That's quite a read! Looks like emacs has done an amazing job of handling text correctly in the face of quite a few challenges. Including cases where there perhaps is no "correct" choice.

Now, I'm certainly not in the "all C code must be rewritten in Rust because security" camp but it does raise the question: With all this complexity how do I know that pasting text from a web page into emacs (or any editor really) isn't going to trigger an undiscovered vulnerability?

Edit: I guess that's rhetorical question because of course the answer is "you don't".

IMO the only/main reason to rewrite C-parts in Rust is to refactor and untagle the code to generate better incentive for people to work on it, because from what I heard - some C parts of Emacs are of nightmarish complexity, and there are literally only a few on this planet who even want to deal with it.

And the same time, I don't think there's enough people who can be persuaded to think that in a decade or two those Rust-rewritten parts wouldn't become problematic either - who can promise today that Zig for example isn't a better choice for that? Or maybe even some close-to-metal Lisp variant?

C style null terminated strings were a mistake. They are almost never the right answer. Even C itself should start transitioning to length prefixed strings instead.

The answer isn't material; You just rewrite it in Rust, and then if anything goes wrong it's not your fault because you did The Right Thing. Rust cannot fail, it can only be failed.