May we have a comment explaining why the library implements its own fixed length hashmap?

[makoConstruct]

around about here

string-cache/src/atom.rs

Line 43 in 4f8310c

buckets: [Option<Box<StringCacheEntry>>; NB_BUCKETS],

I'm mostly just curious. I wonder if it would be worth considering using a parallel hashmap? In my own project, I'm eventually going to need to intern things other than strings. Is there a reason the generic interner from the syntax crate wasn't appropriate for servo's strings? Could the generic interner be improved to the point of being suitable?

[Manishearth]

Is there a reason the generic interner from the syntax crate wasn't appropriate for servo's strings?

It's not something that's exposed (it's unstable).

---

The custom hashtable is presumably because we don't expect hashes to clash often so we're using a chained hashtable instead of a probing one. I'm not sure.

[SimonSapin]

I started typing up how a custom hash map allows Atom::drop to remove a map entry a dynamic atom’s reference count hits zero, but now I think that’d also work with a standard HashMap, only the refcounting needs to be custom.

So I think HashMap would probably work. I don’t know if it’d perform better or worse.