Add an Entry API for HashSet

[cuviper]

This adds HashSet::entry and rustc_entry, mimicking the API of HashMap. The "rustc" variants are without the S type parameter, just like the difference on maps. Set entries have fewer methods since there's no mutable value like in maps, but the basic insert/remove/replace functionality is there.

Resolves #274.

[Amanieu]

Could we avoid introducing another rustc_entry? This was a hack for compatibility with std which didn't have a S parameter on the Entry type. I would prefer if we did this "correctly" and had a S parameter for the version used by std.

[cuviper]

I expect that std::collections::hash_set::Entry<T> should look like a simple wrapper on hash_map::Entry<T, ()>, even if that's not really direct because it goes through hashbrown. I think it would be unfortunate if std were inconsistent about whether these Entry types have an S parameter, but that's ultimately a libs-api decision.

If you still want to push for that S, then I'll drop rustc_entry here and keep it in a branch on the side, just in case. :)

[Amanieu]

I'll repeat that Entry not having S was a mistake that we shouldn't carry over to new APIs. The rustc_entry in the hashbrown crate is actually slightly less efficient because it has to reserve space for a value ahead of time before knowing whether a value will be inserted into the map. It is better to defer the reservation until a value is explicitly inserted, which requires the S parameter.

[cuviper]

OK, I've dropped the S-less one.

[Amanieu]

@bors r+

[bors]

📌 Commit af67f7d has been approved by Amanieu

[bors]

⌛ Testing commit af67f7d with merge e54c160...

[bors]

☀️ Test successful - checks-actions
Approved by: Amanieu
Pushing e54c160 to master...

[SUPERCILEX]

Any idea when this will be released?

[Amanieu]

I just published a release.

[SUPERCILEX]

Thanks!