Fix encoding for columns serialized with YAML

See a good explanation about the problem in '4. YAML serialized columns': http://blog.rayapps.com/2013/03/11/7-things-that-can-go-wrong-with-ruby-19-string-encodings/

Fix encoding for columns serialized with YAML
See a good explanation about the problem in '4. YAML serialized columns': http://blog.rayapps.com/2013/03/11/7-things-that-can-go-wrong-with-ruby-19-string-encodings/
Victor Costa
1 parent 45d9194b
Showing 1 changed file with 30 additions and 0 deletions Show diff stats
db/migrate/20140724134601_fix_yaml_encoding.rb
@@ -0,0 +1,30 @@
+class FixYamlEncoding < ActiveRecord::Migration
+  def self.up
+    fix_encoding(Block, 'settings')
+    fix_encoding(Product, 'data')
+    fix_encoding(Environment, 'settings')
+    fix_encoding(Profile, 'data')
+    fix_encoding(ActionTracker::Record, 'params')
+    fix_encoding(Article, 'setting')
+  end
+
+  def self.down
+    puts "Warning: cannot restore original encoding"
+  end
+
+  private
+
+  def self.fix_encoding(model, param)
+    result = model.find(:all, :conditions => "#{param} LIKE '%!binary%'")
+    puts "Fixing #{result.count} rows of #{model} (#{param})"
+    result.each {|r| r.update_column(param, deep_fix(r.send(param)).to_yaml)}
+  end
+
+  def self.deep_fix(hash)
+    hash.each do |value|
+      value.force_encoding('UTF-8') if value.is_a?(String) && !value.frozen? && value.encoding == Encoding::ASCII_8BIT
+      deep_fix(value) if value.respond_to?(:each)
+    end
+  end
+
+end
...	...	@@ -0,0 +1,30 @@
	1	+class FixYamlEncoding < ActiveRecord::Migration
	2	+ def self.up
	3	+ fix_encoding(Block, 'settings')
	4	+ fix_encoding(Product, 'data')
	5	+ fix_encoding(Environment, 'settings')
	6	+ fix_encoding(Profile, 'data')
	7	+ fix_encoding(ActionTracker::Record, 'params')
	8	+ fix_encoding(Article, 'setting')
	9	+ end
	10	+
	11	+ def self.down
	12	+ puts "Warning: cannot restore original encoding"
	13	+ end
	14	+
	15	+ private
	16	+
	17	+ def self.fix_encoding(model, param)
	18	+ result = model.find(:all, :conditions => "#{param} LIKE '%!binary%'")
	19	+ puts "Fixing #{result.count} rows of #{model} (#{param})"
	20	+ result.each {\|r\| r.update_column(param, deep_fix(r.send(param)).to_yaml)}
	21	+ end
	22	+
	23	+ def self.deep_fix(hash)
	24	+ hash.each do \|value\|
	25	+ value.force_encoding('UTF-8') if value.is_a?(String) && !value.frozen? && value.encoding == Encoding::ASCII_8BIT
	26	+ deep_fix(value) if value.respond_to?(:each)
	27	+ end
	28	+ end
	29	+
	30	+end
...	...