Comments on: iPhone testing, the lay of the land http://www.huesler-informatik.ch/2009/08/18/iphone-testing-the-lay-of-the-land/ Software development and consulting Sat, 12 Feb 2011 18:24:32 +0000 hourly 1 http://wordpress.org/?v=3.0.5 By: Patrick Hüsler http://www.huesler-informatik.ch/2009/08/18/iphone-testing-the-lay-of-the-land/comment-page-1/#comment-19 Patrick Hüsler Thu, 03 Sep 2009 23:46:21 +0000 http://www.huesler-informatik.ch/?p=137#comment-19 I figured that this is due to the dynamic implementation of selectors and matchers. To me, warnings are a bit like broken windows. If you don't fix them, it'll get worse eventually. I understand your point. This makes the implementation easier. It works as advertised, but still there is this notion, that a test framework should not produce that many warnings. A quick google search reveals, that this is quiet a common problem with objective-c. I prefer runtime errors over compile warnings when it comes to tests. You'll notice the errors right away when running the test anyway. To me, that seems like a good trade off. I figured that this is due to the dynamic implementation of selectors and matchers. To me, warnings are a bit like broken windows. If you don’t fix them, it’ll get worse eventually.

I understand your point. This makes the implementation easier. It works as advertised, but still there is this notion, that a test framework should not produce that many warnings. A quick google search reveals, that this is quiet a common problem with objective-c.

I prefer runtime errors over compile warnings when it comes to tests. You’ll notice the errors right away when running the test anyway. To me, that seems like a good trade off.

]]> By: Brian http://www.huesler-informatik.ch/2009/08/18/iphone-testing-the-lay-of-the-land/comment-page-1/#comment-18 Brian Thu, 03 Sep 2009 17:34:27 +0000 http://www.huesler-informatik.ch/?p=137#comment-18 Hi there...great post...was hoping to shed some light on why the warnings are present. Currently UISpec leverages the dynamic nature of ObjectiveC similar to Ruby's "method_missing", which is a very powerful. The entire view filtering logic relies on this. I guess we could create a base class that has the most common properties defined across all UIViews. But this would just get rid of the warnings, you may still get a runtime error that your view doesn't respond to a selector. Hope this helps! Best, Brian Hi there…great post…was hoping to shed some light on why the warnings are present.

Currently UISpec leverages the dynamic nature of ObjectiveC similar to Ruby’s “method_missing”, which is a very powerful. The entire view filtering logic relies on this.

I guess we could create a base class that has the most common properties defined across all UIViews. But this would just get rid of the warnings, you may still get a runtime error that your view doesn’t respond to a selector.

Hope this helps!

Best,

Brian

]]>