A rewrite would tackle the classic C-style W^X problems, at the expense of maintainability and focus on logic bugs and general oversights from inelegant (or lack of) design. And I'd bet PHP (and popular PHP software for reasons rooted in the language) still got plenty of the latter. You know, the "oops, that filename was client provided" kind of code execution that no amount of (un)safe blocks can help with.
Bun isn't an interpreter. It uses JavaScriptCore for that, which, you guessed it, is C++. As far as I know they have no plans of rewriting it in rust either. Your hot take doesn't seem to apply here.
A reminder, because it really needs to be said, that they spent millions of dollars in tokens for the claimed 11 day sprint, and then spent an additional 3.5 months fixing hundreds of issues it introduced before actually releasing it. Long-term maintainability remains to be seen.
Surely you realize that the codebase behind the language powering 60% of the web is a bit more complicated than an experimental JS runtime?
it doesn't have to be done all at once
A rewrite would tackle the classic C-style W^X problems, at the expense of maintainability and focus on logic bugs and general oversights from inelegant (or lack of) design. And I'd bet PHP (and popular PHP software for reasons rooted in the language) still got plenty of the latter. You know, the "oops, that filename was client provided" kind of code execution that no amount of (un)safe blocks can help with.
Bun isn't an interpreter. It uses JavaScriptCore for that, which, you guessed it, is C++. As far as I know they have no plans of rewriting it in rust either. Your hot take doesn't seem to apply here.
A reminder, because it really needs to be said, that they spent millions of dollars in tokens for the claimed 11 day sprint, and then spent an additional 3.5 months fixing hundreds of issues it introduced before actually releasing it. Long-term maintainability remains to be seen.