1 files changed, 143 insertions, 0 deletions
diff --git a/sql/cset_narrowing.h b/sql/cset_narrowing.h
new file mode 100644
index 00000000..bb0a3960
--- /dev/null
+++ b/sql/cset_narrowing.h
@@ -0,0 +1,143 @@
+/*
+   Copyright (c) 2023, MariaDB Corporation.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; version 2 of the License.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335 USA */
+
+#ifndef CSET_NARROWING_H_INCLUDED
+#define CSET_NARROWING_H_INCLUDED
+
+/*
+  A singleton class to provide "utf8mb3_from_mb4.charset()".
+
+  This is a variant of utf8mb3_general_ci that one can use when they have data
+  in MB4 and want to make index lookup keys in MB3.
+*/
+extern
+class Charset_utf8narrow
+{
+  struct my_charset_handler_st cset_handler;
+  struct charset_info_st cset;
+public:
+  Charset_utf8narrow() :
+    cset_handler(*my_charset_utf8mb3_general_ci.cset),
+    cset(my_charset_utf8mb3_general_ci) /* Copy the CHARSET_INFO structure */
+  {
+    /* Insert our function wc_mb */
+    cset_handler.wc_mb= my_wc_mb_utf8mb4_bmp_only;
+    cset.cset=&cset_handler;
+
+    /* Charsets are compared by their name, so assign a different name */
+    LEX_CSTRING tmp= {STRING_WITH_LEN("utf8_mb4_to_mb3")};
+    cset.cs_name= tmp;
+  }
+
+  CHARSET_INFO *charset() { return &cset; }
+
+} utf8mb3_from_mb4;
+
+
+/*
+  A class to temporary change a field that uses utf8mb3_general_ci to enable
+  correct lookup key construction from string value in utf8mb4_general_ci
+
+  Intended usage:
+
+    // can do this in advance:
+    bool do_narrowing= Utf8_narrow::should_do_narrowing(field, value_cset);
+    ...
+
+    // This sets the field to do narrowing if necessary:
+    Utf8_narrow narrow(field, do_narrowing);
+
+    // write to 'field' here
+    // item->save_in_field(field) or something else
+
+    // Stop doing narrowing
+    narrow.stop();
+*/
+
+class Utf8_narrow
+{
+  Field *field;
+  DTCollation save_collation;
+
+public:
+  static bool should_do_narrowing(const THD *thd, CHARSET_INFO *field_cset,
+                                  CHARSET_INFO *value_cset);
+
+  static bool should_do_narrowing(const Field *field, CHARSET_INFO *value_cset)
+  {
+    CHARSET_INFO *field_cset= field->charset();
+    THD *thd= field->table->in_use;
+    return should_do_narrowing(thd, field_cset, value_cset);
+  }
+
+  Utf8_narrow(Field *field_arg, bool is_applicable)
+  {
+    field= NULL;
+    if (is_applicable)
+    {
+      DTCollation mb3_from_mb4= utf8mb3_from_mb4.charset();
+      field= field_arg;
+      save_collation= field->dtcollation();
+      field->change_charset(mb3_from_mb4);
+    }
+  }
+
+  void stop()
+  {
+    if (field)
+     field->change_charset(save_collation);
+#ifndef NDEBUG
+    field= NULL;
+#endif
+  }
+
+  ~Utf8_narrow()
+  {
+    DBUG_ASSERT(!field);
+  }
+};
+
+
+/*
+  @brief
+  Check if two fields can participate in a multiple equality using charset
+  narrowing.
+
+  @detail
+    Normally, check_simple_equality() checks this by calling:
+
+      left_field->eq_def(right_field)
+
+    This function does the same but takes into account we might use charset
+    narrowing:
+     - collations are not the same but rather an utf8mb{3,4}_general_ci pair
+     - for field lengths, should compare # characters, not #bytes.
+*/
+
+inline
+bool fields_equal_using_narrowing(const THD *thd, const Field *left, const Field *right)
+{
+  return
+    dynamic_cast<const Field_longstr*>(left) &&
+    dynamic_cast<const Field_longstr*>(right) &&
+    left->real_type() == right->real_type() &&
+    (Utf8_narrow::should_do_narrowing(left, right->charset()) ||
+     Utf8_narrow::should_do_narrowing(right, left->charset())) &&
+    left->char_length() == right->char_length();
+};
+
+
+#endif /* CSET_NARROWING_H_INCLUDED */