diff options
Diffstat (limited to 'spec/lib')
| -rw-r--r-- | spec/lib/alaveteli_text_masker_spec.rb | 7 | ||||
| -rw-r--r-- | spec/lib/basic_encoding_spec.rb | 24 |
2 files changed, 20 insertions, 11 deletions
diff --git a/spec/lib/alaveteli_text_masker_spec.rb b/spec/lib/alaveteli_text_masker_spec.rb index f2d52c1cc..f8c22a849 100644 --- a/spec/lib/alaveteli_text_masker_spec.rb +++ b/spec/lib/alaveteli_text_masker_spec.rb @@ -31,10 +31,13 @@ describe AlaveteliTextMasker do data.should == "There was a xxxxx called xxxxxxx, he wished that he was xxxx." end - it 'should handle multibyte characters correctly' do + it 'should handle multibyte characters in binary file types as binary data' do data = 'á mouse' + if String.method_defined?(:encode) + data = data.force_encoding("ASCII-8BIT") + end @regex_censor_rule.text = 'á' - apply_masks!(data, "application/octet-stream", :censor_rules => @censor_rules).should == 'x mouse' + apply_masks!(data, "application/octet-stream", :censor_rules => @censor_rules).should == 'xx mouse' end it "should apply censor rules to HTML files" do diff --git a/spec/lib/basic_encoding_spec.rb b/spec/lib/basic_encoding_spec.rb index d77465ad8..6758d60a3 100644 --- a/spec/lib/basic_encoding_spec.rb +++ b/spec/lib/basic_encoding_spec.rb @@ -160,21 +160,24 @@ describe "convert_string_to_utf8" do describe "when passed uninterpretable character data" do - it "should return it as a utf8 string" do + it "should return it as a valid utf8 string with non-utf8 characters removed + and mark it as scrubbed" do converted = convert_string_to_utf8 random_string - converted.should == random_string if String.method_defined?(:encode) - converted.encoding.to_s.should == 'UTF-8' + converted.string.encoding.to_s.should == 'UTF-8' + converted.string.valid_encoding?.should == true end + converted.scrubbed?.should == true converted = convert_string_to_utf8 random_string,'UTF-8' - converted.should == random_string if String.method_defined?(:encode) - converted.encoding.to_s.should == 'UTF-8' + converted.string.encoding.to_s.should == 'UTF-8' + converted.string.valid_encoding?.should == true end + converted.scrubbed?.should == true end end @@ -185,11 +188,13 @@ describe "convert_string_to_utf8" do converted = convert_string_to_utf8 windows_1252_string - converted.should == "DASH – DASH" + converted.string.should == "DASH – DASH" if String.method_defined?(:encode) - converted.encoding.to_s.should == 'UTF-8' + converted.string.encoding.to_s.should == 'UTF-8' end + converted.scrubbed?.should == false + end end @@ -200,11 +205,12 @@ describe "convert_string_to_utf8" do converted = convert_string_to_utf8 gb_18030_spam_string - converted.should start_with("贵公司负责人") + converted.string.should start_with("贵公司负责人") if String.method_defined?(:encode) - converted.encoding.to_s.should == 'UTF-8' + converted.string.encoding.to_s.should == 'UTF-8' end + converted.scrubbed?.should == false end end |