Selenium
Document Actions
Submitted by
jenn.
on 2005-04-14 12:34.
First stabs at using the Selenium automated test framework to do quick "dumb" regression-testing.
As Ross briefly mentioned,
he did the initial startup work for getting Selenium
available on our systems. Selenium is a JavaScript-based web test
automater. There are plenty of automated testers out there, and even
plenty of automated web testers, but I have to say that this
one excites me more than any of the others I've tried. Various people
have done the work necessary to integrate it with Plone, which means
that (after Ross did his thing) it's really a no-brainer for me to go
forward with evaluating it.I confess that I'm a bit curmudgeonly about trying to use bleeding-edge fast-moving open-source software, but Selenium is so well-tailored to the way I think that it's got me subscribing not only to its users and developers mailing lists, but to its cvs commits list!
What does this mean for you? Well, probably not much. For now, it'll mostly replace the quickie tests I run, not the big detailed ones I delegate to others. But in the longer term, if I get the automated tests to be comprehensive enough, we might be able to do some defined percentage of regression testing with Selenium (or whatever), and cut down the suite of test plans that have to be run by humans.
The basic problem is that any test runner that works at such a high level -- the extreme functional-test level, running as a user -- has to be tweaked every time the interface is tweaked. Currently, if a link changes its name or location, the human tester just makes a note of it, and I fix the test plans after the fact. Selenium won't be as forgiving. So it'll be interesting to see in what circumstances this kind of automation can save work, and to what extent it would actually require more work.
Bottom line is that I'm currently working on implementing our "high points test" suite in Selenium, and it looks quite promising!
creating more work
