Clean invalid utf-8 chars from subject

thoughtbot · Sep 29, 2015 · 5057367 · 5057367
1 parent f180df2
commit 5057367
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 1 deletion.
diff --git a/lib/griddler/email.rb b/lib/griddler/email.rb
@@ -11,7 +11,7 @@ def initialize(params)
 
       @to = recipients(:to)
       @from = extract_address(params[:from])
-      @subject = params[:subject]
+      @subject = extract_subject
 
       @body = extract_body
       @raw_text = params[:text]
@@ -44,6 +44,10 @@ def extract_address(address)
       EmailParser.parse_address(clean_text(address))
     end
 
+    def extract_subject
+      clean_text(params[:subject])
+    end
+
     def extract_body
       EmailParser.extract_reply_body(text_or_sanitized_html)
     end

diff --git a/spec/griddler/email_spec.rb b/spec/griddler/email_spec.rb
@@ -639,6 +639,32 @@ def header_from_email(header)
   end
 end
 
+describe Griddler::Email, 'extracting email subject' do
+  before do
+    @address = '[email protected]'
+    @subject = 'A very interesting email'
+  end
+
+  it 'handles normal characters' do
+    email = Griddler::Email.new(
+      to: [@address],
+      from: @address,
+      subject: @subject,
+    )
+    expect(email.subject).to eq @subject
+  end
+
+  it 'handles invalid UTF-8 characters' do
+    email = Griddler::Email.new(
+      to: [@address],
+      from: @address,
+      subject: "\xc0\xc1\xf5\xfa\xfe\xff #{@subject}",
+    )
+    expected = "ÀÁõúþÿ #{@subject}"
+    expect(email.subject).to eq expected
+  end
+end
+
 describe Griddler::Email, 'extracting email addresses from CC field' do
   before do
     @address = '[email protected]'