
+  maybe_null=1; 
+}
+
 
 String* Item_func_dayname::val_str(String* str)
 {
   DBUG_ASSERT(fixed == 1);
   uint weekday=(uint) val_int();		// Always Item_func_daynr()
   const char *day_name;
-  THD *thd= current_thd;
+  uint err;
 
   if (null_value)
     return (String*) 0;
   
-  day_name= thd->variables.lc_time_names->day_names->type_names[weekday];
-  str->set(day_name, strlen(day_name), system_charset_info);
+  day_name= locale->day_names->type_names[weekday];
+  str->copy(day_name, strlen(day_name), &my_charset_utf8_bin,
+	    collation.collation, &err);
   return str;
 }
 

sql/item_timefunc.h

@@ -108,18 +108,13 @@ public:
 
 class Item_func_monthname :public Item_func_month
 {
+  MY_LOCALE *locale;
 public:
   Item_func_monthname(Item *a) :Item_func_month(a) {}
   const char *func_name() const { return "monthname"; }
   String *val_str(String *str);
   enum Item_result result_type () const { return STRING_RESULT; }
-  void fix_length_and_dec() 
-  {
-    collation.set(&my_charset_bin);
-    decimals=0;
-    max_length=10*my_charset_bin.mbmaxlen;
-    maybe_null=1; 
-  }
+  void fix_length_and_dec();
 };
 
 
@@ -272,18 +267,13 @@ public:
 
 class Item_func_dayname :public Item_func_weekday
 {
+  MY_LOCALE *locale;
  public:
   Item_func_dayname(Item *a) :Item_func_weekday(a,0) {}
   const char *func_name() const { return "dayname"; }
   String *val_str(String *str);
   enum Item_result result_type () const { return STRING_RESULT; }
-  void fix_length_and_dec() 
-  { 
-    collation.set(&my_charset_bin);
-    decimals=0; 
-    max_length=9*MY_CHARSET_BIN_MB_MAXLEN;
-    maybe_null=1; 
-  }
+  void fix_length_and_dec();
 };
 
 

sql/mysql_priv.h

@@ -132,15 +132,20 @@ typedef struct my_locale_st
   TYPELIB *ab_month_names;
   TYPELIB *day_names;
   TYPELIB *ab_day_names;
+  uint max_month_name_length;
+  uint max_day_name_length;
 #ifdef __cplusplus 
   my_locale_st(uint number_par,
                const char *name_par, const char *descr_par, bool is_ascii_par,
                TYPELIB *month_names_par, TYPELIB *ab_month_names_par,
-               TYPELIB *day_names_par, TYPELIB *ab_day_names_par) : 
+               TYPELIB *day_names_par, TYPELIB *ab_day_names_par,
+               uint max_month_name_length_par, uint max_day_name_length_par) : 
     number(number_par),
     name(name_par), description(descr_par), is_ascii(is_ascii_par),
     month_names(month_names_par), ab_month_names(ab_month_names_par),
-    day_names(day_names_par), ab_day_names(ab_day_names_par)
+    day_names(day_names_par), ab_day_names(ab_day_names_par),
+    max_month_name_length(max_month_name_length_par),
+    max_day_name_length(max_day_name_length_par)
   {}
 #endif
 } MY_LOCALE;

sql/mysqld.cc

@@ -3616,6 +3616,44 @@ void decrement_handler_count()
 #endif /* defined(__NT__) || defined(HAVE_SMEM) */
 
 
+#ifndef DBUG_OFF
+/*
+  Debugging helper function to keep the locale database
+  (see sql_locale.cc) and max_month_name_length and
+  max_day_name_length variable values in consistent state.
+*/
+static void test_lc_time_sz()
+{
+  DBUG_ENTER("test_lc_time_sz");
+  for (MY_LOCALE **loc= my_locales; *loc; loc++)
+  {
+    uint max_month_len= 0;
+    uint max_day_len = 0;
+    for (const char **month= (*loc)->month_names->type_names; *month; month++)
+    {
+      set_if_bigger(max_month_len,
+                    my_numchars_mb(&my_charset_utf8_general_ci,
+                                   *month, *month + strlen(*month)));
+    }
+    for (const char **day= (*loc)->day_names->type_names; *day; day++)
+    {
+      set_if_bigger(max_day_len,
+                    my_numchars_mb(&my_charset_utf8_general_ci,
+                                   *day, *day + strlen(*day)));
+    }
+    if ((*loc)->max_month_name_length != max_month_len ||
+        (*loc)->max_day_name_length != max_day_len)
+    {
+      DBUG_PRINT("Wrong max day name(or month name) length for locale:",
+                 ("%s", (*loc)->name));
+      DBUG_ASSERT(0);
+    }
+  }
+  DBUG_VOID_RETURN;
+}
+#endif//DBUG_OFF
+
+
 #ifndef EMBEDDED_LIBRARY
 #ifdef __WIN__
 int win_main(int argc, char **argv)
@@ -3712,6 +3750,10 @@ int main(int argc, char **argv)
   openlog(libwrapName, LOG_PID, LOG_AUTH);
 #endif
 
+#ifndef DBUG_OFF
+  test_lc_time_sz();
+#endif
+
   /*
     We have enough space for fiddling with the argv, continue
   */

strings/ctype.c

@@ -337,6 +337,16 @@ my_string_repertoire(CHARSET_INFO *cs, const char *str, ulong length)
 }
 
 
+/*
+  Returns repertoire for charset
+*/
+uint my_charset_repertoire(CHARSET_INFO *cs)
+{
+  return cs->state & MY_CS_PUREASCII ?
+    MY_REPERTOIRE_ASCII : MY_REPERTOIRE_UNICODE30;
+}
+
+
 /*
   Detect whether a character set is ASCII compatible.