AMDG On 03/18/2011 10:23 AM, Vicente Botet wrote:
I know that my question is not directly associated to Boost, but it could be interesting to share the best techniques so we can improve the test of some Boost libraries, existing or future.
A transformation complex is complex if I can not associate mentally the output associated to a given input. Think for example on dividing two bit integers.
One technique could be to make the transformation by hand and keep track of the input -> output association. Another technique consist in re-implementing the algorithm and check that both implementations match.
Do you know other techniques? Does some Boost libraries have such a kind of transformations? How they are tested?
* Compute the inverse transformation and check that you get back the original value. e.g. q, r = div(a, b); assert(r + q * b == a); * Combine multiple transformations in different ways which should produce identical results. e.g. assert(a/b/c == a/(b * c) == a/c/b) In Christ, Steven Watanabe