WorksButNotTested

Fuzzbench results: * https://www.fuzzbench.com/reports/experimental/2023-09-02-aflpp/index.html * `aflplusplus_qemu_tcgcov` * `aflplusplus_qemu`

That should work fine for me. Just wondered if any other environment variables caused issues and something more generic was required? If not then similarity to AFL++ is probably more...

Also, does this line take a double-reference to the `event-sink` into the `user_data`? https://github.com/frida/frida-rust/blob/6d287d10c3f248c6d2e9fd1510d1313df1d44084/frida-gum/src/stalker/event_sink.rs#L142 I had a few problems when I copied the approach for the `StalkerObserver` and end up...

I'm not sure. Looks like the PR which mentioned it above was closed and not merged?

From what I can see, it looks like these 3 APIs mentioned above might all have the same issue, that they take a parameter by reference and expect that it's...

> @WorksButNotTested afl-fuzz sets an env var you can get the directory from Do you know which one I should be looking out for? Is it definitely passed to the...

Looks like FRIDA is triggering ASAN while searching for the auxilliary vector on the stack, [here](https://github.com/frida/frida-gum/blob/dc11cd1c7411b1dc7d2472bbdcaf823270294a5c/gum/backend-linux/gumprocess-linux.c#L458). ASAN is triggered by a call to `memmem` which it hooks, so while we...

Try also setting `AFL_DEBUG_CHILD`.

Sorry. Try AFL_FRIDA_VERBOSE. I think @vanhauser-thc is going to fix this so you don't need any extra flags.

What version of FRIDA are you using? It seems that when searching for the auxilliary vector in the stack, it is starting at the end of the allocation, rather than...