Whats the impact on fuzzing, say a JS interpreter? We found that optimizing the generator had little impact. I have no doubt that you could beat nautilus, but any reason you didn't eval it? I'm really looking forward on how you implement mutations for feedback fuzzing!
Announcing the world’s fastest (grammar) fuzzer, producing millions of test cases per second: https://arxiv.org/abs/1911.07707 Try out our F1 prototype athttps://pypi.org/project/F1-Fuzzer …
-
-
-
(Actually, we know a thing or two on JS fuzzing with grammars;
@mozdeco’s#LangFuzz is in massive daily use testing JS tools and browsers around the globe. The browser you read this in was likely tested by it.) https://www.usenix.org/conference/usenixsecurity12/technical-sessions/presentation/holler … - Još 2 druga odgovora
Novi razgovor -
-
-
Nice work. High-throughput sampling from the space of valid inputs is absolutely exciting! Was wondering what your source of randomness is? You say that F1 when sampling from the CSS grammar achieves 3x higher throughput than piping /dev/random (79MB/sec vs 23MB/sec).
-
The paper seems to reference this PRNG for F1: http://xoshiro.di.unimi.it/xoshiro256starstar.c …. I don't know how /dev/urandom is implemented in MacOS, but I'm guessing it's much slower.
- Još 3 druga odgovora
Novi razgovor -
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.