This adds unittests for the SalityBotnet-attack. These are separate from the rest of the unittests, because, as we discussed in the meetings, we had quite a few problems with the attack generating different results depending on the operating system and the order in which tests were run.
While the actual bug seems to be in libtins (and, in fact, a libtins update on macOS changed the hashes we got), we were able to get rid of the problem by removing the packets triggering this behavior from the resource pcap. We removed 2 IGMP and 3 ICMP packets, all of which shouldn't be crucial to the actual botnet behavior.
With the changes in this branch, we're now getting reproducible results on all our runs across our test systems.
This adds unittests for the SalityBotnet-attack. These are separate from the rest of the unittests, because, as we discussed in the meetings, we had quite a few problems with the attack generating different results depending on the operating system and the order in which tests were run.
While the actual bug seems to be in libtins (and, in fact, a libtins update on macOS changed the hashes we got), we were able to get rid of the problem by removing the packets triggering this behavior from the resource pcap. We removed 2 IGMP and 3 ICMP packets, all of which shouldn't be crucial to the actual botnet behavior.
With the changes in this branch, we're now getting reproducible results on all our runs across our test systems.
This adds unittests for the SalityBotnet-attack. These are separate from the rest of the unittests, because, as we discussed in the meetings, we had quite a few problems with the attack generating different results depending on the operating system and the order in which tests were run.
While the actual bug seems to be in libtins (and, in fact, a libtins update on macOS changed the hashes we got), we were able to get rid of the problem by removing the packets triggering this behavior from the resource pcap. We removed 2 IGMP and 3 ICMP packets, all of which shouldn't be crucial to the actual botnet behavior.
With the changes in this branch, we're now getting reproducible results on all our runs across our test systems.