Merge pull request #2510 from sparklemotion/flavorjones-encoding-read…

…er-performance-v1.13.x improve encoding reader performance (backport to v1.13.x)
sparklemotion · Apr 11, 2022 · 6a20ee4 · 6a20ee4
2 parents b848031 + e444525
commit 6a20ee4
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 1 deletion.
diff --git a/lib/nokogiri/html4/document.rb b/lib/nokogiri/html4/document.rb
@@ -268,7 +268,7 @@ def start_element(name, attrs = [])
         end
 
         def self.detect_encoding(chunk)
-          (m = chunk.match(/\A(<\?xml[ \t\r\n]+[^>]*>)/)) &&
+          (m = chunk.match(/\A(<\?xml[ \t\r\n][^>]*>)/)) &&
             (return Nokogiri.XML(m[1]).encoding)
 
           if Nokogiri.jruby?

diff --git a/test/html4/test_document_encoding.rb b/test/html4/test_document_encoding.rb
@@ -155,6 +155,18 @@ def binopen(file)
             end
           end
         end
+
+        it "does not start backtracking during detection of XHTML encoding" do
+          # this test is a quick and dirty version
+          # of the more complete perf test that is on main.
+          n = 40_000
+          redos_string = "<?xml " + (" " * n)
+          redos_string.encode!("ASCII-8BIT")
+          start_time = Time.now
+          Nokogiri::HTML4(redos_string)
+          elapsed_time = Time.now - start_time
+          assert_operator(elapsed_time, :<, 1)
+        end
       end
     end
   end