fix: check RAND_bytes during shred

[nanookclaw]

What changed

shredFile now checks the return value from RAND_bytes before writing each random overwrite chunk. If OpenSSL reports a failure, the function logs the failed pass, closes the file handle, and returns false instead of continuing with stale buffer contents and later deleting the file.

This mirrors the existing failure handling around seek/write operations in the same loop and leaves the zero-fill pass unchanged.

Closes #7.

Verification

• git diff --check

I could not run the full Windows/MSVC/vcpkg test workflow from this Linux environment, and clang-format is not installed here.

[lextpf]

Hi @nanookclaw,

Thanks for jumping in so quickly, having another pair of eyes on this is always appreciated! 😊
Your changes look good and you have my approval! I'll merge this back ASAP

If you'd like to dive in deeper as a contributer:

The project is a bit opaque about where assets live and it's a bit tricky to set up your own assets which means runtime testing can be quite cumbersome, if you'd like to test locally yourself here are two easy ways to get you up and running in case you want to be more hands-on for future issues:

What you might need	How to get it	Notes
Full pre-built project from your branch	➜ Add me on Discord (@lextpf).	Unzip anywhere and run the executable.
Only the assets (if you want to build yourself)	➜ Add me on Discord (@lextpf).	After cloning the repo, drop the assets/ folder from this ZIP into the repo root, then build.

Let me know if you need support and we can arrange something, and thanks again for volunteering and helping improve Seal, you're always welcome to work on future issues! 🦭

Alex

[nanookclaw]

Appreciate the review and the setup notes. I’ll wait for the current CI run to settle; if anything fails, I’ll follow up on this branch.

I’ll skip Discord for now so the contribution trail stays on GitHub, but the asset-layout context is useful for future Seal issues.