a crude way of looking at it would be to instrument the code to measure code coverage. some coverage tools support distinguishing which branches were executed a

a crude way of looking at it would be to instrument the code to measure code coverage. some coverage tools support distinguishing which branches were executed and which were not (at a finer granularity than statement-level coverage). this may not be exactly what you want - it would tell you "the branches were not taken during this run of the program", but not "it is logically impossible for this branch to be taken for any run of the program"