Contentful Testing at Scotland on Rails 2008


Published on

Slides from my "Contentful Testing - Regression Therapy" talk at Scotland on Rails 2008.

Published in: Technology
1 Comment
  • Notes and other download media can be found here:
    Are you sure you want to  Yes  No
    Your message goes here
  • Be the first to like this

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • I'm going to talk today about a bunch of issues. But, the heart of the talk concerns testing the content of views generated by a Rails app. I found I had quite a lot to say; and that I’m not enough of a speaker to improbise without missing key points, so I’ve decided to work from a script. So, you can expect me to rattle through plausible sounding material whilst not looking you in the eye, and occasionally losing my place and looking around vacantly as I try to remember what it is I’m actually talking about.
  • Contentful Testing at Scotland on Rails 2008

    1. 1. Regression Therapy – Contentful Testing Anthony Bailey 3 April 2008, Scotland on Rails
    2. 2. Teaser <ul><li>The conventional wisdom: </li></ul><ul><li># Assert explicitly and exactly what we care about. </li></ul><ul><li>assert_select(“divs#of_interest”, /my_desires/) </li></ul><ul><li>My dumb straw man alternative: </li></ul><ul><li># Everything should just be exactly like yesterday. </li></ul><ul><li>assert_equals(“expected.html”).read, @response.body ) </li></ul>Regression Therapy – Contentful Testing
    3. 3. Setting the scene <ul><li>I won’t try to behave. </li></ul><ul><li>When I say view content, I mean view content. </li></ul><ul><li>When I don’t, I mean everything… </li></ul><ul><li>… even other languages and frameworks. </li></ul>Regression Therapy – Contentful Testing
    4. 4. Setting the scene: terminology <ul><li>I won’t try to behave </li></ul>Regression Therapy – Contentful Testing
    5. 5. Setting the scene: scope <ul><li>When I say view content, I mean view content. </li></ul><ul><ul><li>not the models the views display </li></ul></ul><ul><ul><li>not the presentation logic they use to display it </li></ul></ul><ul><ul><li>not the controller logic invoked by interacting with the view </li></ul></ul><ul><ul><li>not the workflow that the views comprise </li></ul></ul><ul><ul><li>just the tags </li></ul></ul><ul><ul><li>(at least for now) </li></ul></ul>Regression Therapy – Contentful Testing
    6. 6. Setting the scene: scope <ul><li>When I don’t say views, I mean everything. </li></ul><ul><ul><li>Contentful testing can work in other domains </li></ul></ul>Regression Therapy – Contentful Testing
    7. 7. Setting the scene: scope <ul><li>Other languages and frameworks </li></ul><ul><ul><li>Contentful testing doesn’t need Ruby or Rails, </li></ul></ul><ul><ul><ul><li>used it in a C++ desktop app, a Java/Spring web app </li></ul></ul></ul><ul><ul><li>But this talk will focus on a Rails plug-in: </li></ul></ul><ul><ul><ul><li> </li></ul></ul></ul>Regression Therapy – Contentful Testing
    8. 8. Setting the scene: personal <ul><li>The plug-in was born of genuine need </li></ul><ul><ul><li>but in a small, simple, CRUDdy app. </li></ul></ul>Regression Therapy – Contentful Testing
    9. 9. Why do we test? <ul><li>To maintain known good behavior and catch accidental bad changes </li></ul><ul><li>........................... ! .. ! ..................... ! ......... </li></ul><ul><li>To drive intended changes </li></ul><ul><li>> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > </li></ul><ul><li>To explore and understand emergent changes </li></ul><ul><li>..... !!? ? ? ? ? ? ............ !!? ? ? ? ? ? ........ !!? ? ? ? ? ? ....... </li></ul>Regression Therapy – Contentful Testing
    10. 10. How do Rails devs test the view? <ul><li>Functional test, plus assert_select </li></ul>Regression Therapy – Contentful Testing
    11. 11. How do Rails devs test the view? <ul><li>Decoupled unit test, plus assert_select </li></ul>Regression Therapy – Contentful Testing
    12. 12. But views are not very assertive <ul><li>Assertion tests </li></ul><ul><ul><li>don’t help exploration </li></ul></ul><ul><ul><li>are great for test-driving, </li></ul></ul><ul><ul><ul><li>but view content doesn’t get much value from this. </li></ul></ul></ul><ul><ul><li>have selective coverage </li></ul></ul><ul><ul><li>are verbose when coverage needs to be high; </li></ul></ul><ul><ul><ul><li>translation takes time and space </li></ul></ul></ul><ul><ul><li>are expensive to update when things change </li></ul></ul>Regression Therapy – Contentful Testing
    13. 13. How much do Rails devs test the view? <ul><li>Not much! </li></ul><ul><li>&quot;I find that the View tests that RSpec and Zentest provide are fragile and provide dubious value. Rarely do they find bugs and often they break for unimportant reasons. I just don't find them to have a good return on investment.“ – Jay Fields, Thoughtworks. </li></ul><ul><li>err.the_blog’s groupthink thought similar </li></ul>Regression Therapy – Contentful Testing
    14. 14. assert_ select Regression Therapy – Contentful Testing
    15. 15. Past life regression <ul><li>Regression tests make some things a bit better </li></ul><ul><ul><li>Any one test is simple to express </li></ul></ul><ul><ul><ul><li>“ the expected output is exactly as in this file” </li></ul></ul></ul><ul><ul><li>Expected output needs no translation </li></ul></ul><ul><ul><ul><li>- it’s just a straight copy </li></ul></ul></ul><ul><li>But they make other things much worse </li></ul><ul><ul><li>Very expensive to maintain, because </li></ul></ul><ul><ul><ul><li>each test is brittle, breaking on minor changes </li></ul></ul></ul><ul><ul><ul><li>a single production code change can break many tests </li></ul></ul></ul>Regression Therapy – Contentful Testing
    16. 16. Regression therapy <ul><li>One step back, two steps forward </li></ul><ul><li>Dampen the noise </li></ul><ul><ul><li>diff the DOM, not the text - normalize </li></ul></ul><ul><ul><li>DRY up tested content </li></ul></ul><ul><li>Smooth the workflow </li></ul><ul><ul><li>easily create tests </li></ul></ul><ul><ul><li>conveniently inspect changes </li></ul></ul><ul><ul><li>quickly accept changes </li></ul></ul><ul><ul><li>… and these should all work well in batch </li></ul></ul>Regression Therapy – Contentful Testing
    17. 17. Contentful plug-in <ul><li>% ruby script/plugin install svn:// </li></ul>Regression Therapy – Contentful Testing
    18. 18. Regression Therapy – Contentful Testing
    19. 19. Regression Therapy – Contentful Testing
    20. 20. Regression Therapy – Contentful Testing
    21. 21. Regression Therapy – Contentful Testing
    22. 22. Regression Therapy – Contentful Testing
    23. 23. Regression Therapy – Contentful Testing
    24. 24. Regression Therapy – Contentful Testing
    25. 25. Regression Therapy – Contentful Testing
    26. 26. Regression Therapy – Contentful Testing
    27. 27. Regression Therapy – Contentful Testing
    28. 28. Regression Therapy – Contentful Testing
    29. 29. Regression Therapy – Contentful Testing
    30. 30. Regression Therapy – Contentful Testing
    31. 31. Regression Therapy – Contentful Testing
    32. 32. Other paths to contentfulness <ul><li>That was just one implementation </li></ul><ul><ul><li>Really, I want to sell the contentful testing pattern more than my particular Rails plug-in </li></ul></ul><ul><li>Capture a normalized from of entire output </li></ul><ul><li>and, automate new/test/review/accept somehow </li></ul><ul><ul><ul><li>others Rails devs might prefer different Rake tasks </li></ul></ul></ul><ul><ul><ul><li>other devs might want something other than Rake </li></ul></ul></ul><ul><ul><ul><li>maybe use some IDE other than the command-line </li></ul></ul></ul>Regression Therapy – Contentful Testing
    33. 33. Pushing contentful testing <ul><li>Leveraging the coverage/upkeep trade-off </li></ul><ul><ul><li>I pay the tax on dense coverage to get value from serendipitous discovery of changes </li></ul></ul><ul><li>Testing more than the view content </li></ul><ul><ul><li>cover for exploration, spiking, and follow-on </li></ul></ul><ul><li>Bonus value in regular assertion tests </li></ul><ul><ul><li>If you’re assertion testing controllers anyway, an extra assert_contentful is almost free </li></ul></ul><ul><ul><li>I (guiltily) trade decoupling for convenience </li></ul></ul>Regression Therapy – Contentful Testing
    34. 34. An obvious contentful usecase <ul><li>Use the machinery as a temporary testing vise </li></ul><ul><ul><ul><li>(term due to Michael Feathers) </li></ul></ul></ul><ul><ul><li>Pin down all view content during a big refactor </li></ul></ul><ul><ul><ul><li>extracting partials and helpers </li></ul></ul></ul><ul><ul><ul><li>refactoring form builders </li></ul></ul></ul><ul><ul><ul><li>changing template system </li></ul></ul></ul><ul><li>CONTENTFUL_AUTO = true </li></ul><ul><ul><li>assert_contentful in every functional test </li></ul></ul>Regression Therapy – Contentful Testing
    35. 35. Regression Therapy – Contentful Testing
    1. A particular slide catching your eye?

      Clipping is a handy way to collect important slides you want to go back to later.