0

We are developing some a simple tool based on Matlab and Java. The tool has a couple hundreds of functions which are partly dependent on each other.

The problem we are facing is the testing.

For example, it may happen that we update Function A => Functions B will produce an error as it depends on Function A. However, because function A is rarely used by the users, the bug will come up first after two months or more (unless a senior developer remembers the dependency). The best way to avoid all of this, is to have a sort of an automatic testing.

Testing usually depends on an input/output relationship. Unfortunately, the tool visualizes data or generates a graph.

The question: how can we automatically test our tool before each update without having to sit in front of the computer, and check every button manually?

  • Did you mean your tool visualizes data and generates a graph OR your tool does not visualizes data or generates a graph? – Yu Zhang Jul 27 '16 at 8:03
  • @YuZhang yes, it visualizes data and generates graphs – user2536125 Jul 27 '16 at 10:44
3

There are testing frameworks for Matlab. Suppose you can automate testing of this development just like you would any other software product.

I would not check the visualizations, but check the data instead and trust the render to do its job, unless you ran into issues here in the past. If you need to check the visualizations either compare the output with ImageMagick with a known correct state or use a tool like Sikuli, but do this as a last resort, since updating the known state will be time consuming.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.