Andrea Fioraldi

Ownedref is a very dummy and used pattern not worth a crate (we are not JS devs), for the others I agree

> A working (but unorthodox, as it uses klo routines) QEMU full system fuzzer using LibAFL+kloroutines. It currently supports only aarch64 and a simple interface via `svc #0xaf1` to use...

> The biggest point against merging is, fullsystem qemu is highly async/multithreaded in comparision to single threaded usermode qemu. To seperate these differences I chose to use a new crate...

> Is this superseded by #692 ? yes but I will close this after having a full system fuzzer

> tmin for sure a stage, cmin can be a combination of corpus scheduler and a stage that simply saves as objective the testcase taken from the scheduler A naive...

to check if the pass is working simply print the arg in the pass and try to change it. if llvm 14 changed the way to pass args i'm unaware...

I'm aware of this issue, we already approached in AFL++. The user atm can simply call the asan user poisoning API in the harness, but I guess if would be...

> I think we can make a feature flag that are used combined with asan with that flag InProcessExecutor poisons/unpoisons the input region weak symbols will do the trick without...

@landaire isn't `#[linkage = "extern_weak"]` unstable? We have most of this code in C due to it, but maybe they stabilized it and I missed it. Coll project btw, if...

> for OnDiskCorpus, it should collect all the TestCases and objective outputs from all connected brokers. That's not possible because maybe there are clients with different configurations (so different metadata)...