escape html to support <,>, etc

simonoff · Jan 30, 2009 · a100556 · a100556
1 parent 3114145
commit a100556
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 4 deletions.
diff --git a/lib/roo/openoffice.rb b/lib/roo/openoffice.rb
@@ -5,6 +5,7 @@
 require 'zip/zipfilesystem'
 require 'date'
 require 'base64'
+require 'cgi'
 
 class Openoffice < GenericSpreadsheet
 
@@ -304,9 +305,8 @@ def read_cells(sheet=nil)
                                       str_v = str_v + child.to_s #.text
                                     }
                                   end
-                                  str_v.gsub!(/&quot;/,'"')
-                                  str_v.gsub!(/&amp;/,'&')
-                                  str_v.gsub!(/&apos;/,"'")
+                                  str_v.gsub!(/&apos;/,"'")  # special case not supported by unescapeHTML
+                                  str_v = CGI.unescapeHTML(str_v)
                                 end # == 'p'
                               end
                             elsif vt == 'time'

diff --git a/test/no_spreadsheet_file.txt b/test/no_spreadsheet_file.txt
@@ -1 +1 @@
-this is intionally not a spreadsheet file
+this is intentionally not a spreadsheet file
diff --git a/test/test_roo.rb b/test/test_roo.rb
@@ -4830,4 +4830,17 @@ def SKIP_test_bug_encoding_exported_from_google
       assert_equal 'test', xl.cell(2,'F')
     end
   end
+
+  def test_cell_openoffice_html_escape
+    if OPENOFFICE
+      oo = Openoffice.new(File.join("test","html-escape.ods"))
+      oo.default_sheet = oo.sheets.first
+      assert_equal "'", oo.cell(1,1)
+      assert_equal "&", oo.cell(2,1)
+      assert_equal ">", oo.cell(3,1)
+      assert_equal "<", oo.cell(4,1)
+      assert_equal "`", oo.cell(5,1)
+      # test_openoffice_zipped catches the &quot;
+     end  
+  end    
 end # class