Programming

The article says this isn’t to affect existing code.

Did you read the article at all?

“Putting all new code aside, fortunately, neither this document nor the U.S. government is calling for an immediate migration from C/C++ to Rust — as but one example,” he said. “CISA’s Secure by Design document recognizes that software maintainers simply cannot migrate their code bases en masse like that.”

Companies have until January 1, 2026, to create memory safety roadmaps.

All they are asking for by that date is a roadmap for dealing with memory safety issues, not rewrite everything.

That sounds like a comment written by somebody who has no idea what the article says

Solid detective work Lou.

DARPA has unironically been funding a tool that purports to translate C / C++ into Rust...

https://www.darpa.mil/program/translating-all-c-to-rust

It’s one backed by a lot of data. One example is from the Android project.

The percent of vulnerabilities caused by memory safety issues continues to correlate closely with the development language that’s used for new code. Memory safety issues, which accounted for 76% of Android vulnerabilities in 2019, and are currently 24% in 2024, well below the 70% industry norm, and continuing to drop.

https://security.googleblog.com/2024/09/eliminating-memory-safety-vulnerabilities-Android.html

There’s an argument that critical infrastructure software vendors are already meeting standards for basic, non-memory related items. Yes, there are other categories, but memory safety is one that’s harder to verify. Moving to memory safe languages is an ensure a category of correctness. This excludes usage of unsafe escape hatches.

I have never seen a single C++ codebase do that. It helps but it's not a practical full solution.

Using smart pointers doesn’t eliminate the memory safety issue, it merely addresses one aspect of it. Even with smart pointers, nothing is preventing you from passing references and using them after they’re freed.

I wonder how many issues rewriting everything in another language will create?

I get what you're saying, but I think the issue with optional memory safety features is that it's hard to be sure you're using it in all the places and hard to maintain that when someone can add a new allocation in the future, etc. It's certainly doable, and maybe some static analysis tools out there can prove it's all okay.

Whereas with Rust, it's built from the ground up to prove exactly that, plus other things like no memory being shared between threads by accident etc. Rust makes it difficult and obvious to do the wrong thing, rather than that being the default.

From the original document:

Software manufacturers should build products in a manner that systematically prevents the introduction of memory safety vulnerabilities, such as by using a memory safe language or hardware capabilities that prevent memory safety vulnerabilities. Additionally, software manufacturers should publish a memory safety roadmap by January 1, 2026.

My interpretation is that smart pointers are allowed, as long it’s systematically enforced. Switching to a memory safe language is just one example.