Single regexp for strip_html

[printercu]

Hi!

Benchmark shows almost 2x improvement.

Benchmark source

STRIP_HTML = Regexp.union(
  /<script.*?<\/script>/m,
  /<!--.*?-->/m,
  /<style.*?<\/style>/m,
  /<.*?>/m,
)

def strip_html_old(input)
  empty = ''.freeze
  input.to_s.gsub(/<script.*?<\/script>/m, empty).gsub(/<!--.*?-->/m, empty).gsub(/<style.*?<\/style>/m, empty).gsub(/<.*?>/m, empty)
end

def strip_html_new(input)
  input.to_s.gsub(STRIP_HTML, ''.freeze)
end

require 'benchmark/ips'        

string = <<~HTML
  <div>test</div>
  <div id='test'>test</div>
  <script type='text/javascript'>document.write('some stuff');</script>
  <style type='text/css'>foo bar</style>
  <div\nclass='multiline'>test</div>
  <!-- foo bar \n test -->test
HTML

Benchmark.ips do |x|
  x.report('old') { strip_html_old(string) }
  x.report('new') { strip_html_new(string) }
  x.compare!
end

[pushrax]

I modified this to avoid changing behaviour while keeping a lot of the speedup.