From: James Gray <redmine@...>
Date: 2010-03-22T04:23:58+09:00
Subject: [ruby-core:28841] [Feature #1981] [PATCH] CSV Parsing Speedup

Issue #1981 has been updated by James Gray.


Thanks Timothy, but this patch doesn't include your latest fix from the FasterCSV side, to restore the strict parser behavior.  This tests is currently failing:

  def test_non_regex_edge_cases
    # An early version of the non-regex parser fails this test
    [ [ "foo,\"foo,bar,baz,foo\",\"foo\"",
        ["foo", "foo,bar,baz,foo", "foo"] ] ].each do |edge_case|
      assert_equal(edge_case.last, CSV.parse_line(edge_case.first))
    end

    assert_raise(CSV::MalformedCSVError) do
      CSV.parse_line("1,\"23\"4\"5\", 6")
    end
  end

Can I bother you for one more patch that includes this fix as well?  Thanks!

(I'm applying all of the new FasterCSV fixes and am now missing only this in my local checkout, so a patch adding this after the latest patch you gave me would be best.  You also don't need to add the test, since I already have.)

James Edward Gray II

P.S.  It is quite a bit faster and does seem to play well with m17n, from what I can see.  Great work!

----------------------------------------
http://redmine.ruby-lang.org/issues/show/1981

----------------------------------------
http://redmine.ruby-lang.org