Overly eager regex

Question

I'm trying to avoid emails in my input field boxes, so I'm using multiple regex for this, I got one that is matching more than I want, this is the regex :

(at( )*\w+(\b\.ie\b|\b\ .ie\b))

And its supposed to match attempts like email at domain.ie but I got also this part matching :

also what a tie they are

How can I modify this regex not to match this case?

I tried this :

(at( )[a-z^\s]\w+(\b\.ie\b|\b\ .ie\b))

and couple of more things, still I get the same match, how can I still keep the matching that I want but to avoid this case

Answer 1

To match something like "at foo.ie" or "at foo .ie", something as simple as this will work: /\bat +\w+ ?\.ie\b/

Answer 2

To avoid the unexpected match, try using a literal period \. in your regular expression. Testing that against a few inputs looks like it gets what you want:

tests = [
  {:text => 'at domain.ie',             :expected => true},
  {:text => 'at something.ie',          :expected => true},
  {:text => 'at foo.ie',                :expected => true},
  {:text => 'also what a tie they are', :expected => false}
]

regex = /\bat\s+\w+\s?\.ie\b/

tests.each do |test|
  text         = test[:text]
  expect_match = test[:expected]
  matched = false
  if text =~ regex
    matched = true
  end

  if expect_match == matched
    puts "OK:     expected=#{expect_match} == #{matched} for regex=#{regex} vs text=#{text}"
  else
    puts "NOT OK: expected=#{expect_match} != #{matched} for regex=#{regex} vs text=#{text}"
  end
end

If you have more test cases, it should be easy to add them to the above code - though you may wish to make a more formal test using something like rspec.