People are obsessed with "coverage" in fuzzing and symbolic execution but it's a mirage
-
Co-signed. Coverage is easy to measure, but “state” is what we should be aspiring to use. It’s just currently expensive to capture state, AFAIK.
Partly agree. State with an array containing 20 As vs 20 Bs may yield 100% identical behavior. There is too much useless state. New paths at least trigger some new behavior. Some interesting work by @GuidoVranken about other feedback like stack depthhttps://guidovranken.com/2018/07/06/vrankenfuzz-a-multi-sensor-multi-generator-mutational-fuzz-testing-engine/ …
-
-
Yes, and to some extent it depends on how you define “state”. I think there is definitely a lot of opportunity for improvement here. Godefroid just published a good review re: current state of the art in Communications of the ACM
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.