encoding issues #1113

[mkristian]

feel free to cherry-pick any commit - each should be self contained.

2f43a0c does break

1) Failure:
Nokogiri::XML::TestNode#test_node_context_parsing_of_malformed_html_fragment_with_recover_is_corrected [test/xml/test_node.rb:109]:
Expected: 1
  Actual: 0

  2) Failure:
Nokogiri::XML::TestNode#test_node_context_parsing_of_malformed_html_fragment_without_recover_is_not_corrected [test/xml/test_node.rb:119]:
Expected: 1
  Actual: 0

nekohtml just does not report the missing attribute error anymore (or I did not find a way to tell nekohtml to do so). and I am not sure what to do to resolve this failures.

so any input and or ideas are welcome.

the xerces update it no really needed but I think it is overdue.

[vilius]

👍 would love to see this fixed

[flavorjones]

tagging @jvshahid

[jvshahid]

I was able to fix the regression in nekohtml. I have the commit in my nekohtml fork. Both nekohtml tests as well as nokogiri tests pass and the special case for jruby introduced in this pr isn't required anymore. I'll try to write a test and submit it upstream. I also reverted the xerces.jar file update since this was updated in #1212