Regexps can be fast, but they don't have to

LIghtning talk about regexp performance pitfalls in Ruby and browser. Wroclove.rb 2018

Transcript

1. REGEXPS CAN BE
   REGEXPS CAN BE
   FAST
   FAST
   BUT THEY DON'T HAVE TO
   BUT THEY DON'T HAVE TO
   MACIEK RZĄSA, TEXTMASTER
   MACIEK RZĄSA, TEXTMASTER
   @mjrzasa
   

   View Slide

2. View Slide

3. View Slide

4. BROWSER
   BROWSER
   
   https://regex-performance.github.io/input-en.html
   

   View Slide

5. RUBY
   RUBY
   1231_1231_21312_1231231_12312
   /(\d+_?)+\b/
   

   View Slide

6. 100% CPU ON PRODUCTION
   100% CPU ON PRODUCTION
   Hello! → Witaj! → ¡Hola! → Salut !
   Discourse client, described by Sam Saffron
   ::Typography.to_html_french
   # Insert a thin space before punctuation
   text.gsub(/(\s|)+([!?;]+(\s|\z))/, ' \2\3')
   # <-58 spaces>
   GET /wp-login.php HTTP/1.1 69
   GET /show.aspx HTTP/1.1 15
   

   View Slide

7. DEBUGGER
   DEBUGGER
   http://regex101.com
   

   View Slide

8. WHAT COULD I DO?
   WHAT COULD I DO?
   

   View Slide

9. View Slide

10. JUST KIDDING
    JUST KIDDING
    

    View Slide

11. WHAT COULD I DO?
    WHAT COULD I DO?
    be specific
    test almost matching input
    don't believe in magic, there is just
    software
    I will send links on twitter
    

    View Slide

12. View Slide