Building a Cloud Database from Scratch: Why We Moved from C++ to Rust (2022)

What the "overhead" of unique_ptr is usually attached to is the inability for a compiler to pass the unique_ptr as a value through a register but will instead have to use the stack (memory). And that even doesn't apply in the general case but _only_ for unique_ptrs holding an object with non-trivial copy-constructors or non-trivial destructors. This is due to the platform ABI and not the C++ compiler limitation.

Anyway, calling that an overhead is a far stretch and almost purely theoretical unless someone is able to measure the negative effect of such code transformation in the real-world codebase. And I say this as not particularly heavy user of smart pointers.

Also, many other codegen transformations will not fit into a very limited amount of ABI registers so should we argue about not using those as well?

In my opinion, this was only a "campaign" of Google trying to use a unique_ptr as a leverage to persuade the committee to accept their break-the-world ABI suggestion. We know how it all went.

The camp that thinks Rust is just unnecessary fuss because "you can do it all" in C++ already. They pride themselves with knowing C++ esoterica, and don't like that Rust lowers the entry barrier to writing similar software. They want an exclusive club. They also see ownership as a nuisance. They know you can be equally reckless in Rust too, but they don't like that Rust is so "in your face" about it, too.

The camp that is afraid Rust will marginalise C++ and hence push them towards undesirable (even if high paying) jobs.

There's the camp that's jaded by seeing too many fads come and go, and probably just burnt out at some point. So, they just want to work with what's there and be left alone. They don't want to learn something that they know (incorrectly or not isn't for me to judge here) will be dead in the water soon. And they'll have wasted their energy on it for naught. Or in short, they see it as an unwelcome distraction. This camp also includes "but it has no ISO specification for aviation/etc" too.

I think all these are wrong.

I don't think C++ will be marginalised, it just won't be the default choice for certain domains.

I don't think the extra tooling is a nuisance, if anything in my experience it unburdens the mind from non-domain concerns.

And I also don't think it's a fad, given how productive the language makes people feel, and the fact the Rust Foundation is very serious about the longevity of the project, and that it has many big-player backers.

There's also obviously others too; people come in about 8 billion flavours.

Seven months, 10+ team but let’s average that over time to 10. Let‘s take 20 days/month and we arrive at roughly 200LOC per day. C++, database server code. Interesting…