From 0829e6f394e98145ee64c9e559c5368e61d58fb1 Mon Sep 17 00:00:00 2001
From: Dumitru Ursu <dima@ceata.org>
Date: Fri, 6 Feb 2015 00:18:20 +0200
Subject: [PATCH] Add collation and config check

---
 Changelog.md                                  |  5 +++--
 config/database.yml.example                   |  1 +
 ...20150106050733_set_mysql_to_unicode_mb4.rb | 21 +++++++++++++------
 3 files changed, 19 insertions(+), 8 deletions(-)

diff --git a/Changelog.md b/Changelog.md
index 441983267..e89b2b318 100644
--- a/Changelog.md
+++ b/Changelog.md
@@ -33,8 +33,9 @@ series and run our comprehensive test suite against it.
 The default for including jQuery from a CDN has changed. If you want to continue to include it from a CDN, please explicitly set the `jquery_cdn` setting to `true` in diaspora.yml.
 
 ## Change in database.yml
-For MySQL databases, replace `charset: utf8` with `encoding: utf8mb4` in the file `config/database.yml`. This is enables full UTF8 support (4bytes characters), including standard emoji characters. See `database.yml.example` for reference.
-Also, do not forget to remove `collation: utf8_bin`. It will choose a compatible one automatically.
+For MySQL databases, replace `charset: utf8` with `encoding: utf8mb4m` and  change `collation` from `utf8_bin` to `utf8mb4_bin` in the file `config/database.yml`.
+This is enables full UTF8 support (4bytes characters), including standard emoji characters.
+See `database.yml.example` for reference.
 Please make sure to stop Diaspora prior running this migration!
 
 ## Experimental chat feature
diff --git a/config/database.yml.example b/config/database.yml.example
index e62f7a9a6..f5a1e08cd 100644
--- a/config/database.yml.example
+++ b/config/database.yml.example
@@ -6,6 +6,7 @@ mysql: &mysql
   password: ""
 #  socket: /tmp/mysql.sock
   encoding: utf8mb4
+  collation: utf8mb4_bin
 
 postgres: &postgres
   adapter: postgresql
diff --git a/db/migrate/20150106050733_set_mysql_to_unicode_mb4.rb b/db/migrate/20150106050733_set_mysql_to_unicode_mb4.rb
index c8866bc5e..faf4ce870 100644
--- a/db/migrate/20150106050733_set_mysql_to_unicode_mb4.rb
+++ b/db/migrate/20150106050733_set_mysql_to_unicode_mb4.rb
@@ -5,23 +5,32 @@ class SetMysqlToUnicodeMb4 < ActiveRecord::Migration
   def self.up
     # shorten indexes regardless of the RDBMS provider - for consitency
     shorten_indexes
-    change_encoding('utf8mb4') if AppConfig.mysql?
+    change_encoding('utf8mb4', 'utf8mb4_bin') if AppConfig.mysql?
   end
 
   def self.down
-    change_encoding('utf8') if AppConfig.mysql?
+    change_encoding('utf8', 'utf8_bin') if AppConfig.mysql?
   end
 
-  def change_encoding(encoding)
-    execute "ALTER DATABASE `#{ActiveRecord::Base.connection.current_database}` CHARACTER SET #{encoding};"
+  def check_config(encoding, collation)
+    connection_config = ActiveRecord::Base.connection_config
+    fail "Database encoding is not #{encoding}!"   if connection_config[:encoding] != encoding
+    fail "Database collation is not #{collation}!" if connection_config[:collation] != collation
+  end
+
+  def change_encoding(encoding, collation)
+    # Make sure the podmin changed the database.yml file
+    check_config(encoding, collation)
+
+    execute "ALTER DATABASE `#{ActiveRecord::Base.connection.current_database}` CHARACTER SET #{encoding} COLLATE #{collation};"
 
     tables.each do |table|
-      execute "ALTER TABLE `#{table}` CHARACTER SET = #{encoding}"
+      execute "ALTER TABLE `#{table}` CHARACTER SET = #{encoding} COLLATE #{collation}"
     end
 
     character_columns.each do |table, columns|
       columns.each do |column|
-        execute "ALTER TABLE `#{table}` CHANGE `#{column.name}` `#{column.name}` #{column.sql_type} CHARACTER SET #{encoding} #{column.null ? 'NULL' : 'NOT NULL'} #{"DEFAULT '#{column.default}'" if column.has_default?};"
+        execute "ALTER TABLE `#{table}` CHANGE `#{column.name}` `#{column.name}` #{column.sql_type} CHARACTER SET #{encoding} COLLATE #{collation} #{column.null ? 'NULL' : 'NOT NULL'} #{"DEFAULT '#{column.default}'" if column.has_default?};"
       end
     end
   end