The actual good news is that we have multiple mature languages that entirely remove the need for users to configure and run fuzzers and sanitizers and remember to always obey arcane rules in order to avoid memory corruption.
We’ve been encouraging people to walk, not run, to the exits and switch from C and C++ to memory-safe equivalents. Now it is clear that the time for walking is over and the remaining holdouts should be at the very least jogging if not running.
I work mostly in compiled managed languages, yet have to keep my C and C++ skills sharp, as there is always this or other scenario that requires something written in them, like the runtimes or compiler tools I depend on, because bootstraping is too much work for most implementors.
> Let's say you rewrite your terminal emulator in Rust from C or C++; you'll first undo decades of bug fixes, compatibility decisions, quirks, etc. which you WILL have to deal with eventually. This essentially puts you back at least a decade, because these issues take users using the software to point out to you before they're fixed. The only benefit of this whole game being that you get memory safety. These teams are never the original teams, which is the source of the problem. A rewrite can only succeed if it's a result of all the learned failures and shortcomings of the original, otherwise it's bound to repeat every single mistake and add more on top.
> And you don't even deliver memory safety to the users of the original this way, because people won't use the rewrite, because it's buggy, incompatible, etc.
I would prefer something be impossible by construction.
There are memory-safe options with little/no performance overhead compared to C/C++.
That's not what I said, don't put words in my mouth.
And yes, programmers need to do better. That's not the answer, but the quality of code that the average programmer outputs is abysmal and could be a lot better simply by giving a shit :D
You can avoid a lot of memory issues in C++ already, yet so many people refuse to use these features (like std::array, boost::synchronized_value, std::unique_ptr). If we used these, we wouldn't solve all memory issues, but we would surely have less.
If you interpret me saying "we should do better" as "we don't need Rust because programmers just need to not make mistakes" then I think the ball is in your court.
Now to get people to use them it's an issue.