Bug#35976922 Get character sets/collations by name rather than by global pointers

Fix this TODO:
// TODO(tdidriks) check name rather than address:
extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_gb18030_chinese_ci;
extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_utf16le_general_ci;

Character sets/collations should always be loaded/initialized properly
with get_collation_number() or some other defined function in the
mysys/strings API.

Change-Id: Ifa89308481d9a3db7428c1f3ed4ba375d1681209

Author: Tor Didriksen

include/my_sys.h

@@ -898,7 +898,7 @@ extern bool resolve_collation(const char *cl_name,
                               const CHARSET_INFO *default_cl,
                               const CHARSET_INFO **cl);
 extern char *get_charsets_dir(char *buf);
-extern bool init_compiled_charsets(myf flags);
+
 extern size_t escape_string_for_mysql(const CHARSET_INFO *charset_info,
                                       char *to, size_t to_length,
                                       const char *from, size_t length);

include/mysql/strings/m_ctype.h

@@ -485,9 +485,6 @@ extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_utf8mb3_unicode_ci;
 extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_utf8mb3_bin;
 extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_utf8mb4_bin;
 extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_utf8mb4_general_ci;
-// TODO(tdidriks) check name rather than address:
-extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_gb18030_chinese_ci;
-extern MYSQL_STRINGS_EXPORT CHARSET_INFO my_charset_utf16le_general_ci;
 
 /**
   @note Deprecated function, please call cs->coll->wildcmp(cs...) instead.

sql/mysqld.cc

@@ -816,6 +816,7 @@ MySQL clients support the protocol:
 #include "sql/query_options.h"
 #include "sql/range_optimizer/range_optimizer.h"  // range_optimizer_init
 #include "sql/reference_caching_setup.h"  // Event_reference_caching_channels
+#include "sql/regexp/regexp_facade.h"     // regexp::regexp_lib_charset
 #include "sql/replication.h"              // thd_enter_cond
 #include "sql/resourcegroups/resource_group_mgr.h"  // init, post_init
 #ifdef _WIN32
@@ -6966,6 +6967,13 @@ int init_common_variables() {
                                "--character-set-filesystem");
   global_system_variables.character_set_filesystem = character_set_filesystem;
 
+#ifdef WORDS_BIGENDIAN
+  regexp::regexp_lib_charset = get_charset_by_name("utf16_general_ci", MYF(0));
+#else
+  regexp::regexp_lib_charset =
+      get_charset_by_name("utf16le_general_ci", MYF(0));
+#endif
+
   while (!(my_default_lc_time_names = my_locale_by_name(
                nullptr, lc_time_names_name, strlen(lc_time_names_name)))) {
     LogErr(ERROR_LEVEL, ER_FAILED_TO_FIND_LOCALE_NAME, lc_time_names_name);

sql/regexp/regexp_engine.h

@@ -30,29 +30,17 @@
 #include <string>
 #include <utility>
 
-#include "my_config.h"              // WORDS_BIGENDIAN
-#include "mysql/strings/m_ctype.h"  // CHARSET_INFO.
 #include "sql/current_thd.h"
 #include "sql/regexp/errors.h"
 #include "sql/sql_class.h"  // THD
 #include "template_utils.h"
 
-extern CHARSET_INFO my_charset_utf16le_general_ci;
-extern CHARSET_INFO my_charset_utf16_general_ci;
-
 namespace regexp_engine_unittest {
 class Mock_regexp_engine;
 }
 
 namespace regexp {
 
-static constexpr CHARSET_INFO *regexp_lib_charset =
-#ifdef WORDS_BIGENDIAN
-    &::my_charset_utf16_general_ci;
-#else
-    &::my_charset_utf16le_general_ci;
-#endif
-
 const char *icu_version_string();
 
 /**

sql/regexp/regexp_facade.cc

@@ -27,16 +27,17 @@
 #include <tuple>
 
 #include "my_pointer_arithmetic.h"
+#include "mysql/strings/m_ctype.h"  // CHARSET_INFO.
 #include "sql/item_func.h"
 #include "sql/mysqld.h"  // make_unique_destroy_only
 #include "sql/regexp/regexp_engine.h"
 #include "sql_string.h"
 #include "template_utils.h"
 
-struct CHARSET_INFO;
-
 namespace regexp {
 
+CHARSET_INFO *regexp_lib_charset = nullptr;
+
 /**
   When dealing with the binary character set, we tell ICU that we're using
   CP-1252. This way, comparisons will happen as the user expects; each octet

sql/regexp/regexp_facade.h

@@ -38,11 +38,10 @@
 #include "sql/regexp/regexp_engine.h"
 #include "sql_string.h"
 
-extern int32_t opt_regexp_time_limit;
-extern int32_t opt_regexp_stack_limit;
-
 namespace regexp {
 
+extern CHARSET_INFO *regexp_lib_charset;
+
 /**
   This class handles
 

unittest/gunit/regexp_engine-t.cc

@@ -32,6 +32,7 @@
 
 #include "sql/parse_tree_items.h"
 #include "sql/regexp/regexp_engine.h"
+#include "sql/regexp/regexp_facade.h"
 
 namespace regexp_engine_unittest {