Move sancov_cmp.c into sancov_cmp.rs

[addisoncrump]

sancov_cmp.c does not need to be written in C; we can port this to Rust to unify our codebase some.

Moreover, this addresses some strange linkage behaviour we've been seeing for some time that affects linkage against targets from a .a. I re-encountered this while building an example for #1904 which prevented some symbols from being globally visible when using LibAFL as a .so. Version scripts, visibility attributes, and other strategies did not resolve this, so I think Rust does something internally here to enforce that only Rust-defined symbols are globally visible.

Related, I think we can and should move the other C files in libafl_targets to Rust; most of this code is copied from historical versions of AFL++ and could be written more effectively/idiomatically for Rust.

[addisoncrump]

cc @domenukk, @andreafioraldi for input here.

[addisoncrump]

bump @domenukk, @andreafioraldi: can we consider converting all C sources to Rust for the major refactor, where possible/appropriate? There's a lot of legacy code hanging around that could be improved.

[andreafioraldi]

Yes, more Rust is better, be careful that many things are in C because of weak symbols not supported in Rust Il gio 28 mar 2024, 17:33 Addison Crump ***@***.***> ha scritto:

…

bump @domenukk <https://github.com/domenukk>, @andreafioraldi <https://github.com/andreafioraldi>: can we consider converting all C sources to Rust for the major refactor, where possible/appropriate? There's a lot of legacy code hanging around that could be improved. — Reply to this email directly, view it on GitHub <#1937 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AD3LJ6UUIHCYICOTJLPAPNDY2QZ5RAVCNFSM6AAAAABEVGBCC6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMRVGY2DAOBTG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>