Cmplog fails to solve the most simple sample on v3

[wtdcode]

I'm having 100% stability now, and we should resolve the cmplog issue now. Also see AFLplusplus/LibAFL#3238

Currently enabling cmplog fails to solve the most simple case in sample.rs. You might have a try too.

Note it doesn't require AFLplusplus/AFLplusplus#2419 and you can simply do:

afl-fuzz -i /tmp/
input -o /tmp/test -g 8 -G8 -c ./target/release/examples/sample  -- ./target/release/examples/sample @@

Originally posted by @wtdcode in #40

[wtdcode]

4 bytes variant solves immediately, werid

[wtdcode]

I pushed a new variant for sample.rs and you might check. x86 version (4 bytes magic could be solved immediately while 8 bytes magic is not).

[Evian-Zhang]

Hmmm, does this problem remain here? I tried it locally, both 4-byte version and 8-byte version can be solved.

4-byte version cmdline:

./afl-fuzz -m none -i ./input -o ./output -c 0 -- ./target/release/examples/sample @@

8-byte version cmdline:

./afl-fuzz -m none -i ./input -o ./output -c 0 -- ./target/release/examples/sample @@ true

[wtdcode]

Hmmm, does this problem remain here? I tried it locally, both 4-byte version and 8-byte version can be solved.

4-byte version cmdline:

./afl-fuzz -m none -i ./input -o ./output -c 0 -- ./target/release/examples/sample @@
8-byte version cmdline:

./afl-fuzz -m none -i ./input -o ./output -c 0 -- ./target/release/examples/sample @@ true

Wait, why do you have -c 0?

[Evian-Zhang]

-c 0 is exactly the same as -c ./target/release/examples/sample. 0 means use the same target binary.

Wait me for about 10 minutes, I am running a modified CI in my forked repo (this), which includes the 8-byte version. If it got passed, I will send a PR.

[Evian-Zhang]

See #45