MDEV-26713 set console codepage to what user set in --default-character-set

If someone on whatever reasons uses --default-character-set=cp850,
this will avoid incorrect display, and inserting incorrect data.

Adjusting console codepage sometimes also needs to happen with
--default-charset=auto, on older Windows. This is because autodetection
is not always exact. For example, console codepage on US editions of
Windows is 437. Client autodetects it as cp850, a rather loose
approximation, given 46 code point differences. We change the console
codepage to cp850, so that there is no discrepancy.

That fix is currently Windows-only, and serves people who used combination
of chcp to achieve WYSIWYG effect (although, this would mostly likely used
with utf8 in the past)

Now, --default-character-set would be a replacement for that.

Fix fs_character_set() detection of current codepage.

client/mysql.cc

@@ -3304,6 +3304,21 @@ com_clear(String *buffer,char *line __attribute__((unused)))
   return 0;
 }
 
+static void adjust_console_codepage(const char *name __attribute__((unused)))
+{
+#ifdef _WIN32
+  if (my_set_console_cp(name) < 0)
+  {
+    char buf[128];
+    snprintf(buf, sizeof(buf),
+      "WARNING: Could not determine Windows codepage for charset '%s',"
+      "continue using codepage %u", name, GetConsoleOutputCP());
+    put_info(buf, INFO_INFO);
+  }
+#endif
+}
+
+
 	/* ARGSUSED */
 static int
 com_charset(String *buffer __attribute__((unused)), char *line)
@@ -3325,6 +3340,7 @@ com_charset(String *buffer __attribute__((unused)), char *line)
     mysql_set_character_set(&mysql, charset_info->cs_name.str);
     default_charset= (char *)charset_info->cs_name.str;
     put_info("Charset changed", INFO_INFO);
+    adjust_console_codepage(charset_info->cs_name.str);
   }
   else put_info("Charset is not found", INFO_INFO);
   return 0;
@@ -4873,6 +4889,7 @@ sql_real_connect(char *host,char *database,char *user,char *password,
     put_info(buff, INFO_ERROR);
     return 1;
   }
+  adjust_console_codepage(charset_info->cs_name.str);
   connected=1;
 #ifndef EMBEDDED_LIBRARY
   mysql_options(&mysql, MYSQL_OPT_RECONNECT, &debug_info_flag);

client/mysqladmin.cc

@@ -438,6 +438,7 @@ int main(int argc,char *argv[])
     mysql_options(&mysql,MYSQL_OPT_PROTOCOL,(char*)&opt_protocol);
   if (!strcmp(default_charset,MYSQL_AUTODETECT_CHARSET_NAME))
     default_charset= (char *)my_default_csname();
+  my_set_console_cp(default_charset);
   mysql_options(&mysql, MYSQL_SET_CHARSET_NAME, default_charset);
   error_flags= (myf)(opt_nobeep ? 0 : ME_BELL);
 

client/mysqlcheck.c

@@ -503,6 +503,7 @@ static int get_options(int *argc, char ***argv)
     printf("Unsupported character set: %s\n", default_charset);
     DBUG_RETURN(1);
   }
+  my_set_console_cp(default_charset);
   if (*argc > 0 && opt_alldbs)
   {
     printf("You should give only options, no arguments at all, with option\n");

client/mysqlimport.c

@@ -525,6 +525,7 @@ static MYSQL *db_connect(char *host, char *database,
     mysql_options(mysql, MYSQL_DEFAULT_AUTH, opt_default_auth);
   if (!strcmp(default_charset,MYSQL_AUTODETECT_CHARSET_NAME))
     default_charset= (char *)my_default_csname();
+  my_set_console_cp(default_charset);
   mysql_options(mysql, MYSQL_SET_CHARSET_NAME, my_default_csname());
   mysql_options(mysql, MYSQL_OPT_CONNECT_ATTR_RESET, 0);
   mysql_options4(mysql, MYSQL_OPT_CONNECT_ATTR_ADD,

client/mysqlshow.c

@@ -147,6 +147,7 @@ int main(int argc, char **argv)
 
   if (!strcmp(default_charset,MYSQL_AUTODETECT_CHARSET_NAME))
     default_charset= (char *)my_default_csname();
+  my_set_console_cp(default_charset);
   mysql_options(&mysql, MYSQL_SET_CHARSET_NAME, default_charset);
 
   if (opt_plugin_dir && *opt_plugin_dir)

include/my_sys.h

@@ -1086,6 +1086,9 @@ extern char *get_tty_password(const char *opt_message);
 #define BACKSLASH_MBTAIL
 /* File system character set */
 extern CHARSET_INFO *fs_character_set(void);
+extern int my_set_console_cp(const char *name);
+#else
+#define my_set_console_cp(A) do {} while (0)
 #endif
 extern const char *my_default_csname(void);
 extern size_t escape_quotes_for_mysql(CHARSET_INFO *charset_info,

mysys/charset.c

@@ -1209,30 +1209,17 @@ size_t escape_string_for_mysql(CHARSET_INFO *charset_info,
 
 
 #ifdef BACKSLASH_MBTAIL
-static CHARSET_INFO *fs_cset_cache= NULL;
-
 CHARSET_INFO *fs_character_set()
 {
-  if (!fs_cset_cache)
+  static CHARSET_INFO *fs_cset_cache;
-  {
+  if (fs_cset_cache)
-    char buf[10]= "cp";
+    return fs_cset_cache;
-    GetLocaleInfo(LOCALE_SYSTEM_DEFAULT, LOCALE_IDEFAULTANSICODEPAGE,
+#ifdef HAVE_CHARSET_cp932
-                  buf+2, sizeof(buf)-3);
+  else if (GetACP() == 932)
-    /*
+    return fs_cset_cache= &my_charset_cp932_japanese_ci;
-      We cannot call get_charset_by_name here
+#endif
-      because fs_character_set() is executed before
+  else
-      LOCK_THD_charset mutex initialization, which
+    return fs_cset_cache= &my_charset_bin;
-      is used inside get_charset_by_name.
-      As we're now interested in cp932 only,
-      let's just detect it using strcmp().
-    */
-    fs_cset_cache= 
-                #ifdef HAVE_CHARSET_cp932
-                        !strcmp(buf, "cp932") ? &my_charset_cp932_japanese_ci : 
-                #endif
-                        &my_charset_bin;
-  }
-  return fs_cset_cache;
 }
 #endif
 
@@ -1394,7 +1381,7 @@ static const MY_CSET_OS_NAME charsets[] =
   {"cp54936",        "gb18030",  my_cs_exact},
 #endif
   {"cp65001",        "utf8mb4",  my_cs_exact},
-
+  {"cp65001",        "utf8mb3",  my_cs_approx},
 #else /* not Windows */
 
   {"646",            "latin1",   my_cs_approx}, /* Default on Solaris */
@@ -1534,3 +1521,90 @@ const char* my_default_csname()
 #endif
   return csname ? csname : MYSQL_DEFAULT_CHARSET_NAME;
 }
+
+
+#ifdef _WIN32
+/**
+  Extract codepage number from "cpNNNN" string,
+  and check that this codepage is supported.
+
+  @return 0 - invalid codepage(or unsupported)
+          > 0 - valid codepage number.
+*/
+static UINT get_codepage(const char *s)
+{
+  UINT cp;
+  if (s[0] != 'c' || s[1] != 'p')
+  {
+    DBUG_ASSERT(0);
+    return 0;
+  }
+  cp= strtoul(s + 2, NULL, 10);
+  if (!IsValidCodePage(cp))
+  {
+    /*
+     Can happen also with documented CP, i.e 51936
+     Perhaps differs from one machine to another.
+    */
+    return 0;
+  }
+  return cp;
+}
+
+static UINT mysql_charset_to_codepage(const char *my_cs_name)
+{
+  const MY_CSET_OS_NAME *csp;
+  UINT cp=0,tmp;
+  for (csp= charsets; csp->os_name; csp++)
+  {
+    if (!strcasecmp(csp->my_name, my_cs_name))
+    {
+      switch (csp->param)
+      {
+      case my_cs_exact:
+        tmp= get_codepage(csp->os_name);
+        if (tmp)
+          return tmp;
+        break;
+      case my_cs_approx:
+        /*
+          don't return just yet, perhaps there is a better
+          (exact) match later.
+        */
+        if (!cp)
+          cp= get_codepage(csp->os_name);
+        continue;
+
+      default:
+        return 0;
+      }
+    }
+  }
+  return cp;
+}
+
+/** Set console codepage for MariaDB's charset name */
+int my_set_console_cp(const char *csname)
+{
+  UINT cp;
+  if (fileno(stdout) < 0 || !isatty(fileno(stdout)))
+    return 0;
+  cp= mysql_charset_to_codepage(csname);
+  if (!cp)
+  {
+    /* No compatible os charset.*/
+    return -1;
+  }
+
+  if (GetConsoleOutputCP() != cp && !SetConsoleOutputCP(cp))
+  {
+    return -1;
+  }
+
+  if (GetConsoleCP() != cp && !SetConsoleCP(cp))
+  {
+    return -1;
+  }
+  return 0;
+}
+#endif