Brian Smith

I'm closing this PR because it doesn't really do what we need. Ideally, for each assembly function, we'd have a shim that takes 32-bit arguments with garbage in the high...

Here is my tentative plan: - [ ] Add `doc/supply-chain-integrity.md` (or whatever) that documents the design of the solution we choose. - [ ] Create a new `briansmith/ring-packaged-crate` repository that...

~~I think the change should be reverted and a new release should be made. Then we can have a design decision.~~ ~~(This has broken *ring*'s CI as well.)~~ [Edit: I...

Please ignore my previous comment; I was mistaken about the cause of my CI failure; it wasn't caused by this cc-rs change.

Good points. I think which way is best depends on what the users' goal is. In any case, when I tried to sketch this out, I immediately ran into a...

> In going from a ConfigBuilder to a ServerConfig or ClientConfig, there are a number of other possible builder types along the way, and it's important to go from one...

Why is it important for Rust to support `setjmp`/`longjmp` at all? The motivation says "Users deserve a solution that works and warns or prevents common pitfalls." However, that is not...

> some C libraries emulate exceptions using `setjmp`/`longjmp` and use that as their error reporting mechanism (e.g. Postgres). Some C++ libraries use C++ exceptions too, but Rust doesn't support them...

> If we want Rust to be able to interoperate with all correct and potentially legacy C code out there via its C FFI, we need to support them in...

Here's what compcert's documentation has to say about `setjmp`/`longjmp`: http://compcert.inria.fr/man/manual005.html: > §7.13 Non-local jumps The CompCert C compiler has no special knowledge of the setjmp and longjmp functions, treating them...