Effective test automation

From CitconWiki
Revision as of 13:11, 23 October 2012 by Athoshun (talk | contribs) (adding some valuable ideas mentioned in the discussion)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

--- flakytests --- Flaky tests are common, but not inevitable. There are various ways to identify what's causing the tests to appear flaky. Simon Stewart - creator of WebDriver and lead for the Selenium opensource project has spoken and provided material on the topic on several occasions. Here's a link to a brief Q&A where he suggests 3 possible reasons and how to address them http://www.eurostarconferences.com/blog/2012/10/17/test-leaders-live-webinar-qa-with-simon-stewart and here's a longer article he wrote for the Google Testing blog http://googletesting.blogspot.hu/2009/06/my-selenium-tests-arent-stable.html

Also Jonathan Kohl published an article on 'unreproduceable bugs' http://www.kohl.ca/articles/unrepeatablebug.pdf

There's nothing wrong about night-long running through-the-UI tests, but it's recommended to run some quick environment stability tests first so the unstable environment can be fixed before breaking hours of later tests.

Test code is code as well, so it should be treated like that: for example, decoupling tests from the test framework can lower the cost and the effects of a test framework upgrade.