Testing of applications


#1

I wanted to hear some ideas and stories of how people are testing their applications.

As the modules get more powerful, they become targets of larger apps, with a larger proportion of hardware independent code. Much of the testing is therefore independent of where it will end up running.

For what I can see from the tools provided by wavecom, there are two main ways to test an application:

  • Install it on the module, look at the traces, check if things are going as they should
  • Using the RTE,which I don’t know how different is from the first option (besides being able to debug) since I don’t use it.

Both ways sound quite awkward to me.

What I’m currently doing, is developing a thin layer between wavecom library calls where I need it (for example, to use timers, accessing the RTC, etc.) and my code so I can run the code both on wavecom and linux (where I develop), write test cases using check and check the code with valgrind (maybe someone thinks it’s overkill, but that’s why I want to hear some opinions… of course, these are my first steps writing C code and I want to be as protected as I can in a so unforgiving language).

Thanks!

Thanks!


#2

Sandi,

I’m having the same concerns, I’m doing some consulting services for a company which has an application with 20K+ lines of code and they manually test it :frowning:

Did you had any luck looking for alternatives? I’m seriously considering your approach of building this layer between the wavecom library calls so I can unit test it on linux. Are you still using this approach? Is it too much trouble?

Thanks!