RIIR

I think it's time to consider rewriting this project in Rust: it receives very little maintenance or novel feature development, has had a few memory corruption vulnerabilities discovered through fuzzing over the years, and has a relatively simple C API that could be exposed from a Rust core.

In other words: from an external users' perspective, all that would change is the C API (we'd have to make breaking changes to accommodate internal changes). External users would not need to know Rust.

If we decide to do this, we should do it as a major breaking change, with a new major (3.x) release series.

As part of a potential rewrite, we should reuse an existing high-quality Rust PE parser, rather than writing our own. Two options:

• goblin: https://crates.io/crates/goblin
• object: https://crates.io/crates/object

A rewrite here would also have downstream repercussions for a few of our projects, including winchecksec and uthenticode. It's my opinion that those repercussions would be largely positive, although it's still churn that would need to be performed. It would also be a strong motivating point for rewriting some of those projects in Rust as well.

CCing for opinions: @ekilmer @yardenshafir @ret2libc @alessandrogario

Review the existing C API and compare the proposed goblin and object Rust PE parsers first. Then assess the major-version and downstream impact on winchecksec and uthenticode; done would require a decided rewrite scope, parser choice, compatibility plan, and migration work.