properly remove invalid XML characters \u2028\u0001\u0002\u0003\u0004\u0005\u0006\u0007\u0008\u001f

2 files changed, 10 insertions, 1 deletions

diff --git a/lib/axlsx/workbook/worksheet/worksheet.rb b/lib/axlsx/workbook/worksheet/worksheet.rb
index 741e38b1..51613f75 100644
--- a/lib/axlsx/workbook/worksheet/worksheet.rb
+++ b/lib/axlsx/workbook/worksheet/worksheet.rb
@@ -541,7 +541,8 @@ module Axlsx
        end
        str.concat '</dataValidations>'
      end
-     str + '</worksheet>'
+     str.concat '</worksheet>'
+     str.tr("\u0000-\u001f\u007f\u0028", '')
     end
 
     # The worksheet relationships. This is managed automatically by the worksheet
diff --git a/test/workbook/worksheet/tc_worksheet.rb b/test/workbook/worksheet/tc_worksheet.rb
index ce2873ab..e0ea3b89 100644
--- a/test/workbook/worksheet/tc_worksheet.rb
+++ b/test/workbook/worksheet/tc_worksheet.rb
@@ -310,6 +310,14 @@ class TestWorksheet < Test::Unit::TestCase
     assert(errors.empty?, "error free validation")
   end
 
+  def test_to_xml_string_with_illegal_chars
+    nasties =  "\u2028\u0001\u0002\u0003\u0004\u0005\u0006\u0007\u0008\u001f"
+    @ws.add_row [nasties]
+    assert_equal(nasties, @ws.rows.last.cells.last.value)
+    schema = Nokogiri::XML::Schema(File.open(Axlsx::SML_XSD))
+    doc = Nokogiri::XML(@ws.to_xml_string)
+    assert(schema.validate(doc).map { |e| e.error }.empty?)
+  end
   # Make sure the XML for all optional elements (like pageMargins, autoFilter, ...)
   # is generated in correct order.
   def test_valid_with_optional_elements