[estebarb]

I'm not sure if I understand it correctly. So, if I have a hashtable, adding or removing an element would invalidate existing pointers to any other element in the hashtable? I guess it makes sense from a memory release POV, but... I end up thinking that for databases using GC or RC is a better approach. Maybe I'm biased, but I have found far easier to work on databases written in C or C#. For that kind of programs I felt Rust overrestrictive and forcing to more dangerous patterns such as replacing pointers with offsets.

[verdagon]

Yep, adding or removing an element would invalidate existing pointers to any other element in the hash table. This is generally regarded as a good thing if your elements are stored contiguously in the hash table, because a resize would cause any existing pointers to dangle. This should be true for C, and might be true for C# if you're using `struct`s which put the data inline (memory's a bit fuzzy on C#'s rules for references to structs though, maybe someone can chime in).

This new approach still requires us to be mindful of our data layout. Not caring about data layout is still definitely a strength of GC and RC. I'm actually hoping to find a way to blend Nick's approach seamlessly with reference counting (preferably without risking panics or deadlocks) to get the best of both worlds, so that we can consider it for Mojo. I consider that the holy grail of memory safety, and some recent developments give me some hope for that!

(Also, I probably shouldn't mention it since it's not ready, but Nick's newest model might have found a way to solve that for separate-chaining hash maps where addresses are stable. We might be able to express that to the type system, which would be pretty cool.)

[jfecher]

> I'm actually hoping to find a way to blend Nick's approach seamlessly with reference counting (preferably without risking panics or deadlocks) to get the best of both worlds, so that we can consider it for Mojo. I consider that the holy grail of memory safety, and some recent developments give me some hope for that!

Ante's approach manages to blend a similar scheme for safe, shared mutability with Rc. There are some examples on the most recent blog post on its website of it. IMO combined with its shared types it emulates high-level GC'd code very well.

[estebarb]

Thanks for the answer. For instance, usually those containers are used as indexes in DBs, so they contain pointers to data, not the data itself. That is an scenario where the references shouldn't be invalidated.

Idk if it may be possible to introduce "semantic monitors". Say, a field within a class and an external container must be updated together. In practice the only time when I have needed to break the single ownership is for keeping internal data views. I know it is safe, but convincing Rust of that is painful.

[codedokode]

Rc is slightly expensive and Gc is super expensive. I would prefer a memory safety for free.

As for C, it leaves verifying the safety up to you, it is not easy at all and I would rather spend time on something else using safe languages.

[ntrel]

My understanding is that RC is relatively expensive in time (especially for atomic RC) but uses a lot less memory than state of the art fast GC. And RC doesn't handle cycles.

[pjmlp]

Alternatives have existed since the 1960's, predating C's invention, and also after it came to be, the problem has been mainstream adoption of UNIX, and the side effects from that.