scm: ignore log encoding setting in Subversion and Mercurial (#7597).

git-svn-id: svn+ssh://rubyforge.org/var/svn/redmine/trunk@4962 e93f8b46-1217-0410-a6f0-8f06a7374b81

app/models/changeset.rb

@@ -56,10 +56,6 @@ class Changeset < ActiveRecord::Base
       revision.to_s
     end
   end
-  
-  def comments=(comment)
-    write_attribute(:comments, Changeset.normalize_comments(comment))
-  end
 
   def committed_on=(date)
     self.commit_date = date
@@ -75,10 +71,6 @@ class Changeset < ActiveRecord::Base
     end
   end
   
-  def committer=(arg)
-    write_attribute(:committer, self.class.to_utf8(arg.to_s))
-  end
-
   def project
     repository.project
   end
@@ -88,9 +80,11 @@ class Changeset < ActiveRecord::Base
   end
   
   def before_create
-    self.user = repository.find_committer_user(committer)
+    self.committer = self.class.to_utf8(self.committer, repository.repo_log_encoding)
+    self.comments  = self.class.normalize_comments(self.comments, repository.repo_log_encoding)
+    self.user = repository.find_committer_user(self.committer)
   end
-  
+
   def after_create
     scan_comment_for_issue_ids
   end
@@ -163,11 +157,6 @@ class Changeset < ActiveRecord::Base
     @next ||= Changeset.find(:first, :conditions => ['id > ? AND repository_id = ?', self.id, self.repository_id], :order => 'id ASC')
   end
   
-  # Strips and reencodes a commit log before insertion into the database
-  def self.normalize_comments(str)
-    to_utf8(str.to_s.strip)
-  end
-
   # Creates a new Change from it's common parameters
   def create_change(change)
     Change.create(:changeset => self,
@@ -176,7 +165,7 @@ class Changeset < ActiveRecord::Base
                   :from_path => change[:from_path],
                   :from_revision => change[:from_revision])
   end
-  
+
   private
 
   # Finds an issue that can be referenced by the commit message
@@ -246,9 +235,17 @@ class Changeset < ActiveRecord::Base
     return @short_comments, @long_comments
   end
 
-  def self.to_utf8(str)
+  public
+
+  # Strips and reencodes a commit log before insertion into the database
+  def self.normalize_comments(str, encoding)
+    Changeset.to_utf8(str.to_s.strip, encoding)
+  end
+
+  private
+
+  def self.to_utf8(str, encoding)
     return str if str.blank?
-    encoding = Setting.commit_logs_encoding.to_s.strip
     unless encoding.blank? || encoding == 'UTF-8'
       begin
         str = Iconv.conv('UTF-8', encoding, str)

app/models/repository.rb

@@ -190,6 +190,11 @@ class Repository < ActiveRecord::Base
     end
   end
 
+  def repo_log_encoding
+    encoding = Setting.commit_logs_encoding.to_s.strip
+    encoding.blank? ? 'UTF-8' : encoding
+  end
+
   # Fetches new changesets for all repositories of active projects
   # Can be called periodically by an external script
   # eg. ruby script/runner "Repository.fetch_changesets"
@@ -217,7 +222,7 @@ class Repository < ActiveRecord::Base
   def self.available_scm
     subclasses.collect {|klass| [klass.scm_name, klass.name]}
   end
-  
+
   def self.factory(klass_name, *args)
     klass = "Repository::#{klass_name}".constantize
     klass.new(*args)

app/models/repository/cvs.rb

@@ -107,10 +107,11 @@ class Repository::Cvs < Repository
         tmp_time = revision.time.clone
         unless changes.find_by_path_and_revision(
 	           scm.with_leading_slash(revision.paths[0][:path]), revision.paths[0][:revision])
+          cmt = Changeset.normalize_comments(revision.message, repo_log_encoding)
           cs = changesets.find(:first, :conditions=>{
             :committed_on=>tmp_time - time_delta .. tmp_time + time_delta,
             :committer=>revision.author,
-            :comments=>Changeset.normalize_comments(revision.message)
+            :comments=>cmt
           })
         
           # create a new changeset.... 

app/models/repository/mercurial.rb

@@ -34,6 +34,10 @@ class Repository::Mercurial < Repository
     'Mercurial'
   end
 
+  def repo_log_encoding
+    'UTF-8'
+  end
+
   # Returns the readable identifier for the given mercurial changeset
   def self.format_changeset_identifier(changeset)
     "#{changeset.revision}:#{changeset.scmid}"

app/models/repository/subversion.rb

@@ -30,6 +30,10 @@ class Repository::Subversion < Repository
     'Subversion'
   end
 
+  def repo_log_encoding
+    'UTF-8'
+  end
+
   def latest_changesets(path, rev, limit=10)
     revisions = scm.revisions(path, rev, nil, :limit => limit)
     revisions ? changesets.find_all_by_revision(revisions.collect(&:identifier), :order => "committed_on DESC", :include => :user) : []

test/unit/repository_subversion_test.rb

@@ -137,6 +137,28 @@ class RepositorySubversionTest < ActiveSupport::TestCase
       assert c.event_title.include?('123456789:')
       assert_equal '123456789', c.event_url[:rev]
     end
+
+    def test_log_encoding_ignore_setting
+      with_settings :commit_logs_encoding => 'windows-1252' do
+        s1 = "\xC2\x80"
+        s2 = "\xc3\x82\xc2\x80"
+        if s1.respond_to?(:force_encoding)
+          s3 = s1
+          s4 = s2
+          s1.force_encoding('ASCII-8BIT')
+          s2.force_encoding('ASCII-8BIT')
+          s3.force_encoding('ISO-8859-1')
+          s4.force_encoding('UTF-8')
+          assert_equal s3.encode('UTF-8'), s4
+        end
+        c = Changeset.new(:repository => @repository,
+                          :comments=>s2,
+                          :revision=>'123',
+                          :committed_on => Time.now)
+        assert c.save
+        assert_equal s2, c.comments
+      end
+    end
   else
     puts "Subversion test repository NOT FOUND. Skipping unit tests !!!"
     def test_fake; assert true end