Information Technology Reference
In-Depth Information
0
obs qual([x1:full/inc,x2:zero/std])
1
obs out_pump1_on
2
obs qual([x1:empty..full/inc,x2:full/inc])
obs qual([x1: ...,x2: ...])
98
... (inconcl)
obs out_pump1_off
99
ctr in_water_req(1)
102
obs out_pump2_on
104
obs qual([x1:empty..full/inc,x2:empty..reserve/dec])
106 (pass)
Fig. 10. Example test case for our two-tank system extracted from the test graph
(product LTS) depicted in Figure 7 by applying approach A1 . Several inconclusive
states have been merged for the sake of clarity. Fail verdicts are implicit.
case for our two-tank example has been extracted from the test graph (prod-
uct LTS) depicted in Figure 7. Note that several inconclusive states have been
merged for the sake of clarity and that fail verdicts are implicit. This first ap-
proach yields one test case for each path to an unsafe state and hence a possibly
huge number of test cases is produced.
Our second approach A2 significantly reduces the number of generated test
cases. It works directly on the test graph and in contrast to A1 produces adap-
tive test cases, i.e., the controllables depend on previously made observations.
Figure 11 shows a test case for the two-tank system that has been generated in
this way. The test case contains more than just one path leading to the unsafe
state, which is numbered by 6. For the sake of simplicity, qual events are inter-
nal, i.e., not visible. Our third approach A3 results in even less test cases, since
it works globally over all injected faults. Again, we successively generate test
cases in order to generate a test suite which kills all mutants. But this strategy
checks whether an already generated test case is able to kill other mutants for
which in turn no further test cases need to be generated and added to the test
suite. Again, the extracted test cases are adaptive and have the same structure
as those resulting from A2 . The evaluation of these three approaches is ongoing
work. It has to be assessed, which of the strategies provides sucient coverage
on the implementation under test in order to gain a satisfactory test suite.
 
Search WWH ::




Custom Search