(I'm defining product code to be all parts of the product not specifically written by QA for the purpose of testing.)
What are some of things I should be aware of if I am integrating product code into the suite of QA automated testing code. That is, in order to reduce redundant development, the QA tools will use components that are part of the product code.
What are some of the concerns with tying this code together?
One of the immediate ones that comes to mind, for me, is that if an error occurs in my test library due to a change in the product code, it doesn't necessarily mean the product is broken -- since engineering is not responsible for the QA library they may update many components at the same time with a change that introduces errors into the QA system (not intentionally obviously).
I already understand the notion of false verification by using code to test itself, but I am not using code in this manner. I am strictly using components as utilities to test other components (there is no testing of the components being used in these test cases).
How about time and size. The extra code will add processing time and depending on the code could really slow down a program. Also it will add to the total size of the package. For instance, if the code fits comfortablly on a cd and then you add code you can run the risk on it not fitting on a cd anymore. Also there not only effecting the size but now your cost. Two cd's instead of one.