Replace HashMap implementation with SwissTable (as an external crate) #58623

Amanieu · 2019-02-21T21:04:32Z

This is the same as #56241 except that it imports hashbrown as an external crate instead of copying the implementation into libstd.

This includes a few API changes (all unstable):

try_reserve is added to HashSet.
Some trait bounds have been changed in the raw_entry API.
search_bucket has been removed from the raw_entry API (doesn't work with SwissTable).

rust-highfive · 2019-02-21T21:04:42Z

r? @nikomatsakis

(rust_highfive has picked a reviewer for you, use r? to override)

Amanieu · 2019-02-21T21:05:07Z

r? @alexcrichton

gnzlbg · 2019-02-21T23:16:33Z

Is this blocked on the review of #56241 ?

ghost · 2019-02-23T11:25:49Z

Can we do a crater run to see if there are any regressions?

Amanieu · 2019-02-24T19:16:19Z

Is this blocked on the review of #56241 ?

Sort-of. I've been including most of the feedback from that thread into hashbrown itself.

alexcrichton · 2019-02-25T15:46:59Z

@bors: try

While review progresses let's in parallel start some programmatic analysis! (aka perf runs, crater runs, etc)

Amanieu · 2019-02-25T19:42:27Z

This issue should probably be addressed before merging: https://github.com/Amanieu/hashbrown/issues/47

alexcrichton · 2019-02-25T20:32:24Z

@bors: try

bors · 2019-02-25T20:32:35Z

⌛ Trying commit 76003ba0c21df9b8bc889192d26c04f2568623d5 with merge abade53a649583e40ed07c26ee10652703f09b58...

bors · 2019-02-26T00:32:35Z

💥 Test timed out

alexcrichton · 2019-02-26T02:31:03Z

@rust-timer build abade53a649583e40ed07c26ee10652703f09b58

RalfJung · 2019-04-24T07:22:14Z

@rust-lang/infra shouldn't it have created an issue for the Miri toolstate failure?

kennytm · 2019-04-24T07:41:50Z

@RalfJung "test-fail" won't create an issue since there could be false positive, only "build-fail" would create an issue.

25: Release 0.5.0, deprecated r=cuviper a=cuviper This is the last subtree merge before `hashbrown` replaced the `std` types in rust-lang/rust#58623. I am also marking this crate deprecated -- folks who want to parallelize these types going forward should use `hashbrown`'s own "rayon" feature. Co-authored-by: Florian Hartwig <[email protected]> Co-authored-by: Meltinglava <[email protected]> Co-authored-by: Mazdak Farrokhzad <[email protected]> Co-authored-by: Andy Russell <[email protected]> Co-authored-by: bors <[email protected]> Co-authored-by: Guillaume Gomez <[email protected]> Co-authored-by: Steven Fackler <[email protected]> Co-authored-by: John Kåre Alsaker <[email protected]> Co-authored-by: Hidehito Yabuuchi <[email protected]> Co-authored-by: Corey Farwell <[email protected]> Co-authored-by: kennytm <[email protected]> Co-authored-by: Alexander Regueiro <[email protected]> Co-authored-by: Stein Somers <[email protected]> Co-authored-by: Ryan Marcus <[email protected]> Co-authored-by: Mark Rousskov <[email protected]> Co-authored-by: Wiktor Kuchta <[email protected]> Co-authored-by: Anthony Ramine <[email protected]> Co-authored-by: Scott McMurray <[email protected]>

This includes a `size_of` regression for a few DOM types, due to rust-lang/rust#58623 which replaces the implementation of `HashMap` in the standard library to Hashbrown. Although `size_of<HashMap>` grows, it’s not obvious how total memory usage is going to be impacted: Hashbrown only has one `u8` instead of one `usize` of overhead per hash table bucket for storing (part of) a hash, and so might allocate less memory itself. Hashbrown also typically has better run time performance: https://github.com/rust-lang/hashbrown#performance Still, I’ve filed rust-lang/hashbrown#69 about potentially reducing the `size_of<HashMap>` regression.

Upgrade to rustc 1.36.0-nightly (e305df184 2019-04-24) This includes a `size_of` regression for a few DOM types, due to rust-lang/rust#58623 which replaces the implementation of `HashMap` in the standard library to Hashbrown. Although `size_of<HashMap>` grows, it’s not obvious how total memory usage is going to be impacted: Hashbrown only has one `u8` instead of one `usize` of overhead per hash table bucket for storing (part of) a hash, and so might allocate less memory itself. Hashbrown also typically has better run time performance: https://github.com/rust-lang/hashbrown#performance Still, I’ve filed rust-lang/hashbrown#69 about potentially reducing the `size_of<HashMap>` regression.  --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/23263)

jethrogb · 2019-04-26T17:04:21Z

src/test/ui/issues/issue-21763.stderr

-   = note: required because it appears within the type `std::marker::PhantomData<(std::rc::Rc<()>, std::rc::Rc<()>)>`
-   = note: required because it appears within the type `std::collections::hash::table::RawTable<std::rc::Rc<()>, std::rc::Rc<()>>`
+   = note: required because of the requirements on the impl of `std::marker::Send` for `hashbrown::raw::RawTable<(std::rc::Rc<()>, std::rc::Rc<()>)>`
+   = note: required because it appears within the type `hashbrown::map::HashMap<std::rc::Rc<()>, std::rc::Rc<()>, std::collections::hash_map::RandomState>`


Do I understand from this UI test update that the compiler may "leak" to the user that std::collections::HashMap is actually implemented in the hashbrown crate? I think this might be confusion to users.

It’s not especially different from the current situation of exposing implementation types like RawTable or BTreeNode

Except those are all types from std itself. So the user might see a private type name but at least they know where it's coming from.

This kind of thing already caused a lot of confusion with the "simple" issue of re-exporting of core via std. I feel like I haven't seen that problem in a while; is there some existing solution we can use?

There is a way to do this reexport that produces a path that starts with std, I just don't know how. There are reexports of core / alloc items in std that handle this correctly.

std::HashMap isn't a re-export, though. It's a newtype around hashbrown::Hashmap

You could also reexport hashbrown as an "implementation detail" or something, but I don't know if you can get rustc to use std paths in error messages without it being publicly reachable, which we'd want to avoid.

You can also, for this error in particular, have unsafe impl<K: Send, V: Send, S: Send> Send for HashMap<K, V, S> {} but I'm not sure that's a good idea.

I don't know if you can get rustc to use std paths in error messages without it being publicly reachable, which we'd want to avoid.

What if you re-export them as pub items, but the re-export itself is not actually publicly reachable?

Restioson · 2019-04-30T14:13:58Z

Re: the fxhash crate @archseer. The maintainer seems to be inactive, so no PRs are being reviewed or accepted. Currently it also does not support no_std.

Pkgsrc changes: * NetBSD/sparc64 disabling of "packed" removed ("packed" removed upstream) * Adapt src_libstd_build.rs patch, update sed'ing of .cargo-checksum.json Build verified on NetBSD 8.0/amd64. Upstream changes: Version 1.36.0 (2019-07-04) ========================== Language -------- - [Non-Lexical Lifetimes are now enabled on the 2015 edition.][59114] - [The order of traits in trait objects no longer affects the semantics of that object.][59445] e.g. `dyn Send + fmt::Debug` is now equivalent to `dyn fmt::Debug + Send`, where this was previously not the case. Libraries --------- - [`HashMap`'s implementation has been replaced with `hashbrown::HashMap` implem entation.][58623] - [`TryFromSliceError` now implements `From<Infallible>`.][60318] - [`mem::needs_drop` is now available as a const fn.][60364] - [`alloc::Layout::from_size_align_unchecked` is now available as a const fn.][6 0370] - [`String` now implements `BorrowMut<str>`.][60404] - [`io::Cursor` now implements `Default`.][60234] - [Both `NonNull::{dangling, cast}` are now const fns.][60244] - [The `alloc` crate is now stable.][59675] `alloc` allows you to use a subset of `std` (e.g. `Vec`, `Box`, `Arc`) in `#![no_std]` environments if the environment has access to heap memory allocation. - [`String` now implements `From<&String>`.][59825] - [You can now pass multiple arguments to the `dbg!` macro.][59826] `dbg!` will return a tuple of each argument when there is multiple arguments. - [`Result::{is_err, is_ok}` are now `#[must_use]` and will produce a warning if not used.][59648] Stabilized APIs --------------- - [`VecDeque::rotate_left`] - [`VecDeque::rotate_right`] - [`Iterator::copied`] - [`io::IoSlice`] - [`io::IoSliceMut`] - [`Read::read_vectored`] - [`Write::write_vectored`] - [`str::as_mut_ptr`] - [`mem::MaybeUninit`] - [`pointer::align_offset`] - [`future::Future`] - [`task::Context`] - [`task::RawWaker`] - [`task::RawWakerVTable`] - [`task::Waker`] - [`task::Poll`] Cargo ----- - [Cargo will now produce an error if you attempt to use the name of a required dependency as a feature.][cargo/6860] - [You can now pass the `--offline` flag to run cargo without accessing the netw ork.][cargo/6934] You can find further change's in [Cargo's 1.36.0 release notes][cargo-1-36-0]. Clippy ------ There have been numerous additions and fixes to clippy, see [Clippy's 1.36.0 rel ease notes][clippy-1-36-0] for more details. Misc ---- Compatibility Notes ------------------- - [`std::arch::x86::_rdtsc` returns `u64` instead of `i64`][stdsimd/559] - [`std::arch::x86_64::_mm_shuffle_ps` takes an `i32` instead of `u32` for `mask `][stdsimd/522] - With the stabilisation of `mem::MaybeUninit`, `mem::uninitialized` use is no longer recommended, and will be deprecated in 1.38.0. [60318]: rust-lang/rust#60318 [60364]: rust-lang/rust#60364 [60370]: rust-lang/rust#60370 [60404]: rust-lang/rust#60404 [60234]: rust-lang/rust#60234 [60244]: rust-lang/rust#60244 [58623]: rust-lang/rust#58623 [59648]: rust-lang/rust#59648 [59675]: rust-lang/rust#59675 [59825]: rust-lang/rust#59825 [59826]: rust-lang/rust#59826 [59445]: rust-lang/rust#59445 [59114]: rust-lang/rust#59114 [cargo/6860]: rust-lang/cargo#6860 [cargo/6934]: rust-lang/cargo#6934 [`VecDeque::rotate_left`]: https://doc.rust-lang.org/std/collections/struct.VecDeque.html#method.rotate_left [`VecDeque::rotate_right`]: https://doc.rust-lang.org/std/collections/struct.VecDeque.html#method.rotate_right [`Iterator::copied`]: https://doc.rust-lang.org/std/iter/trait.Iterator.html#tymethod.copied [`io::IoSlice`]: https://doc.rust-lang.org/std/io/struct.IoSlice.html [`io::IoSliceMut`]: https://doc.rust-lang.org/std/io/struct.IoSliceMut.html [`Read::read_vectored`]: https://doc.rust-lang.org/std/io/trait.Read.html#method.read_vectored [`Write::write_vectored`]: https://doc.rust-lang.org/std/io/trait.Write.html#method.write_vectored [`str::as_mut_ptr`]: https://doc.rust-lang.org/std/primitive.str.html#method.as_mut_ptr [`mem::MaybeUninit`]: https://doc.rust-lang.org/std/mem/union.MaybeUninit.html [`pointer::align_offset`]: https://doc.rust-lang.org/std/primitive.pointer.html#method.align_offset [`future::Future`]: https://doc.rust-lang.org/std/future/trait.Future.html [`task::Context`]: https://doc.rust-lang.org/beta/std/task/struct.Context.html [`task::RawWaker`]: https://doc.rust-lang.org/beta/std/task/struct.RawWaker.html [`task::RawWakerVTable`]: https://doc.rust-lang.org/beta/std/task/struct.RawWakerVTable.html [`task::Waker`]: https://doc.rust-lang.org/beta/std/task/struct.Waker.html [`task::Poll`]: https://doc.rust-lang.org/beta/std/task/enum.Poll.html [clippy-1-36-0]: https://github.com/rust-lang/rust-clippy/blob/master/CHANGELOG.md#rust-136 [cargo-1-36-0]: https://github.com/rust-lang/cargo/blob/master/CHANGELOG.md#cargo-136-2019-07-04 [stdsimd/522]: rust-lang/stdarch#522 [stdsimd/559]: rust-lang/stdarch#559

This is the hash map implementation now used in the Rust standard library: * https://blog.rust-lang.org/2019/07/04/Rust-1.36.0.html#a-new-hashmapk-v-implementation * rust-lang/rust#58623 * https://crates.io/crates/hashbrown Differential Revision: https://phabricator.services.mozilla.com/D71740

This is the hash map implementation now used in the Rust standard library: * https://blog.rust-lang.org/2019/07/04/Rust-1.36.0.html#a-new-hashmapk-v-implementation * rust-lang/rust#58623 * https://crates.io/crates/hashbrown Differential Revision: https://phabricator.services.mozilla.com/D71740 UltraBlame original commit: 06dcf8ddbfdefed44cb653c7bf533196ae360da3

rust-highfive assigned nikomatsakis Feb 21, 2019

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Feb 21, 2019