Anay
2c547e60f4
Inet framework
|
пре 6 година | |
|---|---|---|
| .. | ||
| README | пре 6 година | |
| diffingerprints.py | пре 6 година | |
| diffingerprints.sh | пре 6 година | |
| ethernet-bus-reconnect.csv | пре 6 година | |
| ethernet-bus-reconnect.ini | пре 6 година | |
| ethernet-bus-reconnect.xml | пре 6 година | |
| ethernet-hub-reconnect.csv | пре 6 година | |
| ethernet-hub-reconnect.ini | пре 6 година | |
| ethernet-hub-reconnect.xml | пре 6 година | |
| ethernet-hub.csv | пре 6 година | |
| ethernet-hub.ini | пре 6 година | |
| ethernet-switch.csv | пре 6 година | |
| ethernet-switch.ini | пре 6 година | |
| ethernet-twohosts.csv | пре 6 година | |
| ethernet-twohosts.ini | пре 6 година | |
| examples-TODO.csv_off | пре 6 година | |
| examples.csv | пре 6 година | |
| external-tcpip.csv | пре 6 година | |
| fingerprints | пре 6 година | |
| gen_runallexamples.pl | пре 6 година | |
| manet.csv.unstable | пре 6 година | |
| multi.csv | пре 6 година | |
| multi.csv.unstable | пре 6 година | |
| runDefaultTests.sh | пре 6 година | |
| runWirelessTests.sh | пре 6 година | |
| showcases.csv | пре 6 година | |
| test-branch | пре 6 година | |
| tutorials.csv | пре 6 година | |
| voipstream.csv | пре 6 година | |
README
This folder contains fingerprint-based tests for various models.
When a fingerprint test passes, that indicates that with very high
probability, the simulation has followed the same trajectory as when
the fingerprint was recorded, i.e. the times and modules of the events
were the same. That is, a passing fingerprint tests means that the
simulation model logic works the same as before.
However, a passing test does not guarantee that e.g. result recording
has not changed.
Fingerprints are fragile to NED changes and parameter value changes.
When a fingerprint test fails, the simulation's correctness has to be
verified by some other means, and the fingerprints in the tests
updated.
When a fingerprint test passes, that indicates that with very high
probability, the simulation has followed the same trajectory as when
the fingerprint was recorded, i.e. the times and modules of the events
were the same. That is, a passing fingerprint tests means that the
simulation model logic works the same as before.
However, a passing test does not guarantee that e.g. result recording
has not changed.
Fingerprints are fragile to NED changes and parameter value changes.
When a fingerprint test fails, the simulation's correctness has to be
verified by some other means, and the fingerprints in the tests
updated.