diff options
| author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-28 14:29:10 +0000 |
|---|---|---|
| committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-28 14:29:10 +0000 |
| commit | 2aa4a82499d4becd2284cdb482213d541b8804dd (patch) | |
| tree | b80bf8bf13c3766139fbacc530efd0dd9d54394c /third_party/rust/indexmap | |
| parent | Initial commit. (diff) | |
| download | firefox-upstream.tar.xz firefox-upstream.zip | |
Adding upstream version 86.0.1.upstream/86.0.1upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to '')
25 files changed, 8055 insertions, 0 deletions
diff --git a/third_party/rust/indexmap/.cargo-checksum.json b/third_party/rust/indexmap/.cargo-checksum.json new file mode 100644 index 0000000000..7cf443999b --- /dev/null +++ b/third_party/rust/indexmap/.cargo-checksum.json @@ -0,0 +1 @@ +{"files":{"Cargo.toml":"b2e03e7a524d10bb253021375ccc06a3d2a060d2c7c045a1a60d8bb3739b46a9","LICENSE-APACHE":"a60eea817514531668d7e00765731449fe14d059d3249e0bc93b36de45f759f2","LICENSE-MIT":"ecc269ef87fd38a1d98e30bfac9ba964a9dbd9315c3770fed98d4d7cb5882055","README.rst":"6751ed067142c9445ffb9c201457a40bb73af76aa3c7e525fb415508abec74b1","benches/bench.rs":"dda3c7e3cd68b8f5c74dbd60189e89f0e43339cdc25dae7b1e5a9d25b1fb80a9","benches/faststring.rs":"2472b9d0d52031988cd31401357908fd850d15f1efa9f9761aa6b6779aad1889","build.rs":"558b4d0b9e9b3a44f7e1a2b69f7a7567ea721cd45cb54f4e458e850bf702f35c","src/equivalent.rs":"2e6ae24ef09a09b917f4e2b0f6288f901878e42f5080f61b1bd1afdcc90aba87","src/lib.rs":"4c4fa0d4882387d2c1681fa29a333354230de157101783bea2af6af04951d973","src/macros.rs":"cb2e9742b7e2bcadc4eb356a4ca4eb3317d1a80e293f37a3cd5a0155c5347c50","src/map.rs":"21c0c04acfb25d526da41b447b2e572d14e52778b84910514869d1cd2dce267c","src/map/core.rs":"9d9df231fbc02e3539429451c6f8cfd142c97d6e47b0c1f58672f75e3806d578","src/map/core/raw.rs":"9842320404357ec15e536240a9ab32d969e61f31508d3c485c0c76077618ef63","src/mutable_keys.rs":"99fdec7c0182902ba19524f342c220b51475bcd41c44c2cb2c3f62eacb029335","src/rayon/map.rs":"cdbe6cad8e5753aca3f07b621d2bb3b480f1ae3e10697568b8c16aeb865de65b","src/rayon/mod.rs":"bf2b0fc074f20135a6734db600b04828e88dd263b3e361606be4efe8e916eafc","src/rayon/set.rs":"789efd10d41f03fa73268bae8afdc3b6582767420218a4b28f3c83c528ef726c","src/serde.rs":"2f7ce29e63e92ddbe14d3ad0b7e83bdc9662b226565ec0aa707d5cd92cb1e263","src/set.rs":"a88c5e902e8c67d014c7c1680a7e4b9156af67d2ac58e7dcf96698b742abc88b","src/util.rs":"c415261cff9f610d7331192feba0839cb05e04d3d375a5fa2f8190a29661994e","tests/equivalent_trait.rs":"efe9393069e3cfc893d2c9c0343679979578e437fdb98a10baefeced027ba310","tests/macros_full_path.rs":"c33c86d7341581fdd08e2e6375a4afca507fa603540c54a3b9e51c4cd011cd71","tests/quick.rs":"6efe8c6dfa5bdd466b3f5c491d9dfbf57fd730301849a710abff9358968218c5","tests/tests.rs":"f6dbeeb0e2950402b0e66ac52bf74c9e4197d3c5d9c0dde64a7998a2ef74d327"},"package":"55e2e4c765aa53a0424761bf9f41aa7a6ac1efa87238f59560640e27fca028f2"}
\ No newline at end of file diff --git a/third_party/rust/indexmap/Cargo.toml b/third_party/rust/indexmap/Cargo.toml new file mode 100644 index 0000000000..96876f030e --- /dev/null +++ b/third_party/rust/indexmap/Cargo.toml @@ -0,0 +1,75 @@ +# THIS FILE IS AUTOMATICALLY GENERATED BY CARGO +# +# When uploading crates to the registry Cargo will automatically +# "normalize" Cargo.toml files for maximal compatibility +# with all versions of Cargo and also rewrite `path` dependencies +# to registry (e.g., crates.io) dependencies +# +# If you believe there's an error in this file please file an +# issue against the rust-lang/cargo repository. If you're +# editing this file be aware that the upstream Cargo.toml +# will likely look very different (and much more reasonable) + +[package] +edition = "2018" +name = "indexmap" +version = "1.6.0" +authors = ["bluss", "Josh Stone <cuviper@gmail.com>"] +build = "build.rs" +description = "A hash table with consistent order and fast iteration.\n\nThe indexmap is a hash table where the iteration order of the key-value\npairs is independent of the hash values of the keys. It has the usual\nhash table functionality, it preserves insertion order except after\nremovals, and it allows lookup of its elements by either hash table key\nor numerical index. A corresponding hash set type is also provided.\n\nThis crate was initially published under the name ordermap, but it was renamed to\nindexmap.\n" +documentation = "https://docs.rs/indexmap/" +keywords = ["hashmap", "no_std"] +categories = ["data-structures", "no-std"] +license = "Apache-2.0/MIT" +repository = "https://github.com/bluss/indexmap" +[package.metadata.docs.rs] +features = ["serde-1", "rayon"] + +[package.metadata.release] +no-dev-version = true +tag-name = "{{version}}" +[profile.bench] +debug = true + +[lib] +bench = false +[dependencies.hashbrown] +version = "0.9.0" +features = ["raw"] +default-features = false + +[dependencies.rayon] +version = "1.0" +optional = true + +[dependencies.serde] +version = "1.0" +optional = true +default-features = false +[dev-dependencies.fnv] +version = "1.0" + +[dev-dependencies.fxhash] +version = "0.2.1" + +[dev-dependencies.itertools] +version = "0.9" + +[dev-dependencies.lazy_static] +version = "1.3" + +[dev-dependencies.quickcheck] +version = "0.9" +default-features = false + +[dev-dependencies.rand] +version = "0.7" +features = ["small_rng"] +[build-dependencies.autocfg] +version = "1" + +[features] +serde-1 = ["serde"] +std = [] +test_debug = [] +test_low_transition_point = [] diff --git a/third_party/rust/indexmap/LICENSE-APACHE b/third_party/rust/indexmap/LICENSE-APACHE new file mode 100644 index 0000000000..16fe87b06e --- /dev/null +++ b/third_party/rust/indexmap/LICENSE-APACHE @@ -0,0 +1,201 @@ + Apache License + Version 2.0, January 2004 + http://www.apache.org/licenses/ + +TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION + +1. Definitions. + + "License" shall mean the terms and conditions for use, reproduction, + and distribution as defined by Sections 1 through 9 of this document. + + "Licensor" shall mean the copyright owner or entity authorized by + the copyright owner that is granting the License. + + "Legal Entity" shall mean the union of the acting entity and all + other entities that control, are controlled by, or are under common + control with that entity. For the purposes of this definition, + "control" means (i) the power, direct or indirect, to cause the + direction or management of such entity, whether by contract or + otherwise, or (ii) ownership of fifty percent (50%) or more of the + outstanding shares, or (iii) beneficial ownership of such entity. + + "You" (or "Your") shall mean an individual or Legal Entity + exercising permissions granted by this License. + + "Source" form shall mean the preferred form for making modifications, + including but not limited to software source code, documentation + source, and configuration files. + + "Object" form shall mean any form resulting from mechanical + transformation or translation of a Source form, including but + not limited to compiled object code, generated documentation, + and conversions to other media types. + + "Work" shall mean the work of authorship, whether in Source or + Object form, made available under the License, as indicated by a + copyright notice that is included in or attached to the work + (an example is provided in the Appendix below). + + "Derivative Works" shall mean any work, whether in Source or Object + form, that is based on (or derived from) the Work and for which the + editorial revisions, annotations, elaborations, or other modifications + represent, as a whole, an original work of authorship. For the purposes + of this License, Derivative Works shall not include works that remain + separable from, or merely link (or bind by name) to the interfaces of, + the Work and Derivative Works thereof. + + "Contribution" shall mean any work of authorship, including + the original version of the Work and any modifications or additions + to that Work or Derivative Works thereof, that is intentionally + submitted to Licensor for inclusion in the Work by the copyright owner + or by an individual or Legal Entity authorized to submit on behalf of + the copyright owner. For the purposes of this definition, "submitted" + means any form of electronic, verbal, or written communication sent + to the Licensor or its representatives, including but not limited to + communication on electronic mailing lists, source code control systems, + and issue tracking systems that are managed by, or on behalf of, the + Licensor for the purpose of discussing and improving the Work, but + excluding communication that is conspicuously marked or otherwise + designated in writing by the copyright owner as "Not a Contribution." + + "Contributor" shall mean Licensor and any individual or Legal Entity + on behalf of whom a Contribution has been received by Licensor and + subsequently incorporated within the Work. + +2. Grant of Copyright License. Subject to the terms and conditions of + this License, each Contributor hereby grants to You a perpetual, + worldwide, non-exclusive, no-charge, royalty-free, irrevocable + copyright license to reproduce, prepare Derivative Works of, + publicly display, publicly perform, sublicense, and distribute the + Work and such Derivative Works in Source or Object form. + +3. Grant of Patent License. Subject to the terms and conditions of + this License, each Contributor hereby grants to You a perpetual, + worldwide, non-exclusive, no-charge, royalty-free, irrevocable + (except as stated in this section) patent license to make, have made, + use, offer to sell, sell, import, and otherwise transfer the Work, + where such license applies only to those patent claims licensable + by such Contributor that are necessarily infringed by their + Contribution(s) alone or by combination of their Contribution(s) + with the Work to which such Contribution(s) was submitted. If You + institute patent litigation against any entity (including a + cross-claim or counterclaim in a lawsuit) alleging that the Work + or a Contribution incorporated within the Work constitutes direct + or contributory patent infringement, then any patent licenses + granted to You under this License for that Work shall terminate + as of the date such litigation is filed. + +4. Redistribution. You may reproduce and distribute copies of the + Work or Derivative Works thereof in any medium, with or without + modifications, and in Source or Object form, provided that You + meet the following conditions: + + (a) You must give any other recipients of the Work or + Derivative Works a copy of this License; and + + (b) You must cause any modified files to carry prominent notices + stating that You changed the files; and + + (c) You must retain, in the Source form of any Derivative Works + that You distribute, all copyright, patent, trademark, and + attribution notices from the Source form of the Work, + excluding those notices that do not pertain to any part of + the Derivative Works; and + + (d) If the Work includes a "NOTICE" text file as part of its + distribution, then any Derivative Works that You distribute must + include a readable copy of the attribution notices contained + within such NOTICE file, excluding those notices that do not + pertain to any part of the Derivative Works, in at least one + of the following places: within a NOTICE text file distributed + as part of the Derivative Works; within the Source form or + documentation, if provided along with the Derivative Works; or, + within a display generated by the Derivative Works, if and + wherever such third-party notices normally appear. The contents + of the NOTICE file are for informational purposes only and + do not modify the License. You may add Your own attribution + notices within Derivative Works that You distribute, alongside + or as an addendum to the NOTICE text from the Work, provided + that such additional attribution notices cannot be construed + as modifying the License. + + You may add Your own copyright statement to Your modifications and + may provide additional or different license terms and conditions + for use, reproduction, or distribution of Your modifications, or + for any such Derivative Works as a whole, provided Your use, + reproduction, and distribution of the Work otherwise complies with + the conditions stated in this License. + +5. Submission of Contributions. Unless You explicitly state otherwise, + any Contribution intentionally submitted for inclusion in the Work + by You to the Licensor shall be under the terms and conditions of + this License, without any additional terms or conditions. + Notwithstanding the above, nothing herein shall supersede or modify + the terms of any separate license agreement you may have executed + with Licensor regarding such Contributions. + +6. Trademarks. This License does not grant permission to use the trade + names, trademarks, service marks, or product names of the Licensor, + except as required for reasonable and customary use in describing the + origin of the Work and reproducing the content of the NOTICE file. + +7. Disclaimer of Warranty. Unless required by applicable law or + agreed to in writing, Licensor provides the Work (and each + Contributor provides its Contributions) on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or + implied, including, without limitation, any warranties or conditions + of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A + PARTICULAR PURPOSE. You are solely responsible for determining the + appropriateness of using or redistributing the Work and assume any + risks associated with Your exercise of permissions under this License. + +8. Limitation of Liability. In no event and under no legal theory, + whether in tort (including negligence), contract, or otherwise, + unless required by applicable law (such as deliberate and grossly + negligent acts) or agreed to in writing, shall any Contributor be + liable to You for damages, including any direct, indirect, special, + incidental, or consequential damages of any character arising as a + result of this License or out of the use or inability to use the + Work (including but not limited to damages for loss of goodwill, + work stoppage, computer failure or malfunction, or any and all + other commercial damages or losses), even if such Contributor + has been advised of the possibility of such damages. + +9. Accepting Warranty or Additional Liability. While redistributing + the Work or Derivative Works thereof, You may choose to offer, + and charge a fee for, acceptance of support, warranty, indemnity, + or other liability obligations and/or rights consistent with this + License. However, in accepting such obligations, You may act only + on Your own behalf and on Your sole responsibility, not on behalf + of any other Contributor, and only if You agree to indemnify, + defend, and hold each Contributor harmless for any liability + incurred by, or claims asserted against, such Contributor by reason + of your accepting any such warranty or additional liability. + +END OF TERMS AND CONDITIONS + +APPENDIX: How to apply the Apache License to your work. + + To apply the Apache License to your work, attach the following + boilerplate notice, with the fields enclosed by brackets "[]" + replaced with your own identifying information. (Don't include + the brackets!) The text should be enclosed in the appropriate + comment syntax for the file format. We also recommend that a + file or class name and description of purpose be included on the + same "printed page" as the copyright notice for easier + identification within third-party archives. + +Copyright [yyyy] [name of copyright owner] + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. diff --git a/third_party/rust/indexmap/LICENSE-MIT b/third_party/rust/indexmap/LICENSE-MIT new file mode 100644 index 0000000000..8b8181068b --- /dev/null +++ b/third_party/rust/indexmap/LICENSE-MIT @@ -0,0 +1,25 @@ +Copyright (c) 2016--2017 + +Permission is hereby granted, free of charge, to any +person obtaining a copy of this software and associated +documentation files (the "Software"), to deal in the +Software without restriction, including without +limitation the rights to use, copy, modify, merge, +publish, distribute, sublicense, and/or sell copies of +the Software, and to permit persons to whom the Software +is furnished to do so, subject to the following +conditions: + +The above copyright notice and this permission notice +shall be included in all copies or substantial portions +of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF +ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED +TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A +PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT +SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY +CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION +OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR +IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +DEALINGS IN THE SOFTWARE. diff --git a/third_party/rust/indexmap/README.rst b/third_party/rust/indexmap/README.rst new file mode 100644 index 0000000000..2c489850df --- /dev/null +++ b/third_party/rust/indexmap/README.rst @@ -0,0 +1,356 @@ +indexmap +======== + +|build_status|_ |crates|_ |docs|_ |rustc|_ + +.. |crates| image:: https://img.shields.io/crates/v/indexmap.svg +.. _crates: https://crates.io/crates/indexmap + +.. |build_status| image:: https://travis-ci.org/bluss/indexmap.svg +.. _build_status: https://travis-ci.org/bluss/indexmap + +.. |docs| image:: https://docs.rs/indexmap/badge.svg +.. _docs: https://docs.rs/indexmap + +.. |rustc| image:: https://img.shields.io/badge/rust-1.32%2B-orange.svg +.. _rustc: https://img.shields.io/badge/rust-1.32%2B-orange.svg + +A pure-Rust hash table which preserves (in a limited sense) insertion order. + +This crate implements compact map and set data-structures, +where the iteration order of the keys is independent from their hash or +value. It preserves insertion order (except after removals), and it +allows lookup of entries by either hash table key or numerical index. + +Note: this crate was originally released under the name ``ordermap``, +but it was renamed to ``indexmap`` to better reflect its features. + +Background +========== + +This was inspired by Python 3.6's new dict implementation (which remembers +the insertion order and is fast to iterate, and is compact in memory). + +Some of those features were translated to Rust, and some were not. The result +was indexmap, a hash table that has following properties: + +- Order is **independent of hash function** and hash values of keys. +- Fast to iterate. +- Indexed in compact space. +- Preserves insertion order **as long** as you don't call ``.remove()``. +- Uses hashbrown for the inner table, just like Rust's libstd ``HashMap`` does. + +Performance +----------- + +``IndexMap`` derives a couple of performance facts directly from how it is constructed, +which is roughly: + + A raw hash table of key-value indices, and a vector of key-value pairs. + +- Iteration is very fast since it is on the dense key-values. +- Removal is fast since it moves memory areas only in the table, + and uses a single swap in the vector. +- Lookup is fast-ish because the initial 7-bit hash lookup uses SIMD, and indices are + densely stored. Lookup also is slow-ish since the actual key-value pairs are stored + separately. (Visible when cpu caches size is limiting.) + +- In practice, ``IndexMap`` has been tested out as the hashmap in rustc in PR45282_ and + the performance was roughly on par across the whole workload. +- If you want the properties of ``IndexMap``, or its strongest performance points + fits your workload, it might be the best hash table implementation. + +.. _PR45282: https://github.com/rust-lang/rust/pull/45282 + + +Recent Changes +============== + +- 1.6.0 + + - **MSRV**: Rust 1.36 or later is now required. + + - The ``hashbrown`` dependency has been updated to version 0.9. + +- 1.5.2 + + - The new "std" feature will force the use of ``std`` for users that explicitly + want the default ``S = RandomState``, bypassing the autodetection added in 1.3.0, + by @cuviper in PR 145_. + +.. _145: https://github.com/bluss/indexmap/pull/145 + +- 1.5.1 + + - Values can now be indexed by their ``usize`` position by @cuviper in PR 132_. + + - Some of the generic bounds have been relaxed to match ``std`` by @cuviper in PR 141_. + + - ``drain`` now accepts any ``R: RangeBounds<usize>`` by @cuviper in PR 142_. + +.. _132: https://github.com/bluss/indexmap/pull/132 +.. _141: https://github.com/bluss/indexmap/pull/141 +.. _142: https://github.com/bluss/indexmap/pull/142 + +- 1.5.0 + + - **MSRV**: Rust 1.32 or later is now required. + + - The inner hash table is now based on ``hashbrown`` by @cuviper in PR 131_. + This also completes the method ``reserve`` and adds ``shrink_to_fit``. + + - Add new methods ``get_key_value``, ``remove_entry``, ``swap_remove_entry``, + and ``shift_remove_entry``, by @cuviper in PR 136_ + + - ``Clone::clone_from`` reuses allocations by @cuviper in PR 125_ + + - Add new method ``reverse`` by @linclelinkpart5 in PR 128_ + +.. _125: https://github.com/bluss/indexmap/pull/125 +.. _128: https://github.com/bluss/indexmap/pull/128 +.. _131: https://github.com/bluss/indexmap/pull/131 +.. _136: https://github.com/bluss/indexmap/pull/136 + +- 1.4.0 + + - Add new method ``get_index_of`` by @Thermatrix in PR 115_ and 120_ + + - Fix build script rebuild-if-changed configuration to use "build.rs"; + fixes issue 123_. Fix by @cuviper. + + - Dev-dependencies (rand and quickcheck) have been updated. The crate's tests + now run using Rust 1.32 or later (MSRV for building the crate has not changed). + by @kjeremy and @bluss + +.. _123: https://github.com/bluss/indexmap/issues/123 +.. _115: https://github.com/bluss/indexmap/pull/115 +.. _120: https://github.com/bluss/indexmap/pull/120 + +- 1.3.2 + + - Maintenance update to regenerate the published `Cargo.toml`. + +- 1.3.1 + + - Maintenance update for formatting and ``autocfg`` 1.0. + +- 1.3.0 + + - The deprecation messages in the previous version have been removed. + (The methods have not otherwise changed.) Docs for removal methods have been + improved. + - From Rust 1.36, this crate supports being built **without std**, requiring + ``alloc`` instead. This is enabled automatically when it is detected that + ``std`` is not available. There is no crate feature to enable/disable to + trigger this. The new build-dep ``autocfg`` enables this. + +- 1.2.0 + + - Plain ``.remove()`` now has a deprecation message, it informs the user + about picking one of the removal functions ``swap_remove`` and ``shift_remove`` + which have different performance and order semantics. + Plain ``.remove()`` will not be removed, the warning message and method + will remain until further. + + - Add new method ``shift_remove`` for order preserving removal on the map, + and ``shift_take`` for the corresponding operation on the set. + + - Add methods ``swap_remove``, ``swap_remove_entry`` to ``Entry``. + + - Fix indexset/indexmap to support full paths, like ``indexmap::indexmap!()`` + + - Internal improvements: fix warnings, deprecations and style lints + +- 1.1.0 + + - Added optional feature `"rayon"` that adds parallel iterator support + to `IndexMap` and `IndexSet` using Rayon. This includes all the regular + iterators in parallel versions, and parallel sort. + + - Implemented ``Clone`` for ``map::{Iter, Keys, Values}`` and + ``set::{Difference, Intersection, Iter, SymmetricDifference, Union}`` + + - Implemented ``Debug`` for ``map::{Entry, IntoIter, Iter, Keys, Values}`` and + ``set::{Difference, Intersection, IntoIter, Iter, SymmetricDifference, Union}`` + + - Serde trait ``IntoDeserializer`` are implemented for ``IndexMap`` and ``IndexSet``. + + - Minimum Rust version requirement increased to Rust 1.30 for development builds. + +- 1.0.2 + + - The new methods ``IndexMap::insert_full`` and ``IndexSet::insert_full`` are + both like ``insert`` with the index included in the return value. + + - The new method ``Entry::and_modify`` can be used to modify occupied + entries, matching the new methods of ``std`` maps in Rust 1.26. + + - The new method ``Entry::or_default`` inserts a default value in unoccupied + entries, matching the new methods of ``std`` maps in Rust 1.28. + +- 1.0.1 + + - Document Rust version policy for the crate (see rustdoc) + +- 1.0.0 + + - This is the 1.0 release for ``indexmap``! (the crate and datastructure + formerly known as “ordermap”) + - ``OccupiedEntry::insert`` changed its signature, to use ``&mut self`` for + the method receiver, matching the equivalent method for a standard + ``HashMap``. Thanks to @dtolnay for finding this bug. + - The deprecated old names from ordermap were removed: ``OrderMap``, + ``OrderSet``, ``ordermap!{}``, ``orderset!{}``. Use the new ``IndexMap`` + etc names instead. + +- 0.4.1 + + - Renamed crate to ``indexmap``; the ``ordermap`` crate is now deprecated + and the types ``OrderMap/Set`` now have a deprecation notice. + +- 0.4.0 + + - This is the last release series for this ``ordermap`` under that name, + because the crate is **going to be renamed** to ``indexmap`` (with types + ``IndexMap``, ``IndexSet``) and no change in functionality! + - The map and its associated structs moved into the ``map`` submodule of the + crate, so that the map and set are symmetric + + + The iterators, ``Entry`` and other structs are now under ``ordermap::map::`` + + - Internally refactored ``OrderMap<K, V, S>`` so that all the main algorithms + (insertion, lookup, removal etc) that don't use the ``S`` parameter (the + hasher) are compiled without depending on ``S``, which reduces generics bloat. + + - ``Entry<K, V>`` no longer has a type parameter ``S``, which is just like + the standard ``HashMap``'s entry. + + - Minimum Rust version requirement increased to Rust 1.18 + +- 0.3.5 + + - Documentation improvements + +- 0.3.4 + + - The ``.retain()`` methods for ``OrderMap`` and ``OrderSet`` now + traverse the elements in order, and the retained elements **keep their order** + - Added new methods ``.sort_by()``, ``.sort_keys()`` to ``OrderMap`` and + ``.sort_by()``, ``.sort()`` to ``OrderSet``. These methods allow you to + sort the maps in place efficiently. + +- 0.3.3 + + - Document insertion behaviour better by @lucab + - Updated dependences (no feature changes) by @ignatenkobrain + +- 0.3.2 + + - Add ``OrderSet`` by @cuviper! + - ``OrderMap::drain`` is now (too) a double ended iterator. + +- 0.3.1 + + - In all ordermap iterators, forward the ``collect`` method to the underlying + iterator as well. + - Add crates.io categories. + +- 0.3.0 + + - The methods ``get_pair``, ``get_pair_index`` were both replaced by + ``get_full`` (and the same for the mutable case). + - Method ``swap_remove_pair`` replaced by ``swap_remove_full``. + - Add trait ``MutableKeys`` for opt-in mutable key access. Mutable key access + is only possible through the methods of this extension trait. + - Add new trait ``Equivalent`` for key equivalence. This extends the + ``Borrow`` trait mechanism for ``OrderMap::get`` in a backwards compatible + way, just some minor type inference related issues may become apparent. + See `#10`__ for more information. + - Implement ``Extend<(&K, &V)>`` by @xfix. + +__ https://github.com/bluss/ordermap/pull/10 + +- 0.2.13 + + - Fix deserialization to support custom hashers by @Techcable. + - Add methods ``.index()`` on the entry types by @garro95. + +- 0.2.12 + + - Add methods ``.with_hasher()``, ``.hasher()``. + +- 0.2.11 + + - Support ``ExactSizeIterator`` for the iterators. By @Binero. + - Use ``Box<[Pos]>`` internally, saving a word in the ``OrderMap`` struct. + - Serde support, with crate feature ``"serde-1"``. By @xfix. + +- 0.2.10 + + - Add iterator ``.drain(..)`` by @stevej. + +- 0.2.9 + + - Add method ``.is_empty()`` by @overvenus. + - Implement ``PartialEq, Eq`` by @overvenus. + - Add method ``.sorted_by()``. + +- 0.2.8 + + - Add iterators ``.values()`` and ``.values_mut()``. + - Fix compatibility with 32-bit platforms. + +- 0.2.7 + + - Add ``.retain()``. + +- 0.2.6 + + - Add ``OccupiedEntry::remove_entry`` and other minor entry methods, + so that it now has all the features of ``HashMap``'s entries. + +- 0.2.5 + + - Improved ``.pop()`` slightly. + +- 0.2.4 + + - Improved performance of ``.insert()`` (`#3`__) by @pczarn. + +__ https://github.com/bluss/ordermap/pull/3 + +- 0.2.3 + + - Generalize ``Entry`` for now, so that it works on hashmaps with non-default + hasher. However, there's a lingering compat issue since libstd ``HashMap`` + does not parameterize its entries by the hasher (``S`` typarm). + - Special case some iterator methods like ``.nth()``. + +- 0.2.2 + + - Disable the verbose ``Debug`` impl by default. + +- 0.2.1 + + - Fix doc links and clarify docs. + +- 0.2.0 + + - Add more ``HashMap`` methods & compat with its API. + - Experimental support for ``.entry()`` (the simplest parts of the API). + - Add ``.reserve()`` (placeholder impl). + - Add ``.remove()`` as synonym for ``.swap_remove()``. + - Changed ``.insert()`` to swap value if the entry already exists, and + return ``Option``. + - Experimental support as an *indexed* hash map! Added methods + ``.get_index()``, ``.get_index_mut()``, ``.swap_remove_index()``, + ``.get_pair_index()``, ``.get_pair_index_mut()``. + +- 0.1.2 + + - Implement the 32/32 split idea for ``Pos`` which improves cache utilization + and lookup performance. + +- 0.1.1 + + - Initial release. diff --git a/third_party/rust/indexmap/benches/bench.rs b/third_party/rust/indexmap/benches/bench.rs new file mode 100644 index 0000000000..102cd49c8c --- /dev/null +++ b/third_party/rust/indexmap/benches/bench.rs @@ -0,0 +1,764 @@ +#![feature(test)] + +extern crate test; +#[macro_use] +extern crate lazy_static; + +use fnv::FnvHasher; +use std::hash::BuildHasherDefault; +use std::hash::Hash; +type FnvBuilder = BuildHasherDefault<FnvHasher>; + +use test::black_box; +use test::Bencher; + +use indexmap::IndexMap; + +use std::collections::HashMap; +use std::iter::FromIterator; + +use rand::rngs::SmallRng; +use rand::seq::SliceRandom; +use rand::SeedableRng; + +/// Use a consistently seeded Rng for benchmark stability +fn small_rng() -> SmallRng { + let seed = u64::from_le_bytes(*b"indexmap"); + SmallRng::seed_from_u64(seed) +} + +#[bench] +fn new_hashmap(b: &mut Bencher) { + b.iter(|| HashMap::<String, String>::new()); +} + +#[bench] +fn new_indexmap(b: &mut Bencher) { + b.iter(|| IndexMap::<String, String>::new()); +} + +#[bench] +fn with_capacity_10e5_hashmap(b: &mut Bencher) { + b.iter(|| HashMap::<String, String>::with_capacity(10_000)); +} + +#[bench] +fn with_capacity_10e5_indexmap(b: &mut Bencher) { + b.iter(|| IndexMap::<String, String>::with_capacity(10_000)); +} + +#[bench] +fn insert_hashmap_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.insert(x, ()); + } + map + }); +} + +#[bench] +fn insert_indexmap_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for x in 0..c { + map.insert(x, ()); + } + map + }); +} + +#[bench] +fn insert_hashmap_string_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.insert(x.to_string(), ()); + } + map + }); +} + +#[bench] +fn insert_indexmap_string_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for x in 0..c { + map.insert(x.to_string(), ()); + } + map + }); +} + +#[bench] +fn insert_hashmap_str_10_000(b: &mut Bencher) { + let c = 10_000; + let ss = Vec::from_iter((0..c).map(|x| x.to_string())); + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for key in &ss { + map.insert(&key[..], ()); + } + map + }); +} + +#[bench] +fn insert_indexmap_str_10_000(b: &mut Bencher) { + let c = 10_000; + let ss = Vec::from_iter((0..c).map(|x| x.to_string())); + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for key in &ss { + map.insert(&key[..], ()); + } + map + }); +} + +#[bench] +fn insert_hashmap_int_bigvalue_10_000(b: &mut Bencher) { + let c = 10_000; + let value = [0u64; 10]; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for i in 0..c { + map.insert(i, value); + } + map + }); +} + +#[bench] +fn insert_indexmap_int_bigvalue_10_000(b: &mut Bencher) { + let c = 10_000; + let value = [0u64; 10]; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for i in 0..c { + map.insert(i, value); + } + map + }); +} + +#[bench] +fn insert_hashmap_100_000(b: &mut Bencher) { + let c = 100_000; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.insert(x, ()); + } + map + }); +} + +#[bench] +fn insert_indexmap_100_000(b: &mut Bencher) { + let c = 100_000; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for x in 0..c { + map.insert(x, ()); + } + map + }); +} + +#[bench] +fn insert_hashmap_150(b: &mut Bencher) { + let c = 150; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.insert(x, ()); + } + map + }); +} + +#[bench] +fn insert_indexmap_150(b: &mut Bencher) { + let c = 150; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for x in 0..c { + map.insert(x, ()); + } + map + }); +} + +#[bench] +fn entry_hashmap_150(b: &mut Bencher) { + let c = 150; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.entry(x).or_insert(()); + } + map + }); +} + +#[bench] +fn entry_indexmap_150(b: &mut Bencher) { + let c = 150; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for x in 0..c { + map.entry(x).or_insert(()); + } + map + }); +} + +#[bench] +fn iter_sum_hashmap_10_000(b: &mut Bencher) { + let c = 10_000; + let mut map = HashMap::with_capacity(c); + let len = c - c / 10; + for x in 0..len { + map.insert(x, ()); + } + assert_eq!(map.len(), len); + b.iter(|| map.keys().sum::<usize>()); +} + +#[bench] +fn iter_sum_indexmap_10_000(b: &mut Bencher) { + let c = 10_000; + let mut map = IndexMap::with_capacity(c); + let len = c - c / 10; + for x in 0..len { + map.insert(x, ()); + } + assert_eq!(map.len(), len); + b.iter(|| map.keys().sum::<usize>()); +} + +#[bench] +fn iter_black_box_hashmap_10_000(b: &mut Bencher) { + let c = 10_000; + let mut map = HashMap::with_capacity(c); + let len = c - c / 10; + for x in 0..len { + map.insert(x, ()); + } + assert_eq!(map.len(), len); + b.iter(|| { + for &key in map.keys() { + black_box(key); + } + }); +} + +#[bench] +fn iter_black_box_indexmap_10_000(b: &mut Bencher) { + let c = 10_000; + let mut map = IndexMap::with_capacity(c); + let len = c - c / 10; + for x in 0..len { + map.insert(x, ()); + } + assert_eq!(map.len(), len); + b.iter(|| { + for &key in map.keys() { + black_box(key); + } + }); +} + +fn shuffled_keys<I>(iter: I) -> Vec<I::Item> +where + I: IntoIterator, +{ + let mut v = Vec::from_iter(iter); + let mut rng = small_rng(); + v.shuffle(&mut rng); + v +} + +#[bench] +fn lookup_hashmap_10_000_exist(b: &mut Bencher) { + let c = 10_000; + let mut map = HashMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(key, 1); + } + b.iter(|| { + let mut found = 0; + for key in 5000..c { + found += map.get(&key).is_some() as i32; + } + found + }); +} + +#[bench] +fn lookup_hashmap_10_000_noexist(b: &mut Bencher) { + let c = 10_000; + let mut map = HashMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(key, 1); + } + b.iter(|| { + let mut found = 0; + for key in c..15000 { + found += map.get(&key).is_some() as i32; + } + found + }); +} + +#[bench] +fn lookup_indexmap_10_000_exist(b: &mut Bencher) { + let c = 10_000; + let mut map = IndexMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(key, 1); + } + b.iter(|| { + let mut found = 0; + for key in 5000..c { + found += map.get(&key).is_some() as i32; + } + found + }); +} + +#[bench] +fn lookup_indexmap_10_000_noexist(b: &mut Bencher) { + let c = 10_000; + let mut map = IndexMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(key, 1); + } + b.iter(|| { + let mut found = 0; + for key in c..15000 { + found += map.get(&key).is_some() as i32; + } + found + }); +} + +// number of items to look up +const LOOKUP_MAP_SIZE: u32 = 100_000_u32; +const LOOKUP_SAMPLE_SIZE: u32 = 5000; +const SORT_MAP_SIZE: usize = 10_000; + +// use lazy_static so that comparison benchmarks use the exact same inputs +lazy_static! { + static ref KEYS: Vec<u32> = shuffled_keys(0..LOOKUP_MAP_SIZE); +} + +lazy_static! { + static ref HMAP_100K: HashMap<u32, u32> = { + let c = LOOKUP_MAP_SIZE; + let mut map = HashMap::with_capacity(c as usize); + let keys = &*KEYS; + for &key in keys { + map.insert(key, key); + } + map + }; +} + +lazy_static! { + static ref IMAP_100K: IndexMap<u32, u32> = { + let c = LOOKUP_MAP_SIZE; + let mut map = IndexMap::with_capacity(c as usize); + let keys = &*KEYS; + for &key in keys { + map.insert(key, key); + } + map + }; +} + +lazy_static! { + static ref IMAP_SORT_U32: IndexMap<u32, u32> = { + let mut map = IndexMap::with_capacity(SORT_MAP_SIZE); + for &key in &KEYS[..SORT_MAP_SIZE] { + map.insert(key, key); + } + map + }; +} +lazy_static! { + static ref IMAP_SORT_S: IndexMap<String, String> = { + let mut map = IndexMap::with_capacity(SORT_MAP_SIZE); + for &key in &KEYS[..SORT_MAP_SIZE] { + map.insert(format!("{:^16x}", &key), String::new()); + } + map + }; +} + +#[bench] +fn lookup_hashmap_100_000_multi(b: &mut Bencher) { + let map = &*HMAP_100K; + b.iter(|| { + let mut found = 0; + for key in 0..LOOKUP_SAMPLE_SIZE { + found += map.get(&key).is_some() as u32; + } + found + }); +} + +#[bench] +fn lookup_indexmap_100_000_multi(b: &mut Bencher) { + let map = &*IMAP_100K; + b.iter(|| { + let mut found = 0; + for key in 0..LOOKUP_SAMPLE_SIZE { + found += map.get(&key).is_some() as u32; + } + found + }); +} + +// inorder: Test looking up keys in the same order as they were inserted +#[bench] +fn lookup_hashmap_100_000_inorder_multi(b: &mut Bencher) { + let map = &*HMAP_100K; + let keys = &*KEYS; + b.iter(|| { + let mut found = 0; + for key in &keys[0..LOOKUP_SAMPLE_SIZE as usize] { + found += map.get(key).is_some() as u32; + } + found + }); +} + +#[bench] +fn lookup_indexmap_100_000_inorder_multi(b: &mut Bencher) { + let map = &*IMAP_100K; + let keys = &*KEYS; + b.iter(|| { + let mut found = 0; + for key in &keys[0..LOOKUP_SAMPLE_SIZE as usize] { + found += map.get(key).is_some() as u32; + } + found + }); +} + +#[bench] +fn lookup_hashmap_100_000_single(b: &mut Bencher) { + let map = &*HMAP_100K; + let mut iter = (0..LOOKUP_MAP_SIZE + LOOKUP_SAMPLE_SIZE).cycle(); + b.iter(|| { + let key = iter.next().unwrap(); + map.get(&key).is_some() + }); +} + +#[bench] +fn lookup_indexmap_100_000_single(b: &mut Bencher) { + let map = &*IMAP_100K; + let mut iter = (0..LOOKUP_MAP_SIZE + LOOKUP_SAMPLE_SIZE).cycle(); + b.iter(|| { + let key = iter.next().unwrap(); + map.get(&key).is_some() + }); +} + +const GROW_SIZE: usize = 100_000; +type GrowKey = u32; + +// Test grow/resize without preallocation +#[bench] +fn grow_fnv_hashmap_100_000(b: &mut Bencher) { + b.iter(|| { + let mut map: HashMap<_, _, FnvBuilder> = HashMap::default(); + for x in 0..GROW_SIZE { + map.insert(x as GrowKey, x as GrowKey); + } + map + }); +} + +#[bench] +fn grow_fnv_indexmap_100_000(b: &mut Bencher) { + b.iter(|| { + let mut map: IndexMap<_, _, FnvBuilder> = IndexMap::default(); + for x in 0..GROW_SIZE { + map.insert(x as GrowKey, x as GrowKey); + } + map + }); +} + +const MERGE: u64 = 10_000; +#[bench] +fn hashmap_merge_simple(b: &mut Bencher) { + let first_map: HashMap<u64, _> = (0..MERGE).map(|i| (i, ())).collect(); + let second_map: HashMap<u64, _> = (MERGE..MERGE * 2).map(|i| (i, ())).collect(); + b.iter(|| { + let mut merged = first_map.clone(); + merged.extend(second_map.iter().map(|(&k, &v)| (k, v))); + merged + }); +} + +#[bench] +fn hashmap_merge_shuffle(b: &mut Bencher) { + let first_map: HashMap<u64, _> = (0..MERGE).map(|i| (i, ())).collect(); + let second_map: HashMap<u64, _> = (MERGE..MERGE * 2).map(|i| (i, ())).collect(); + let mut v = Vec::new(); + let mut rng = small_rng(); + b.iter(|| { + let mut merged = first_map.clone(); + v.extend(second_map.iter().map(|(&k, &v)| (k, v))); + v.shuffle(&mut rng); + merged.extend(v.drain(..)); + + merged + }); +} + +#[bench] +fn indexmap_merge_simple(b: &mut Bencher) { + let first_map: IndexMap<u64, _> = (0..MERGE).map(|i| (i, ())).collect(); + let second_map: IndexMap<u64, _> = (MERGE..MERGE * 2).map(|i| (i, ())).collect(); + b.iter(|| { + let mut merged = first_map.clone(); + merged.extend(second_map.iter().map(|(&k, &v)| (k, v))); + merged + }); +} + +#[bench] +fn indexmap_merge_shuffle(b: &mut Bencher) { + let first_map: IndexMap<u64, _> = (0..MERGE).map(|i| (i, ())).collect(); + let second_map: IndexMap<u64, _> = (MERGE..MERGE * 2).map(|i| (i, ())).collect(); + let mut v = Vec::new(); + let mut rng = small_rng(); + b.iter(|| { + let mut merged = first_map.clone(); + v.extend(second_map.iter().map(|(&k, &v)| (k, v))); + v.shuffle(&mut rng); + merged.extend(v.drain(..)); + + merged + }); +} + +#[bench] +fn swap_remove_indexmap_100_000(b: &mut Bencher) { + let map = IMAP_100K.clone(); + let mut keys = Vec::from_iter(map.keys().cloned()); + let mut rng = small_rng(); + keys.shuffle(&mut rng); + + b.iter(|| { + let mut map = map.clone(); + for key in &keys { + map.swap_remove(key); + } + assert_eq!(map.len(), 0); + map + }); +} + +#[bench] +fn shift_remove_indexmap_100_000_few(b: &mut Bencher) { + let map = IMAP_100K.clone(); + let mut keys = Vec::from_iter(map.keys().cloned()); + let mut rng = small_rng(); + keys.shuffle(&mut rng); + keys.truncate(50); + + b.iter(|| { + let mut map = map.clone(); + for key in &keys { + map.shift_remove(key); + } + assert_eq!(map.len(), IMAP_100K.len() - keys.len()); + map + }); +} + +#[bench] +fn shift_remove_indexmap_2_000_full(b: &mut Bencher) { + let mut keys = KEYS[..2_000].to_vec(); + let mut map = IndexMap::with_capacity(keys.len()); + for &key in &keys { + map.insert(key, key); + } + let mut rng = small_rng(); + keys.shuffle(&mut rng); + + b.iter(|| { + let mut map = map.clone(); + for key in &keys { + map.shift_remove(key); + } + assert_eq!(map.len(), 0); + map + }); +} + +#[bench] +fn pop_indexmap_100_000(b: &mut Bencher) { + let map = IMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + while !map.is_empty() { + map.pop(); + } + assert_eq!(map.len(), 0); + map + }); +} + +#[bench] +fn few_retain_indexmap_100_000(b: &mut Bencher) { + let map = IMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + map.retain(|k, _| *k % 7 == 0); + map + }); +} + +#[bench] +fn few_retain_hashmap_100_000(b: &mut Bencher) { + let map = HMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + map.retain(|k, _| *k % 7 == 0); + map + }); +} + +#[bench] +fn half_retain_indexmap_100_000(b: &mut Bencher) { + let map = IMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + map.retain(|k, _| *k % 2 == 0); + map + }); +} + +#[bench] +fn half_retain_hashmap_100_000(b: &mut Bencher) { + let map = HMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + map.retain(|k, _| *k % 2 == 0); + map + }); +} + +#[bench] +fn many_retain_indexmap_100_000(b: &mut Bencher) { + let map = IMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + map.retain(|k, _| *k % 100 != 0); + map + }); +} + +#[bench] +fn many_retain_hashmap_100_000(b: &mut Bencher) { + let map = HMAP_100K.clone(); + + b.iter(|| { + let mut map = map.clone(); + map.retain(|k, _| *k % 100 != 0); + map + }); +} + +// simple sort impl for comparison +pub fn simple_sort<K: Ord + Hash, V>(m: &mut IndexMap<K, V>) { + let mut ordered: Vec<_> = m.drain(..).collect(); + ordered.sort_by(|left, right| left.0.cmp(&right.0)); + m.extend(ordered); +} + +#[bench] +fn indexmap_sort_s(b: &mut Bencher) { + let map = IMAP_SORT_S.clone(); + + // there's a map clone there, but it's still useful to profile this + b.iter(|| { + let mut map = map.clone(); + map.sort_keys(); + map + }); +} + +#[bench] +fn indexmap_simple_sort_s(b: &mut Bencher) { + let map = IMAP_SORT_S.clone(); + + // there's a map clone there, but it's still useful to profile this + b.iter(|| { + let mut map = map.clone(); + simple_sort(&mut map); + map + }); +} + +#[bench] +fn indexmap_sort_u32(b: &mut Bencher) { + let map = IMAP_SORT_U32.clone(); + + // there's a map clone there, but it's still useful to profile this + b.iter(|| { + let mut map = map.clone(); + map.sort_keys(); + map + }); +} + +#[bench] +fn indexmap_simple_sort_u32(b: &mut Bencher) { + let map = IMAP_SORT_U32.clone(); + + // there's a map clone there, but it's still useful to profile this + b.iter(|| { + let mut map = map.clone(); + simple_sort(&mut map); + map + }); +} + +// measure the fixed overhead of cloning in sort benchmarks +#[bench] +fn indexmap_clone_for_sort_s(b: &mut Bencher) { + let map = IMAP_SORT_S.clone(); + + b.iter(|| map.clone()); +} + +#[bench] +fn indexmap_clone_for_sort_u32(b: &mut Bencher) { + let map = IMAP_SORT_U32.clone(); + + b.iter(|| map.clone()); +} diff --git a/third_party/rust/indexmap/benches/faststring.rs b/third_party/rust/indexmap/benches/faststring.rs new file mode 100644 index 0000000000..86b7e9cf71 --- /dev/null +++ b/third_party/rust/indexmap/benches/faststring.rs @@ -0,0 +1,186 @@ +#![feature(test)] + +extern crate test; + +use test::Bencher; + +use indexmap::IndexMap; + +use std::collections::HashMap; +use std::iter::FromIterator; + +use rand::rngs::SmallRng; +use rand::seq::SliceRandom; +use rand::SeedableRng; + +use std::hash::{Hash, Hasher}; + +use std::borrow::Borrow; +use std::ops::Deref; + +/// Use a consistently seeded Rng for benchmark stability +fn small_rng() -> SmallRng { + let seed = u64::from_le_bytes(*b"indexmap"); + SmallRng::seed_from_u64(seed) +} + +#[derive(PartialEq, Eq, Copy, Clone)] +#[repr(transparent)] +pub struct OneShot<T: ?Sized>(pub T); + +impl Hash for OneShot<str> { + fn hash<H: Hasher>(&self, h: &mut H) { + h.write(self.0.as_bytes()) + } +} + +impl<'a, S> From<&'a S> for &'a OneShot<str> +where + S: AsRef<str>, +{ + fn from(s: &'a S) -> Self { + let s: &str = s.as_ref(); + unsafe { &*(s as *const str as *const OneShot<str>) } + } +} + +impl Hash for OneShot<String> { + fn hash<H: Hasher>(&self, h: &mut H) { + h.write(self.0.as_bytes()) + } +} + +impl Borrow<OneShot<str>> for OneShot<String> { + fn borrow(&self) -> &OneShot<str> { + <&OneShot<str>>::from(&self.0) + } +} + +impl<T> Deref for OneShot<T> { + type Target = T; + fn deref(&self) -> &T { + &self.0 + } +} + +fn shuffled_keys<I>(iter: I) -> Vec<I::Item> +where + I: IntoIterator, +{ + let mut v = Vec::from_iter(iter); + let mut rng = small_rng(); + v.shuffle(&mut rng); + v +} + +#[bench] +fn insert_hashmap_string_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.insert(x.to_string(), ()); + } + map + }); +} + +#[bench] +fn insert_hashmap_string_oneshot_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = HashMap::with_capacity(c); + for x in 0..c { + map.insert(OneShot(x.to_string()), ()); + } + map + }); +} + +#[bench] +fn insert_indexmap_string_10_000(b: &mut Bencher) { + let c = 10_000; + b.iter(|| { + let mut map = IndexMap::with_capacity(c); + for x in 0..c { + map.insert(x.to_string(), ()); + } + map + }); +} + +#[bench] +fn lookup_hashmap_10_000_exist_string(b: &mut Bencher) { + let c = 10_000; + let mut map = HashMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(key.to_string(), 1); + } + let lookups = (5000..c).map(|x| x.to_string()).collect::<Vec<_>>(); + b.iter(|| { + let mut found = 0; + for key in &lookups { + found += map.get(key).is_some() as i32; + } + found + }); +} + +#[bench] +fn lookup_hashmap_10_000_exist_string_oneshot(b: &mut Bencher) { + let c = 10_000; + let mut map = HashMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(OneShot(key.to_string()), 1); + } + let lookups = (5000..c) + .map(|x| OneShot(x.to_string())) + .collect::<Vec<_>>(); + b.iter(|| { + let mut found = 0; + for key in &lookups { + found += map.get(key).is_some() as i32; + } + found + }); +} + +#[bench] +fn lookup_indexmap_10_000_exist_string(b: &mut Bencher) { + let c = 10_000; + let mut map = IndexMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(key.to_string(), 1); + } + let lookups = (5000..c).map(|x| x.to_string()).collect::<Vec<_>>(); + b.iter(|| { + let mut found = 0; + for key in &lookups { + found += map.get(key).is_some() as i32; + } + found + }); +} + +#[bench] +fn lookup_indexmap_10_000_exist_string_oneshot(b: &mut Bencher) { + let c = 10_000; + let mut map = IndexMap::with_capacity(c); + let keys = shuffled_keys(0..c); + for &key in &keys { + map.insert(OneShot(key.to_string()), 1); + } + let lookups = (5000..c) + .map(|x| OneShot(x.to_string())) + .collect::<Vec<_>>(); + b.iter(|| { + let mut found = 0; + for key in &lookups { + found += map.get(key).is_some() as i32; + } + found + }); +} diff --git a/third_party/rust/indexmap/build.rs b/third_party/rust/indexmap/build.rs new file mode 100644 index 0000000000..9f9fa054f8 --- /dev/null +++ b/third_party/rust/indexmap/build.rs @@ -0,0 +1,8 @@ +fn main() { + // If "std" is explicitly requested, don't bother probing the target for it. + match std::env::var_os("CARGO_FEATURE_STD") { + Some(_) => autocfg::emit("has_std"), + None => autocfg::new().emit_sysroot_crate("std"), + } + autocfg::rerun_path("build.rs"); +} diff --git a/third_party/rust/indexmap/src/equivalent.rs b/third_party/rust/indexmap/src/equivalent.rs new file mode 100644 index 0000000000..ad6635ffac --- /dev/null +++ b/third_party/rust/indexmap/src/equivalent.rs @@ -0,0 +1,27 @@ +use core::borrow::Borrow; + +/// Key equivalence trait. +/// +/// This trait allows hash table lookup to be customized. +/// It has one blanket implementation that uses the regular `Borrow` solution, +/// just like `HashMap` and `BTreeMap` do, so that you can pass `&str` to lookup +/// into a map with `String` keys and so on. +/// +/// # Contract +/// +/// The implementor **must** hash like `K`, if it is hashable. +pub trait Equivalent<K: ?Sized> { + /// Compare self to `key` and return `true` if they are equal. + fn equivalent(&self, key: &K) -> bool; +} + +impl<Q: ?Sized, K: ?Sized> Equivalent<K> for Q +where + Q: Eq, + K: Borrow<Q>, +{ + #[inline] + fn equivalent(&self, key: &K) -> bool { + *self == *key.borrow() + } +} diff --git a/third_party/rust/indexmap/src/lib.rs b/third_party/rust/indexmap/src/lib.rs new file mode 100644 index 0000000000..2df2ceed8a --- /dev/null +++ b/third_party/rust/indexmap/src/lib.rs @@ -0,0 +1,191 @@ +// We *mostly* avoid unsafe code, but `map::core::raw` allows it to use `RawTable` buckets. +#![deny(unsafe_code)] +#![warn(rust_2018_idioms)] +#![doc(html_root_url = "https://docs.rs/indexmap/1/")] +#![no_std] + +//! [`IndexMap`] is a hash table where the iteration order of the key-value +//! pairs is independent of the hash values of the keys. +//! +//! [`IndexSet`] is a corresponding hash set using the same implementation and +//! with similar properties. +//! +//! [`IndexMap`]: map/struct.IndexMap.html +//! [`IndexSet`]: set/struct.IndexSet.html +//! +//! +//! ### Feature Highlights +//! +//! [`IndexMap`] and [`IndexSet`] are drop-in compatible with the std `HashMap` +//! and `HashSet`, but they also have some features of note: +//! +//! - The ordering semantics (see their documentation for details) +//! - Sorting methods and the [`.pop()`][IndexMap::pop] methods. +//! - The [`Equivalent`] trait, which offers more flexible equality definitions +//! between borrowed and owned versions of keys. +//! - The [`MutableKeys`][map::MutableKeys] trait, which gives opt-in mutable +//! access to hash map keys. +//! +//! ### Alternate Hashers +//! +//! [`IndexMap`] and [`IndexSet`] have a default hasher type `S = RandomState`, +//! just like the standard `HashMap` and `HashSet`, which is resistant to +//! HashDoS attacks but not the most performant. Type aliases can make it easier +//! to use alternate hashers: +//! +//! ``` +//! use fnv::FnvBuildHasher; +//! use fxhash::FxBuildHasher; +//! use indexmap::{IndexMap, IndexSet}; +//! +//! type FnvIndexMap<K, V> = IndexMap<K, V, FnvBuildHasher>; +//! type FnvIndexSet<T> = IndexSet<T, FnvBuildHasher>; +//! +//! type FxIndexMap<K, V> = IndexMap<K, V, FxBuildHasher>; +//! type FxIndexSet<T> = IndexSet<T, FxBuildHasher>; +//! +//! let std: IndexSet<i32> = (0..100).collect(); +//! let fnv: FnvIndexSet<i32> = (0..100).collect(); +//! let fx: FxIndexSet<i32> = (0..100).collect(); +//! assert_eq!(std, fnv); +//! assert_eq!(std, fx); +//! ``` +//! +//! ### Rust Version +//! +//! This version of indexmap requires Rust 1.32 or later, or Rust 1.36+ for +//! using with `alloc` (without `std`), see below. +//! +//! The indexmap 1.x release series will use a carefully considered version +//! upgrade policy, where in a later 1.x version, we will raise the minimum +//! required Rust version. +//! +//! ## No Standard Library Targets +//! +//! From Rust 1.36, this crate supports being built without `std`, requiring +//! `alloc` instead. This is enabled automatically when it is detected that +//! `std` is not available. There is no crate feature to enable/disable to +//! trigger this. It can be tested by building for a std-less target. +//! +//! - Creating maps and sets using [`new`][IndexMap::new] and +//! [`with_capacity`][IndexMap::with_capacity] is unavailable without `std`. +//! Use methods [`IndexMap::default`][def], +//! [`with_hasher`][IndexMap::with_hasher], +//! [`with_capacity_and_hasher`][IndexMap::with_capacity_and_hasher] instead. +//! A no-std compatible hasher will be needed as well, for example +//! from the crate `twox-hash`. +//! - Macros [`indexmap!`] and [`indexset!`] are unavailable without `std`. +//! +//! [def]: map/struct.IndexMap.html#impl-Default + +#[cfg(not(has_std))] +extern crate alloc; + +#[cfg(has_std)] +#[macro_use] +extern crate std; + +#[cfg(not(has_std))] +use alloc::vec::{self, Vec}; + +#[cfg(has_std)] +use std::vec::{self, Vec}; + +#[macro_use] +mod macros; +mod equivalent; +mod mutable_keys; +#[cfg(feature = "serde-1")] +mod serde; +mod util; + +pub mod map; +pub mod set; + +// Placed after `map` and `set` so new `rayon` methods on the types +// are documented after the "normal" methods. +#[cfg(feature = "rayon")] +mod rayon; + +pub use crate::equivalent::Equivalent; +pub use crate::map::IndexMap; +pub use crate::set::IndexSet; + +// shared private items + +/// Hash value newtype. Not larger than usize, since anything larger +/// isn't used for selecting position anyway. +#[derive(Clone, Copy, Debug, PartialEq)] +struct HashValue(usize); + +impl HashValue { + #[inline(always)] + fn get(self) -> u64 { + self.0 as u64 + } +} + +#[derive(Copy, Debug)] +struct Bucket<K, V> { + hash: HashValue, + key: K, + value: V, +} + +impl<K, V> Clone for Bucket<K, V> +where + K: Clone, + V: Clone, +{ + fn clone(&self) -> Self { + Bucket { + hash: self.hash, + key: self.key.clone(), + value: self.value.clone(), + } + } + + fn clone_from(&mut self, other: &Self) { + self.hash = other.hash; + self.key.clone_from(&other.key); + self.value.clone_from(&other.value); + } +} + +impl<K, V> Bucket<K, V> { + // field accessors -- used for `f` instead of closures in `.map(f)` + fn key_ref(&self) -> &K { + &self.key + } + fn value_ref(&self) -> &V { + &self.value + } + fn value_mut(&mut self) -> &mut V { + &mut self.value + } + fn key(self) -> K { + self.key + } + fn key_value(self) -> (K, V) { + (self.key, self.value) + } + fn refs(&self) -> (&K, &V) { + (&self.key, &self.value) + } + fn ref_mut(&mut self) -> (&K, &mut V) { + (&self.key, &mut self.value) + } + fn muts(&mut self) -> (&mut K, &mut V) { + (&mut self.key, &mut self.value) + } +} + +trait Entries { + type Entry; + fn into_entries(self) -> Vec<Self::Entry>; + fn as_entries(&self) -> &[Self::Entry]; + fn as_entries_mut(&mut self) -> &mut [Self::Entry]; + fn with_entries<F>(&mut self, f: F) + where + F: FnOnce(&mut [Self::Entry]); +} diff --git a/third_party/rust/indexmap/src/macros.rs b/third_party/rust/indexmap/src/macros.rs new file mode 100644 index 0000000000..c4d84217d1 --- /dev/null +++ b/third_party/rust/indexmap/src/macros.rs @@ -0,0 +1,120 @@ +#[cfg(has_std)] +#[macro_export] +/// Create an `IndexMap` from a list of key-value pairs +/// +/// ## Example +/// +/// ``` +/// use indexmap::indexmap; +/// +/// let map = indexmap!{ +/// "a" => 1, +/// "b" => 2, +/// }; +/// assert_eq!(map["a"], 1); +/// assert_eq!(map["b"], 2); +/// assert_eq!(map.get("c"), None); +/// +/// // "a" is the first key +/// assert_eq!(map.keys().next(), Some(&"a")); +/// ``` +macro_rules! indexmap { + (@single $($x:tt)*) => (()); + (@count $($rest:expr),*) => (<[()]>::len(&[$($crate::indexmap!(@single $rest)),*])); + + ($($key:expr => $value:expr,)+) => { $crate::indexmap!($($key => $value),+) }; + ($($key:expr => $value:expr),*) => { + { + let _cap = $crate::indexmap!(@count $($key),*); + let mut _map = $crate::IndexMap::with_capacity(_cap); + $( + _map.insert($key, $value); + )* + _map + } + }; +} + +#[cfg(has_std)] +#[macro_export] +/// Create an `IndexSet` from a list of values +/// +/// ## Example +/// +/// ``` +/// use indexmap::indexset; +/// +/// let set = indexset!{ +/// "a", +/// "b", +/// }; +/// assert!(set.contains("a")); +/// assert!(set.contains("b")); +/// assert!(!set.contains("c")); +/// +/// // "a" is the first value +/// assert_eq!(set.iter().next(), Some(&"a")); +/// ``` +macro_rules! indexset { + (@single $($x:tt)*) => (()); + (@count $($rest:expr),*) => (<[()]>::len(&[$($crate::indexset!(@single $rest)),*])); + + ($($value:expr,)+) => { $crate::indexset!($($value),+) }; + ($($value:expr),*) => { + { + let _cap = $crate::indexset!(@count $($value),*); + let mut _set = $crate::IndexSet::with_capacity(_cap); + $( + _set.insert($value); + )* + _set + } + }; +} + +// generate all the Iterator methods by just forwarding to the underlying +// self.iter and mapping its element. +macro_rules! iterator_methods { + // $map_elt is the mapping function from the underlying iterator's element + // same mapping function for both options and iterators + ($map_elt:expr) => { + fn next(&mut self) -> Option<Self::Item> { + self.iter.next().map($map_elt) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } + + fn count(self) -> usize { + self.iter.len() + } + + fn nth(&mut self, n: usize) -> Option<Self::Item> { + self.iter.nth(n).map($map_elt) + } + + fn last(mut self) -> Option<Self::Item> { + self.next_back() + } + + fn collect<C>(self) -> C + where + C: FromIterator<Self::Item>, + { + // NB: forwarding this directly to standard iterators will + // allow it to leverage unstable traits like `TrustedLen`. + self.iter.map($map_elt).collect() + } + }; +} + +macro_rules! double_ended_iterator_methods { + // $map_elt is the mapping function from the underlying iterator's element + // same mapping function for both options and iterators + ($map_elt:expr) => { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map($map_elt) + } + }; +} diff --git a/third_party/rust/indexmap/src/map.rs b/third_party/rust/indexmap/src/map.rs new file mode 100644 index 0000000000..6829ef94b8 --- /dev/null +++ b/third_party/rust/indexmap/src/map.rs @@ -0,0 +1,1627 @@ +//! `IndexMap` is a hash table where the iteration order of the key-value +//! pairs is independent of the hash values of the keys. + +mod core; + +pub use crate::mutable_keys::MutableKeys; + +#[cfg(feature = "rayon")] +pub use crate::rayon::map as rayon; + +use crate::vec::{self, Vec}; +use ::core::cmp::Ordering; +use ::core::fmt; +use ::core::hash::{BuildHasher, Hash, Hasher}; +use ::core::iter::FromIterator; +use ::core::ops::{Index, IndexMut, RangeBounds}; +use ::core::slice::{Iter as SliceIter, IterMut as SliceIterMut}; + +#[cfg(has_std)] +use std::collections::hash_map::RandomState; + +use self::core::IndexMapCore; +use crate::equivalent::Equivalent; +use crate::util::third; +use crate::{Bucket, Entries, HashValue}; + +pub use self::core::{Entry, OccupiedEntry, VacantEntry}; + +/// A hash table where the iteration order of the key-value pairs is independent +/// of the hash values of the keys. +/// +/// The interface is closely compatible with the standard `HashMap`, but also +/// has additional features. +/// +/// # Order +/// +/// The key-value pairs have a consistent order that is determined by +/// the sequence of insertion and removal calls on the map. The order does +/// not depend on the keys or the hash function at all. +/// +/// All iterators traverse the map in *the order*. +/// +/// The insertion order is preserved, with **notable exceptions** like the +/// `.remove()` or `.swap_remove()` methods. Methods such as `.sort_by()` of +/// course result in a new order, depending on the sorting order. +/// +/// # Indices +/// +/// The key-value pairs are indexed in a compact range without holes in the +/// range `0..self.len()`. For example, the method `.get_full` looks up the +/// index for a key, and the method `.get_index` looks up the key-value pair by +/// index. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexMap; +/// +/// // count the frequency of each letter in a sentence. +/// let mut letters = IndexMap::new(); +/// for ch in "a short treatise on fungi".chars() { +/// *letters.entry(ch).or_insert(0) += 1; +/// } +/// +/// assert_eq!(letters[&'s'], 2); +/// assert_eq!(letters[&'t'], 3); +/// assert_eq!(letters[&'u'], 1); +/// assert_eq!(letters.get(&'y'), None); +/// ``` +#[cfg(has_std)] +pub struct IndexMap<K, V, S = RandomState> { + core: IndexMapCore<K, V>, + hash_builder: S, +} +#[cfg(not(has_std))] +pub struct IndexMap<K, V, S> { + core: IndexMapCore<K, V>, + hash_builder: S, +} + +impl<K, V, S> Clone for IndexMap<K, V, S> +where + K: Clone, + V: Clone, + S: Clone, +{ + fn clone(&self) -> Self { + IndexMap { + core: self.core.clone(), + hash_builder: self.hash_builder.clone(), + } + } + + fn clone_from(&mut self, other: &Self) { + self.core.clone_from(&other.core); + self.hash_builder.clone_from(&other.hash_builder); + } +} + +impl<K, V, S> Entries for IndexMap<K, V, S> { + type Entry = Bucket<K, V>; + + #[inline] + fn into_entries(self) -> Vec<Self::Entry> { + self.core.into_entries() + } + + #[inline] + fn as_entries(&self) -> &[Self::Entry] { + self.core.as_entries() + } + + #[inline] + fn as_entries_mut(&mut self) -> &mut [Self::Entry] { + self.core.as_entries_mut() + } + + fn with_entries<F>(&mut self, f: F) + where + F: FnOnce(&mut [Self::Entry]), + { + self.core.with_entries(f); + } +} + +impl<K, V, S> fmt::Debug for IndexMap<K, V, S> +where + K: fmt::Debug, + V: fmt::Debug, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + if cfg!(not(feature = "test_debug")) { + f.debug_map().entries(self.iter()).finish() + } else { + // Let the inner `IndexMapCore` print all of its details + f.debug_struct("IndexMap") + .field("core", &self.core) + .finish() + } + } +} + +#[cfg(has_std)] +impl<K, V> IndexMap<K, V> { + /// Create a new map. (Does not allocate.) + #[inline] + pub fn new() -> Self { + Self::with_capacity(0) + } + + /// Create a new map with capacity for `n` key-value pairs. (Does not + /// allocate if `n` is zero.) + /// + /// Computes in **O(n)** time. + #[inline] + pub fn with_capacity(n: usize) -> Self { + Self::with_capacity_and_hasher(n, <_>::default()) + } +} + +impl<K, V, S> IndexMap<K, V, S> { + /// Create a new map with capacity for `n` key-value pairs. (Does not + /// allocate if `n` is zero.) + /// + /// Computes in **O(n)** time. + #[inline] + pub fn with_capacity_and_hasher(n: usize, hash_builder: S) -> Self { + if n == 0 { + IndexMap { + core: IndexMapCore::new(), + hash_builder, + } + } else { + IndexMap { + core: IndexMapCore::with_capacity(n), + hash_builder, + } + } + } + + /// Create a new map with `hash_builder` + pub fn with_hasher(hash_builder: S) -> Self { + Self::with_capacity_and_hasher(0, hash_builder) + } + + /// Computes in **O(1)** time. + pub fn capacity(&self) -> usize { + self.core.capacity() + } + + /// Return a reference to the map's `BuildHasher`. + pub fn hasher(&self) -> &S { + &self.hash_builder + } + + /// Return the number of key-value pairs in the map. + /// + /// Computes in **O(1)** time. + #[inline] + pub fn len(&self) -> usize { + self.core.len() + } + + /// Returns true if the map contains no elements. + /// + /// Computes in **O(1)** time. + #[inline] + pub fn is_empty(&self) -> bool { + self.len() == 0 + } + + /// Return an iterator over the key-value pairs of the map, in their order + pub fn iter(&self) -> Iter<'_, K, V> { + Iter { + iter: self.as_entries().iter(), + } + } + + /// Return an iterator over the key-value pairs of the map, in their order + pub fn iter_mut(&mut self) -> IterMut<'_, K, V> { + IterMut { + iter: self.as_entries_mut().iter_mut(), + } + } + + /// Return an iterator over the keys of the map, in their order + pub fn keys(&self) -> Keys<'_, K, V> { + Keys { + iter: self.as_entries().iter(), + } + } + + /// Return an iterator over the values of the map, in their order + pub fn values(&self) -> Values<'_, K, V> { + Values { + iter: self.as_entries().iter(), + } + } + + /// Return an iterator over mutable references to the the values of the map, + /// in their order + pub fn values_mut(&mut self) -> ValuesMut<'_, K, V> { + ValuesMut { + iter: self.as_entries_mut().iter_mut(), + } + } + + /// Remove all key-value pairs in the map, while preserving its capacity. + /// + /// Computes in **O(n)** time. + pub fn clear(&mut self) { + self.core.clear(); + } + + /// Clears the `IndexMap` in the given index range, returning those + /// key-value pairs as a drain iterator. + /// + /// The range may be any type that implements `RangeBounds<usize>`, + /// including all of the `std::ops::Range*` types, or even a tuple pair of + /// `Bound` start and end values. To drain the map entirely, use `RangeFull` + /// like `map.drain(..)`. + /// + /// This shifts down all entries following the drained range to fill the + /// gap, and keeps the allocated memory for reuse. + /// + /// ***Panics*** if the starting point is greater than the end point or if + /// the end point is greater than the length of the map. + pub fn drain<R>(&mut self, range: R) -> Drain<'_, K, V> + where + R: RangeBounds<usize>, + { + Drain { + iter: self.core.drain(range), + } + } +} + +impl<K, V, S> IndexMap<K, V, S> +where + K: Hash + Eq, + S: BuildHasher, +{ + /// Reserve capacity for `additional` more key-value pairs. + /// + /// Computes in **O(n)** time. + pub fn reserve(&mut self, additional: usize) { + self.core.reserve(additional); + } + + /// Shrink the capacity of the map as much as possible. + /// + /// Computes in **O(n)** time. + pub fn shrink_to_fit(&mut self) { + self.core.shrink_to_fit(); + } + + fn hash<Q: ?Sized + Hash>(&self, key: &Q) -> HashValue { + let mut h = self.hash_builder.build_hasher(); + key.hash(&mut h); + HashValue(h.finish() as usize) + } + + /// Insert a key-value pair in the map. + /// + /// If an equivalent key already exists in the map: the key remains and + /// retains in its place in the order, its corresponding value is updated + /// with `value` and the older value is returned inside `Some(_)`. + /// + /// If no equivalent key existed in the map: the new key-value pair is + /// inserted, last in order, and `None` is returned. + /// + /// Computes in **O(1)** time (amortized average). + /// + /// See also [`entry`](#method.entry) if you you want to insert *or* modify + /// or if you need to get the index of the corresponding key-value pair. + pub fn insert(&mut self, key: K, value: V) -> Option<V> { + self.insert_full(key, value).1 + } + + /// Insert a key-value pair in the map, and get their index. + /// + /// If an equivalent key already exists in the map: the key remains and + /// retains in its place in the order, its corresponding value is updated + /// with `value` and the older value is returned inside `(index, Some(_))`. + /// + /// If no equivalent key existed in the map: the new key-value pair is + /// inserted, last in order, and `(index, None)` is returned. + /// + /// Computes in **O(1)** time (amortized average). + /// + /// See also [`entry`](#method.entry) if you you want to insert *or* modify + /// or if you need to get the index of the corresponding key-value pair. + pub fn insert_full(&mut self, key: K, value: V) -> (usize, Option<V>) { + let hash = self.hash(&key); + self.core.insert_full(hash, key, value) + } + + /// Get the given key’s corresponding entry in the map for insertion and/or + /// in-place manipulation. + /// + /// Computes in **O(1)** time (amortized average). + pub fn entry(&mut self, key: K) -> Entry<'_, K, V> { + let hash = self.hash(&key); + self.core.entry(hash, key) + } + + /// Return `true` if an equivalent to `key` exists in the map. + /// + /// Computes in **O(1)** time (average). + pub fn contains_key<Q: ?Sized>(&self, key: &Q) -> bool + where + Q: Hash + Equivalent<K>, + { + self.get_index_of(key).is_some() + } + + /// Return a reference to the value stored for `key`, if it is present, + /// else `None`. + /// + /// Computes in **O(1)** time (average). + pub fn get<Q: ?Sized>(&self, key: &Q) -> Option<&V> + where + Q: Hash + Equivalent<K>, + { + if let Some(i) = self.get_index_of(key) { + let entry = &self.as_entries()[i]; + Some(&entry.value) + } else { + None + } + } + + /// Return references to the key-value pair stored for `key`, + /// if it is present, else `None`. + /// + /// Computes in **O(1)** time (average). + pub fn get_key_value<Q: ?Sized>(&self, key: &Q) -> Option<(&K, &V)> + where + Q: Hash + Equivalent<K>, + { + if let Some(i) = self.get_index_of(key) { + let entry = &self.as_entries()[i]; + Some((&entry.key, &entry.value)) + } else { + None + } + } + + /// Return item index, key and value + pub fn get_full<Q: ?Sized>(&self, key: &Q) -> Option<(usize, &K, &V)> + where + Q: Hash + Equivalent<K>, + { + if let Some(i) = self.get_index_of(key) { + let entry = &self.as_entries()[i]; + Some((i, &entry.key, &entry.value)) + } else { + None + } + } + + /// Return item index, if it exists in the map + pub fn get_index_of<Q: ?Sized>(&self, key: &Q) -> Option<usize> + where + Q: Hash + Equivalent<K>, + { + if self.is_empty() { + None + } else { + let hash = self.hash(key); + self.core.get_index_of(hash, key) + } + } + + pub fn get_mut<Q: ?Sized>(&mut self, key: &Q) -> Option<&mut V> + where + Q: Hash + Equivalent<K>, + { + if let Some(i) = self.get_index_of(key) { + let entry = &mut self.as_entries_mut()[i]; + Some(&mut entry.value) + } else { + None + } + } + + pub fn get_full_mut<Q: ?Sized>(&mut self, key: &Q) -> Option<(usize, &K, &mut V)> + where + Q: Hash + Equivalent<K>, + { + if let Some(i) = self.get_index_of(key) { + let entry = &mut self.as_entries_mut()[i]; + Some((i, &entry.key, &mut entry.value)) + } else { + None + } + } + + pub(crate) fn get_full_mut2_impl<Q: ?Sized>( + &mut self, + key: &Q, + ) -> Option<(usize, &mut K, &mut V)> + where + Q: Hash + Equivalent<K>, + { + if let Some(i) = self.get_index_of(key) { + let entry = &mut self.as_entries_mut()[i]; + Some((i, &mut entry.key, &mut entry.value)) + } else { + None + } + } + + /// Remove the key-value pair equivalent to `key` and return + /// its value. + /// + /// **NOTE:** This is equivalent to `.swap_remove(key)`, if you need to + /// preserve the order of the keys in the map, use `.shift_remove(key)` + /// instead. + /// + /// Computes in **O(1)** time (average). + pub fn remove<Q: ?Sized>(&mut self, key: &Q) -> Option<V> + where + Q: Hash + Equivalent<K>, + { + self.swap_remove(key) + } + + /// Remove and return the key-value pair equivalent to `key`. + /// + /// **NOTE:** This is equivalent to `.swap_remove_entry(key)`, if you need to + /// preserve the order of the keys in the map, use `.shift_remove_entry(key)` + /// instead. + /// + /// Computes in **O(1)** time (average). + pub fn remove_entry<Q: ?Sized>(&mut self, key: &Q) -> Option<(K, V)> + where + Q: Hash + Equivalent<K>, + { + self.swap_remove_entry(key) + } + + /// Remove the key-value pair equivalent to `key` and return + /// its value. + /// + /// Like `Vec::swap_remove`, the pair is removed by swapping it with the + /// last element of the map and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Return `None` if `key` is not in map. + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove<Q: ?Sized>(&mut self, key: &Q) -> Option<V> + where + Q: Hash + Equivalent<K>, + { + self.swap_remove_full(key).map(third) + } + + /// Remove and return the key-value pair equivalent to `key`. + /// + /// Like `Vec::swap_remove`, the pair is removed by swapping it with the + /// last element of the map and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Return `None` if `key` is not in map. + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove_entry<Q: ?Sized>(&mut self, key: &Q) -> Option<(K, V)> + where + Q: Hash + Equivalent<K>, + { + match self.swap_remove_full(key) { + Some((_, key, value)) => Some((key, value)), + None => None, + } + } + + /// Remove the key-value pair equivalent to `key` and return it and + /// the index it had. + /// + /// Like `Vec::swap_remove`, the pair is removed by swapping it with the + /// last element of the map and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Return `None` if `key` is not in map. + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove_full<Q: ?Sized>(&mut self, key: &Q) -> Option<(usize, K, V)> + where + Q: Hash + Equivalent<K>, + { + if self.is_empty() { + return None; + } + let hash = self.hash(key); + self.core.swap_remove_full(hash, key) + } + + /// Remove the key-value pair equivalent to `key` and return + /// its value. + /// + /// Like `Vec::remove`, the pair is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Return `None` if `key` is not in map. + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove<Q: ?Sized>(&mut self, key: &Q) -> Option<V> + where + Q: Hash + Equivalent<K>, + { + self.shift_remove_full(key).map(third) + } + + /// Remove and return the key-value pair equivalent to `key`. + /// + /// Like `Vec::remove`, the pair is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Return `None` if `key` is not in map. + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove_entry<Q: ?Sized>(&mut self, key: &Q) -> Option<(K, V)> + where + Q: Hash + Equivalent<K>, + { + match self.shift_remove_full(key) { + Some((_, key, value)) => Some((key, value)), + None => None, + } + } + + /// Remove the key-value pair equivalent to `key` and return it and + /// the index it had. + /// + /// Like `Vec::remove`, the pair is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Return `None` if `key` is not in map. + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove_full<Q: ?Sized>(&mut self, key: &Q) -> Option<(usize, K, V)> + where + Q: Hash + Equivalent<K>, + { + if self.is_empty() { + return None; + } + let hash = self.hash(key); + self.core.shift_remove_full(hash, key) + } + + /// Remove the last key-value pair + /// + /// Computes in **O(1)** time (average). + pub fn pop(&mut self) -> Option<(K, V)> { + self.core.pop() + } + + /// Scan through each key-value pair in the map and keep those where the + /// closure `keep` returns `true`. + /// + /// The elements are visited in order, and remaining elements keep their + /// order. + /// + /// Computes in **O(n)** time (average). + pub fn retain<F>(&mut self, mut keep: F) + where + F: FnMut(&K, &mut V) -> bool, + { + self.core.retain_in_order(move |k, v| keep(k, v)); + } + + pub(crate) fn retain_mut<F>(&mut self, keep: F) + where + F: FnMut(&mut K, &mut V) -> bool, + { + self.core.retain_in_order(keep); + } + + /// Sort the map’s key-value pairs by the default ordering of the keys. + /// + /// See `sort_by` for details. + pub fn sort_keys(&mut self) + where + K: Ord, + { + self.with_entries(|entries| { + entries.sort_by(|a, b| Ord::cmp(&a.key, &b.key)); + }); + } + + /// Sort the map’s key-value pairs in place using the comparison + /// function `compare`. + /// + /// The comparison function receives two key and value pairs to compare (you + /// can sort by keys or values or their combination as needed). + /// + /// Computes in **O(n log n + c)** time and **O(n)** space where *n* is + /// the length of the map and *c* the capacity. The sort is stable. + pub fn sort_by<F>(&mut self, mut cmp: F) + where + F: FnMut(&K, &V, &K, &V) -> Ordering, + { + self.with_entries(move |entries| { + entries.sort_by(move |a, b| cmp(&a.key, &a.value, &b.key, &b.value)); + }); + } + + /// Sort the key-value pairs of the map and return a by value iterator of + /// the key-value pairs with the result. + /// + /// The sort is stable. + pub fn sorted_by<F>(self, mut cmp: F) -> IntoIter<K, V> + where + F: FnMut(&K, &V, &K, &V) -> Ordering, + { + let mut entries = self.into_entries(); + entries.sort_by(move |a, b| cmp(&a.key, &a.value, &b.key, &b.value)); + IntoIter { + iter: entries.into_iter(), + } + } + + /// Reverses the order of the map’s key-value pairs in place. + /// + /// Computes in **O(n)** time and **O(1)** space. + pub fn reverse(&mut self) { + self.core.reverse() + } +} + +impl<K, V, S> IndexMap<K, V, S> { + /// Get a key-value pair by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Computes in **O(1)** time. + pub fn get_index(&self, index: usize) -> Option<(&K, &V)> { + self.as_entries().get(index).map(Bucket::refs) + } + + /// Get a key-value pair by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Computes in **O(1)** time. + pub fn get_index_mut(&mut self, index: usize) -> Option<(&mut K, &mut V)> { + self.as_entries_mut().get_mut(index).map(Bucket::muts) + } + + /// Remove the key-value pair by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Like `Vec::swap_remove`, the pair is removed by swapping it with the + /// last element of the map and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove_index(&mut self, index: usize) -> Option<(K, V)> { + self.core.swap_remove_index(index) + } + + /// Remove the key-value pair by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Like `Vec::remove`, the pair is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove_index(&mut self, index: usize) -> Option<(K, V)> { + self.core.shift_remove_index(index) + } +} + +/// An iterator over the keys of a `IndexMap`. +/// +/// This `struct` is created by the [`keys`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`keys`]: struct.IndexMap.html#method.keys +/// [`IndexMap`]: struct.IndexMap.html +pub struct Keys<'a, K, V> { + pub(crate) iter: SliceIter<'a, Bucket<K, V>>, +} + +impl<'a, K, V> Iterator for Keys<'a, K, V> { + type Item = &'a K; + + iterator_methods!(Bucket::key_ref); +} + +impl<K, V> DoubleEndedIterator for Keys<'_, K, V> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::key_ref) + } +} + +impl<K, V> ExactSizeIterator for Keys<'_, K, V> { + fn len(&self) -> usize { + self.iter.len() + } +} + +// FIXME(#26925) Remove in favor of `#[derive(Clone)]` +impl<K, V> Clone for Keys<'_, K, V> { + fn clone(&self) -> Self { + Keys { + iter: self.iter.clone(), + } + } +} + +impl<K: fmt::Debug, V> fmt::Debug for Keys<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// An iterator over the values of a `IndexMap`. +/// +/// This `struct` is created by the [`values`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`values`]: struct.IndexMap.html#method.values +/// [`IndexMap`]: struct.IndexMap.html +pub struct Values<'a, K, V> { + iter: SliceIter<'a, Bucket<K, V>>, +} + +impl<'a, K, V> Iterator for Values<'a, K, V> { + type Item = &'a V; + + iterator_methods!(Bucket::value_ref); +} + +impl<K, V> DoubleEndedIterator for Values<'_, K, V> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::value_ref) + } +} + +impl<K, V> ExactSizeIterator for Values<'_, K, V> { + fn len(&self) -> usize { + self.iter.len() + } +} + +// FIXME(#26925) Remove in favor of `#[derive(Clone)]` +impl<K, V> Clone for Values<'_, K, V> { + fn clone(&self) -> Self { + Values { + iter: self.iter.clone(), + } + } +} + +impl<K, V: fmt::Debug> fmt::Debug for Values<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A mutable iterator over the values of a `IndexMap`. +/// +/// This `struct` is created by the [`values_mut`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`values_mut`]: struct.IndexMap.html#method.values_mut +/// [`IndexMap`]: struct.IndexMap.html +pub struct ValuesMut<'a, K, V> { + iter: SliceIterMut<'a, Bucket<K, V>>, +} + +impl<'a, K, V> Iterator for ValuesMut<'a, K, V> { + type Item = &'a mut V; + + iterator_methods!(Bucket::value_mut); +} + +impl<K, V> DoubleEndedIterator for ValuesMut<'_, K, V> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::value_mut) + } +} + +impl<K, V> ExactSizeIterator for ValuesMut<'_, K, V> { + fn len(&self) -> usize { + self.iter.len() + } +} + +/// An iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`iter`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`iter`]: struct.IndexMap.html#method.iter +/// [`IndexMap`]: struct.IndexMap.html +pub struct Iter<'a, K, V> { + iter: SliceIter<'a, Bucket<K, V>>, +} + +impl<'a, K, V> Iterator for Iter<'a, K, V> { + type Item = (&'a K, &'a V); + + iterator_methods!(Bucket::refs); +} + +impl<K, V> DoubleEndedIterator for Iter<'_, K, V> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::refs) + } +} + +impl<K, V> ExactSizeIterator for Iter<'_, K, V> { + fn len(&self) -> usize { + self.iter.len() + } +} + +// FIXME(#26925) Remove in favor of `#[derive(Clone)]` +impl<K, V> Clone for Iter<'_, K, V> { + fn clone(&self) -> Self { + Iter { + iter: self.iter.clone(), + } + } +} + +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for Iter<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A mutable iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`iter_mut`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`iter_mut`]: struct.IndexMap.html#method.iter_mut +/// [`IndexMap`]: struct.IndexMap.html +pub struct IterMut<'a, K, V> { + iter: SliceIterMut<'a, Bucket<K, V>>, +} + +impl<'a, K, V> Iterator for IterMut<'a, K, V> { + type Item = (&'a K, &'a mut V); + + iterator_methods!(Bucket::ref_mut); +} + +impl<K, V> DoubleEndedIterator for IterMut<'_, K, V> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::ref_mut) + } +} + +impl<K, V> ExactSizeIterator for IterMut<'_, K, V> { + fn len(&self) -> usize { + self.iter.len() + } +} + +/// An owning iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`into_iter`] method on [`IndexMap`] +/// (provided by the `IntoIterator` trait). See its documentation for more. +/// +/// [`into_iter`]: struct.IndexMap.html#method.into_iter +/// [`IndexMap`]: struct.IndexMap.html +pub struct IntoIter<K, V> { + pub(crate) iter: vec::IntoIter<Bucket<K, V>>, +} + +impl<K, V> Iterator for IntoIter<K, V> { + type Item = (K, V); + + iterator_methods!(Bucket::key_value); +} + +impl<K, V> DoubleEndedIterator for IntoIter<K, V> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::key_value) + } +} + +impl<K, V> ExactSizeIterator for IntoIter<K, V> { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for IntoIter<K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.iter.as_slice().iter().map(Bucket::refs); + f.debug_list().entries(iter).finish() + } +} + +/// A draining iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`drain`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`drain`]: struct.IndexMap.html#method.drain +/// [`IndexMap`]: struct.IndexMap.html +pub struct Drain<'a, K, V> { + pub(crate) iter: vec::Drain<'a, Bucket<K, V>>, +} + +impl<K, V> Iterator for Drain<'_, K, V> { + type Item = (K, V); + + iterator_methods!(Bucket::key_value); +} + +impl<K, V> DoubleEndedIterator for Drain<'_, K, V> { + double_ended_iterator_methods!(Bucket::key_value); +} + +impl<'a, K, V, S> IntoIterator for &'a IndexMap<K, V, S> { + type Item = (&'a K, &'a V); + type IntoIter = Iter<'a, K, V>; + fn into_iter(self) -> Self::IntoIter { + self.iter() + } +} + +impl<'a, K, V, S> IntoIterator for &'a mut IndexMap<K, V, S> { + type Item = (&'a K, &'a mut V); + type IntoIter = IterMut<'a, K, V>; + fn into_iter(self) -> Self::IntoIter { + self.iter_mut() + } +} + +impl<K, V, S> IntoIterator for IndexMap<K, V, S> { + type Item = (K, V); + type IntoIter = IntoIter<K, V>; + fn into_iter(self) -> Self::IntoIter { + IntoIter { + iter: self.into_entries().into_iter(), + } + } +} + +/// Access `IndexMap` values corresponding to a key. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// for word in "Lorem ipsum dolor sit amet".split_whitespace() { +/// map.insert(word.to_lowercase(), word.to_uppercase()); +/// } +/// assert_eq!(map["lorem"], "LOREM"); +/// assert_eq!(map["ipsum"], "IPSUM"); +/// ``` +/// +/// ```should_panic +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// map.insert("foo", 1); +/// println!("{:?}", map["bar"]); // panics! +/// ``` +impl<K, V, Q: ?Sized, S> Index<&Q> for IndexMap<K, V, S> +where + Q: Hash + Equivalent<K>, + K: Hash + Eq, + S: BuildHasher, +{ + type Output = V; + + /// Returns a reference to the value corresponding to the supplied `key`. + /// + /// ***Panics*** if `key` is not present in the map. + fn index(&self, key: &Q) -> &V { + self.get(key).expect("IndexMap: key not found") + } +} + +/// Access `IndexMap` values corresponding to a key. +/// +/// Mutable indexing allows changing / updating values of key-value +/// pairs that are already present. +/// +/// You can **not** insert new pairs with index syntax, use `.insert()`. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// for word in "Lorem ipsum dolor sit amet".split_whitespace() { +/// map.insert(word.to_lowercase(), word.to_string()); +/// } +/// let lorem = &mut map["lorem"]; +/// assert_eq!(lorem, "Lorem"); +/// lorem.retain(char::is_lowercase); +/// assert_eq!(map["lorem"], "orem"); +/// ``` +/// +/// ```should_panic +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// map.insert("foo", 1); +/// map["bar"] = 1; // panics! +/// ``` +impl<K, V, Q: ?Sized, S> IndexMut<&Q> for IndexMap<K, V, S> +where + Q: Hash + Equivalent<K>, + K: Hash + Eq, + S: BuildHasher, +{ + /// Returns a mutable reference to the value corresponding to the supplied `key`. + /// + /// ***Panics*** if `key` is not present in the map. + fn index_mut(&mut self, key: &Q) -> &mut V { + self.get_mut(key).expect("IndexMap: key not found") + } +} + +/// Access `IndexMap` values at indexed positions. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// for word in "Lorem ipsum dolor sit amet".split_whitespace() { +/// map.insert(word.to_lowercase(), word.to_uppercase()); +/// } +/// assert_eq!(map[0], "LOREM"); +/// assert_eq!(map[1], "IPSUM"); +/// map.reverse(); +/// assert_eq!(map[0], "AMET"); +/// assert_eq!(map[1], "SIT"); +/// map.sort_keys(); +/// assert_eq!(map[0], "AMET"); +/// assert_eq!(map[1], "DOLOR"); +/// ``` +/// +/// ```should_panic +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// map.insert("foo", 1); +/// println!("{:?}", map[10]); // panics! +/// ``` +impl<K, V, S> Index<usize> for IndexMap<K, V, S> { + type Output = V; + + /// Returns a reference to the value at the supplied `index`. + /// + /// ***Panics*** if `index` is out of bounds. + fn index(&self, index: usize) -> &V { + self.get_index(index) + .expect("IndexMap: index out of bounds") + .1 + } +} + +/// Access `IndexMap` values at indexed positions. +/// +/// Mutable indexing allows changing / updating indexed values +/// that are already present. +/// +/// You can **not** insert new values with index syntax, use `.insert()`. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// for word in "Lorem ipsum dolor sit amet".split_whitespace() { +/// map.insert(word.to_lowercase(), word.to_string()); +/// } +/// let lorem = &mut map[0]; +/// assert_eq!(lorem, "Lorem"); +/// lorem.retain(char::is_lowercase); +/// assert_eq!(map["lorem"], "orem"); +/// ``` +/// +/// ```should_panic +/// use indexmap::IndexMap; +/// +/// let mut map = IndexMap::new(); +/// map.insert("foo", 1); +/// map[10] = 1; // panics! +/// ``` +impl<K, V, S> IndexMut<usize> for IndexMap<K, V, S> { + /// Returns a mutable reference to the value at the supplied `index`. + /// + /// ***Panics*** if `index` is out of bounds. + fn index_mut(&mut self, index: usize) -> &mut V { + self.get_index_mut(index) + .expect("IndexMap: index out of bounds") + .1 + } +} + +impl<K, V, S> FromIterator<(K, V)> for IndexMap<K, V, S> +where + K: Hash + Eq, + S: BuildHasher + Default, +{ + /// Create an `IndexMap` from the sequence of key-value pairs in the + /// iterable. + /// + /// `from_iter` uses the same logic as `extend`. See + /// [`extend`](#method.extend) for more details. + fn from_iter<I: IntoIterator<Item = (K, V)>>(iterable: I) -> Self { + let iter = iterable.into_iter(); + let (low, _) = iter.size_hint(); + let mut map = Self::with_capacity_and_hasher(low, <_>::default()); + map.extend(iter); + map + } +} + +impl<K, V, S> Extend<(K, V)> for IndexMap<K, V, S> +where + K: Hash + Eq, + S: BuildHasher, +{ + /// Extend the map with all key-value pairs in the iterable. + /// + /// This is equivalent to calling [`insert`](#method.insert) for each of + /// them in order, which means that for keys that already existed + /// in the map, their value is updated but it keeps the existing order. + /// + /// New keys are inserted in the order they appear in the sequence. If + /// equivalents of a key occur more than once, the last corresponding value + /// prevails. + fn extend<I: IntoIterator<Item = (K, V)>>(&mut self, iterable: I) { + // (Note: this is a copy of `std`/`hashbrown`'s reservation logic.) + // Keys may be already present or show multiple times in the iterator. + // Reserve the entire hint lower bound if the map is empty. + // Otherwise reserve half the hint (rounded up), so the map + // will only resize twice in the worst case. + let iter = iterable.into_iter(); + let reserve = if self.is_empty() { + iter.size_hint().0 + } else { + (iter.size_hint().0 + 1) / 2 + }; + self.reserve(reserve); + iter.for_each(move |(k, v)| { + self.insert(k, v); + }); + } +} + +impl<'a, K, V, S> Extend<(&'a K, &'a V)> for IndexMap<K, V, S> +where + K: Hash + Eq + Copy, + V: Copy, + S: BuildHasher, +{ + /// Extend the map with all key-value pairs in the iterable. + /// + /// See the first extend method for more details. + fn extend<I: IntoIterator<Item = (&'a K, &'a V)>>(&mut self, iterable: I) { + self.extend(iterable.into_iter().map(|(&key, &value)| (key, value))); + } +} + +impl<K, V, S> Default for IndexMap<K, V, S> +where + S: Default, +{ + /// Return an empty `IndexMap` + fn default() -> Self { + Self::with_capacity_and_hasher(0, S::default()) + } +} + +impl<K, V1, S1, V2, S2> PartialEq<IndexMap<K, V2, S2>> for IndexMap<K, V1, S1> +where + K: Hash + Eq, + V1: PartialEq<V2>, + S1: BuildHasher, + S2: BuildHasher, +{ + fn eq(&self, other: &IndexMap<K, V2, S2>) -> bool { + if self.len() != other.len() { + return false; + } + + self.iter() + .all(|(key, value)| other.get(key).map_or(false, |v| *value == *v)) + } +} + +impl<K, V, S> Eq for IndexMap<K, V, S> +where + K: Eq + Hash, + V: Eq, + S: BuildHasher, +{ +} + +#[cfg(test)] +mod tests { + use super::*; + use crate::util::enumerate; + use std::string::String; + + #[test] + fn it_works() { + let mut map = IndexMap::new(); + assert_eq!(map.is_empty(), true); + map.insert(1, ()); + map.insert(1, ()); + assert_eq!(map.len(), 1); + assert!(map.get(&1).is_some()); + assert_eq!(map.is_empty(), false); + } + + #[test] + fn new() { + let map = IndexMap::<String, String>::new(); + println!("{:?}", map); + assert_eq!(map.capacity(), 0); + assert_eq!(map.len(), 0); + assert_eq!(map.is_empty(), true); + } + + #[test] + fn insert() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5]; + let not_present = [1, 3, 6, 9, 10]; + let mut map = IndexMap::with_capacity(insert.len()); + + for (i, &elt) in enumerate(&insert) { + assert_eq!(map.len(), i); + map.insert(elt, elt); + assert_eq!(map.len(), i + 1); + assert_eq!(map.get(&elt), Some(&elt)); + assert_eq!(map[&elt], elt); + } + println!("{:?}", map); + + for &elt in ¬_present { + assert!(map.get(&elt).is_none()); + } + } + + #[test] + fn insert_full() { + let insert = vec![9, 2, 7, 1, 4, 6, 13]; + let present = vec![1, 6, 2]; + let mut map = IndexMap::with_capacity(insert.len()); + + for (i, &elt) in enumerate(&insert) { + assert_eq!(map.len(), i); + let (index, existing) = map.insert_full(elt, elt); + assert_eq!(existing, None); + assert_eq!(Some(index), map.get_full(&elt).map(|x| x.0)); + assert_eq!(map.len(), i + 1); + } + + let len = map.len(); + for &elt in &present { + let (index, existing) = map.insert_full(elt, elt); + assert_eq!(existing, Some(elt)); + assert_eq!(Some(index), map.get_full(&elt).map(|x| x.0)); + assert_eq!(map.len(), len); + } + } + + #[test] + fn insert_2() { + let mut map = IndexMap::with_capacity(16); + + let mut keys = vec![]; + keys.extend(0..16); + keys.extend(128..267); + + for &i in &keys { + let old_map = map.clone(); + map.insert(i, ()); + for key in old_map.keys() { + if map.get(key).is_none() { + println!("old_map: {:?}", old_map); + println!("map: {:?}", map); + panic!("did not find {} in map", key); + } + } + } + + for &i in &keys { + assert!(map.get(&i).is_some(), "did not find {}", i); + } + } + + #[test] + fn insert_order() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut map = IndexMap::new(); + + for &elt in &insert { + map.insert(elt, ()); + } + + assert_eq!(map.keys().count(), map.len()); + assert_eq!(map.keys().count(), insert.len()); + for (a, b) in insert.iter().zip(map.keys()) { + assert_eq!(a, b); + } + for (i, k) in (0..insert.len()).zip(map.keys()) { + assert_eq!(map.get_index(i).unwrap().0, k); + } + } + + #[test] + fn grow() { + let insert = [0, 4, 2, 12, 8, 7, 11]; + let not_present = [1, 3, 6, 9, 10]; + let mut map = IndexMap::with_capacity(insert.len()); + + for (i, &elt) in enumerate(&insert) { + assert_eq!(map.len(), i); + map.insert(elt, elt); + assert_eq!(map.len(), i + 1); + assert_eq!(map.get(&elt), Some(&elt)); + assert_eq!(map[&elt], elt); + } + + println!("{:?}", map); + for &elt in &insert { + map.insert(elt * 10, elt); + } + for &elt in &insert { + map.insert(elt * 100, elt); + } + for (i, &elt) in insert.iter().cycle().enumerate().take(100) { + map.insert(elt * 100 + i as i32, elt); + } + println!("{:?}", map); + for &elt in ¬_present { + assert!(map.get(&elt).is_none()); + } + } + + #[test] + fn reserve() { + let mut map = IndexMap::<usize, usize>::new(); + assert_eq!(map.capacity(), 0); + map.reserve(100); + let capacity = map.capacity(); + assert!(capacity >= 100); + for i in 0..capacity { + assert_eq!(map.len(), i); + map.insert(i, i * i); + assert_eq!(map.len(), i + 1); + assert_eq!(map.capacity(), capacity); + assert_eq!(map.get(&i), Some(&(i * i))); + } + map.insert(capacity, std::usize::MAX); + assert_eq!(map.len(), capacity + 1); + assert!(map.capacity() > capacity); + assert_eq!(map.get(&capacity), Some(&std::usize::MAX)); + } + + #[test] + fn shrink_to_fit() { + let mut map = IndexMap::<usize, usize>::new(); + assert_eq!(map.capacity(), 0); + for i in 0..100 { + assert_eq!(map.len(), i); + map.insert(i, i * i); + assert_eq!(map.len(), i + 1); + assert!(map.capacity() >= i + 1); + assert_eq!(map.get(&i), Some(&(i * i))); + map.shrink_to_fit(); + assert_eq!(map.len(), i + 1); + assert_eq!(map.capacity(), i + 1); + assert_eq!(map.get(&i), Some(&(i * i))); + } + } + + #[test] + fn remove() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut map = IndexMap::new(); + + for &elt in &insert { + map.insert(elt, elt); + } + + assert_eq!(map.keys().count(), map.len()); + assert_eq!(map.keys().count(), insert.len()); + for (a, b) in insert.iter().zip(map.keys()) { + assert_eq!(a, b); + } + + let remove_fail = [99, 77]; + let remove = [4, 12, 8, 7]; + + for &key in &remove_fail { + assert!(map.swap_remove_full(&key).is_none()); + } + println!("{:?}", map); + for &key in &remove { + //println!("{:?}", map); + let index = map.get_full(&key).unwrap().0; + assert_eq!(map.swap_remove_full(&key), Some((index, key, key))); + } + println!("{:?}", map); + + for key in &insert { + assert_eq!(map.get(key).is_some(), !remove.contains(key)); + } + assert_eq!(map.len(), insert.len() - remove.len()); + assert_eq!(map.keys().count(), insert.len() - remove.len()); + } + + #[test] + fn remove_to_empty() { + let mut map = indexmap! { 0 => 0, 4 => 4, 5 => 5 }; + map.swap_remove(&5).unwrap(); + map.swap_remove(&4).unwrap(); + map.swap_remove(&0).unwrap(); + assert!(map.is_empty()); + } + + #[test] + fn swap_remove_index() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut map = IndexMap::new(); + + for &elt in &insert { + map.insert(elt, elt * 2); + } + + let mut vector = insert.to_vec(); + let remove_sequence = &[3, 3, 10, 4, 5, 4, 3, 0, 1]; + + // check that the same swap remove sequence on vec and map + // have the same result. + for &rm in remove_sequence { + let out_vec = vector.swap_remove(rm); + let (out_map, _) = map.swap_remove_index(rm).unwrap(); + assert_eq!(out_vec, out_map); + } + assert_eq!(vector.len(), map.len()); + for (a, b) in vector.iter().zip(map.keys()) { + assert_eq!(a, b); + } + } + + #[test] + fn partial_eq_and_eq() { + let mut map_a = IndexMap::new(); + map_a.insert(1, "1"); + map_a.insert(2, "2"); + let mut map_b = map_a.clone(); + assert_eq!(map_a, map_b); + map_b.swap_remove(&1); + assert_ne!(map_a, map_b); + + let map_c: IndexMap<_, String> = map_b.into_iter().map(|(k, v)| (k, v.into())).collect(); + assert_ne!(map_a, map_c); + assert_ne!(map_c, map_a); + } + + #[test] + fn extend() { + let mut map = IndexMap::new(); + map.extend(vec![(&1, &2), (&3, &4)]); + map.extend(vec![(5, 6)]); + assert_eq!( + map.into_iter().collect::<Vec<_>>(), + vec![(1, 2), (3, 4), (5, 6)] + ); + } + + #[test] + fn entry() { + let mut map = IndexMap::new(); + + map.insert(1, "1"); + map.insert(2, "2"); + { + let e = map.entry(3); + assert_eq!(e.index(), 2); + let e = e.or_insert("3"); + assert_eq!(e, &"3"); + } + + let e = map.entry(2); + assert_eq!(e.index(), 1); + assert_eq!(e.key(), &2); + match e { + Entry::Occupied(ref e) => assert_eq!(e.get(), &"2"), + Entry::Vacant(_) => panic!(), + } + assert_eq!(e.or_insert("4"), &"2"); + } + + #[test] + fn entry_and_modify() { + let mut map = IndexMap::new(); + + map.insert(1, "1"); + map.entry(1).and_modify(|x| *x = "2"); + assert_eq!(Some(&"2"), map.get(&1)); + + map.entry(2).and_modify(|x| *x = "doesn't exist"); + assert_eq!(None, map.get(&2)); + } + + #[test] + fn entry_or_default() { + let mut map = IndexMap::new(); + + #[derive(Debug, PartialEq)] + enum TestEnum { + DefaultValue, + NonDefaultValue, + } + + impl Default for TestEnum { + fn default() -> Self { + TestEnum::DefaultValue + } + } + + map.insert(1, TestEnum::NonDefaultValue); + assert_eq!(&mut TestEnum::NonDefaultValue, map.entry(1).or_default()); + + assert_eq!(&mut TestEnum::DefaultValue, map.entry(2).or_default()); + } + + #[test] + fn keys() { + let vec = vec![(1, 'a'), (2, 'b'), (3, 'c')]; + let map: IndexMap<_, _> = vec.into_iter().collect(); + let keys: Vec<_> = map.keys().cloned().collect(); + assert_eq!(keys.len(), 3); + assert!(keys.contains(&1)); + assert!(keys.contains(&2)); + assert!(keys.contains(&3)); + } + + #[test] + fn values() { + let vec = vec![(1, 'a'), (2, 'b'), (3, 'c')]; + let map: IndexMap<_, _> = vec.into_iter().collect(); + let values: Vec<_> = map.values().cloned().collect(); + assert_eq!(values.len(), 3); + assert!(values.contains(&'a')); + assert!(values.contains(&'b')); + assert!(values.contains(&'c')); + } + + #[test] + fn values_mut() { + let vec = vec![(1, 1), (2, 2), (3, 3)]; + let mut map: IndexMap<_, _> = vec.into_iter().collect(); + for value in map.values_mut() { + *value *= 2 + } + let values: Vec<_> = map.values().cloned().collect(); + assert_eq!(values.len(), 3); + assert!(values.contains(&2)); + assert!(values.contains(&4)); + assert!(values.contains(&6)); + } +} diff --git a/third_party/rust/indexmap/src/map/core.rs b/third_party/rust/indexmap/src/map/core.rs new file mode 100644 index 0000000000..7adb69a133 --- /dev/null +++ b/third_party/rust/indexmap/src/map/core.rs @@ -0,0 +1,410 @@ +//! This is the core implementation that doesn't depend on the hasher at all. +//! +//! The methods of `IndexMapCore` don't use any Hash properties of K. +//! +//! It's cleaner to separate them out, then the compiler checks that we are not +//! using Hash at all in these methods. +//! +//! However, we should probably not let this show in the public API or docs. + +mod raw; + +use hashbrown::raw::RawTable; + +use crate::vec::{Drain, Vec}; +use core::cmp; +use core::fmt; +use core::mem::replace; +use core::ops::RangeBounds; + +use crate::equivalent::Equivalent; +use crate::util::{enumerate, simplify_range}; +use crate::{Bucket, Entries, HashValue}; + +/// Core of the map that does not depend on S +pub(crate) struct IndexMapCore<K, V> { + /// indices mapping from the entry hash to its index. + indices: RawTable<usize>, + /// entries is a dense vec of entries in their order. + entries: Vec<Bucket<K, V>>, +} + +#[inline(always)] +fn get_hash<K, V>(entries: &[Bucket<K, V>]) -> impl Fn(&usize) -> u64 + '_ { + move |&i| entries[i].hash.get() +} + +impl<K, V> Clone for IndexMapCore<K, V> +where + K: Clone, + V: Clone, +{ + fn clone(&self) -> Self { + let indices = self.indices.clone(); + let mut entries = Vec::with_capacity(indices.capacity()); + entries.clone_from(&self.entries); + IndexMapCore { indices, entries } + } + + fn clone_from(&mut self, other: &Self) { + let hasher = get_hash(&other.entries); + self.indices.clone_from_with_hasher(&other.indices, hasher); + if self.entries.capacity() < other.entries.len() { + // If we must resize, match the indices capacity + self.reserve_entries(); + } + self.entries.clone_from(&other.entries); + } +} + +impl<K, V> fmt::Debug for IndexMapCore<K, V> +where + K: fmt::Debug, + V: fmt::Debug, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_struct("IndexMapCore") + .field("indices", &raw::DebugIndices(&self.indices)) + .field("entries", &self.entries) + .finish() + } +} + +impl<K, V> Entries for IndexMapCore<K, V> { + type Entry = Bucket<K, V>; + + #[inline] + fn into_entries(self) -> Vec<Self::Entry> { + self.entries + } + + #[inline] + fn as_entries(&self) -> &[Self::Entry] { + &self.entries + } + + #[inline] + fn as_entries_mut(&mut self) -> &mut [Self::Entry] { + &mut self.entries + } + + fn with_entries<F>(&mut self, f: F) + where + F: FnOnce(&mut [Self::Entry]), + { + f(&mut self.entries); + self.rebuild_hash_table(); + } +} + +impl<K, V> IndexMapCore<K, V> { + #[inline] + pub(crate) fn new() -> Self { + IndexMapCore { + indices: RawTable::new(), + entries: Vec::new(), + } + } + + #[inline] + pub(crate) fn with_capacity(n: usize) -> Self { + IndexMapCore { + indices: RawTable::with_capacity(n), + entries: Vec::with_capacity(n), + } + } + + #[inline] + pub(crate) fn len(&self) -> usize { + self.indices.len() + } + + #[inline] + pub(crate) fn capacity(&self) -> usize { + cmp::min(self.indices.capacity(), self.entries.capacity()) + } + + pub(crate) fn clear(&mut self) { + self.indices.clear(); + self.entries.clear(); + } + + pub(crate) fn drain<R>(&mut self, range: R) -> Drain<'_, Bucket<K, V>> + where + R: RangeBounds<usize>, + { + let range = simplify_range(range, self.entries.len()); + self.erase_indices(range.start, range.end); + self.entries.drain(range) + } + + /// Reserve capacity for `additional` more key-value pairs. + pub(crate) fn reserve(&mut self, additional: usize) { + self.indices.reserve(additional, get_hash(&self.entries)); + self.reserve_entries(); + } + + /// Reserve entries capacity to match the indices + fn reserve_entries(&mut self) { + let additional = self.indices.capacity() - self.entries.len(); + self.entries.reserve_exact(additional); + } + + /// Shrink the capacity of the map as much as possible. + pub(crate) fn shrink_to_fit(&mut self) { + self.indices.shrink_to(0, get_hash(&self.entries)); + self.entries.shrink_to_fit(); + } + + /// Remove the last key-value pair + pub(crate) fn pop(&mut self) -> Option<(K, V)> { + if let Some(entry) = self.entries.pop() { + let last = self.entries.len(); + self.erase_index(entry.hash, last); + Some((entry.key, entry.value)) + } else { + None + } + } + + /// Append a key-value pair, *without* checking whether it already exists, + /// and return the pair's new index. + fn push(&mut self, hash: HashValue, key: K, value: V) -> usize { + let i = self.entries.len(); + self.indices.insert(hash.get(), i, get_hash(&self.entries)); + if i == self.entries.capacity() { + // Reserve our own capacity synced to the indices, + // rather than letting `Vec::push` just double it. + self.reserve_entries(); + } + self.entries.push(Bucket { hash, key, value }); + i + } + + pub(crate) fn insert_full(&mut self, hash: HashValue, key: K, value: V) -> (usize, Option<V>) + where + K: Eq, + { + match self.get_index_of(hash, &key) { + Some(i) => (i, Some(replace(&mut self.entries[i].value, value))), + None => (self.push(hash, key, value), None), + } + } + + pub(crate) fn retain_in_order<F>(&mut self, mut keep: F) + where + F: FnMut(&mut K, &mut V) -> bool, + { + // Like Vec::retain in self.entries, but with mutable K and V. + // We swap-shift all the items we want to keep, truncate the rest, + // then rebuild the raw hash table with the new indexes. + let len = self.entries.len(); + let mut n_deleted = 0; + for i in 0..len { + let will_keep = { + let entry = &mut self.entries[i]; + keep(&mut entry.key, &mut entry.value) + }; + if !will_keep { + n_deleted += 1; + } else if n_deleted > 0 { + self.entries.swap(i - n_deleted, i); + } + } + if n_deleted > 0 { + self.entries.truncate(len - n_deleted); + self.rebuild_hash_table(); + } + } + + fn rebuild_hash_table(&mut self) { + self.indices.clear(); + debug_assert!(self.indices.capacity() >= self.entries.len()); + for (i, entry) in enumerate(&self.entries) { + // We should never have to reallocate, so there's no need for a real hasher. + self.indices.insert_no_grow(entry.hash.get(), i); + } + } +} + +/// Entry for an existing key-value pair or a vacant location to +/// insert one. +pub enum Entry<'a, K, V> { + /// Existing slot with equivalent key. + Occupied(OccupiedEntry<'a, K, V>), + /// Vacant slot (no equivalent key in the map). + Vacant(VacantEntry<'a, K, V>), +} + +impl<'a, K, V> Entry<'a, K, V> { + /// Computes in **O(1)** time (amortized average). + pub fn or_insert(self, default: V) -> &'a mut V { + match self { + Entry::Occupied(entry) => entry.into_mut(), + Entry::Vacant(entry) => entry.insert(default), + } + } + + /// Computes in **O(1)** time (amortized average). + pub fn or_insert_with<F>(self, call: F) -> &'a mut V + where + F: FnOnce() -> V, + { + match self { + Entry::Occupied(entry) => entry.into_mut(), + Entry::Vacant(entry) => entry.insert(call()), + } + } + + pub fn key(&self) -> &K { + match *self { + Entry::Occupied(ref entry) => entry.key(), + Entry::Vacant(ref entry) => entry.key(), + } + } + + /// Return the index where the key-value pair exists or will be inserted. + pub fn index(&self) -> usize { + match *self { + Entry::Occupied(ref entry) => entry.index(), + Entry::Vacant(ref entry) => entry.index(), + } + } + + /// Modifies the entry if it is occupied. + pub fn and_modify<F>(self, f: F) -> Self + where + F: FnOnce(&mut V), + { + match self { + Entry::Occupied(mut o) => { + f(o.get_mut()); + Entry::Occupied(o) + } + x => x, + } + } + + /// Inserts a default-constructed value in the entry if it is vacant and returns a mutable + /// reference to it. Otherwise a mutable reference to an already existent value is returned. + /// + /// Computes in **O(1)** time (amortized average). + pub fn or_default(self) -> &'a mut V + where + V: Default, + { + match self { + Entry::Occupied(entry) => entry.into_mut(), + Entry::Vacant(entry) => entry.insert(V::default()), + } + } +} + +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for Entry<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match *self { + Entry::Vacant(ref v) => f.debug_tuple(stringify!(Entry)).field(v).finish(), + Entry::Occupied(ref o) => f.debug_tuple(stringify!(Entry)).field(o).finish(), + } + } +} + +pub use self::raw::OccupiedEntry; + +// Extra methods that don't threaten the unsafe encapsulation. +impl<K, V> OccupiedEntry<'_, K, V> { + /// Sets the value of the entry to `value`, and returns the entry's old value. + pub fn insert(&mut self, value: V) -> V { + replace(self.get_mut(), value) + } + + /// Remove the key, value pair stored in the map for this entry, and return the value. + /// + /// **NOTE:** This is equivalent to `.swap_remove()`. + pub fn remove(self) -> V { + self.swap_remove() + } + + /// Remove the key, value pair stored in the map for this entry, and return the value. + /// + /// Like `Vec::swap_remove`, the pair is removed by swapping it with the + /// last element of the map and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove(self) -> V { + self.swap_remove_entry().1 + } + + /// Remove the key, value pair stored in the map for this entry, and return the value. + /// + /// Like `Vec::remove`, the pair is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove(self) -> V { + self.shift_remove_entry().1 + } + + /// Remove and return the key, value pair stored in the map for this entry + /// + /// **NOTE:** This is equivalent to `.swap_remove_entry()`. + pub fn remove_entry(self) -> (K, V) { + self.swap_remove_entry() + } +} + +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for OccupiedEntry<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_struct(stringify!(OccupiedEntry)) + .field("key", self.key()) + .field("value", self.get()) + .finish() + } +} + +/// A view into a vacant entry in a `IndexMap`. +/// It is part of the [`Entry`] enum. +/// +/// [`Entry`]: enum.Entry.html +pub struct VacantEntry<'a, K, V> { + map: &'a mut IndexMapCore<K, V>, + hash: HashValue, + key: K, +} + +impl<'a, K, V> VacantEntry<'a, K, V> { + pub fn key(&self) -> &K { + &self.key + } + + pub fn into_key(self) -> K { + self.key + } + + /// Return the index where the key-value pair will be inserted. + pub fn index(&self) -> usize { + self.map.len() + } + + pub fn insert(self, value: V) -> &'a mut V { + let i = self.map.push(self.hash, self.key, value); + &mut self.map.entries[i].value + } +} + +impl<K: fmt::Debug, V> fmt::Debug for VacantEntry<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple(stringify!(VacantEntry)) + .field(self.key()) + .finish() + } +} + +#[test] +fn assert_send_sync() { + fn assert_send_sync<T: Send + Sync>() {} + assert_send_sync::<IndexMapCore<i32, i32>>(); + assert_send_sync::<Entry<'_, i32, i32>>(); +} diff --git a/third_party/rust/indexmap/src/map/core/raw.rs b/third_party/rust/indexmap/src/map/core/raw.rs new file mode 100644 index 0000000000..fdfadb91c1 --- /dev/null +++ b/third_party/rust/indexmap/src/map/core/raw.rs @@ -0,0 +1,335 @@ +#![allow(unsafe_code)] +//! This module encapsulates the `unsafe` access to `hashbrown::raw::RawTable`, +//! mostly in dealing with its bucket "pointers". + +use super::{Entry, Equivalent, HashValue, IndexMapCore, VacantEntry}; +use crate::util::enumerate; +use core::fmt; +use core::mem::replace; +use hashbrown::raw::RawTable; + +type RawBucket = hashbrown::raw::Bucket<usize>; + +pub(super) struct DebugIndices<'a>(pub &'a RawTable<usize>); +impl fmt::Debug for DebugIndices<'_> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let indices = unsafe { self.0.iter().map(|raw_bucket| raw_bucket.read()) }; + f.debug_list().entries(indices).finish() + } +} + +impl<K, V> IndexMapCore<K, V> { + /// Return the raw bucket with an equivalent key + fn find_equivalent<Q>(&self, hash: HashValue, key: &Q) -> Option<RawBucket> + where + Q: ?Sized + Equivalent<K>, + { + self.indices.find(hash.get(), { + move |&i| Q::equivalent(key, &self.entries[i].key) + }) + } + + /// Return the raw bucket for the given index + fn find_index(&self, hash: HashValue, index: usize) -> Option<RawBucket> { + self.indices.find(hash.get(), move |&i| i == index) + } + + /// Return the index in `entries` where an equivalent key can be found + pub(crate) fn get_index_of<Q>(&self, hash: HashValue, key: &Q) -> Option<usize> + where + Q: ?Sized + Equivalent<K>, + { + match self.find_equivalent(hash, key) { + Some(raw_bucket) => Some(unsafe { raw_bucket.read() }), + None => None, + } + } + + /// Erase the given index from `indices`. + /// + /// The index doesn't need to be valid in `entries` while calling this. No other index + /// adjustments are made -- this is only used by `pop` for the greatest index. + pub(super) fn erase_index(&mut self, hash: HashValue, index: usize) { + debug_assert_eq!(index, self.indices.len() - 1); + let raw_bucket = self.find_index(hash, index).unwrap(); + unsafe { self.indices.erase(raw_bucket) }; + } + + /// Erase `start..end` from `indices`, and shift `end..` indices down to `start..` + /// + /// All of these items should still be at their original location in `entries`. + /// This is used by `drain`, which will let `Vec::drain` do the work on `entries`. + pub(super) fn erase_indices(&mut self, start: usize, end: usize) { + let (init, shifted_entries) = self.entries.split_at(end); + let (start_entries, erased_entries) = init.split_at(start); + + let erased = erased_entries.len(); + let shifted = shifted_entries.len(); + let half_capacity = self.indices.buckets() / 2; + + // Use a heuristic between different strategies + if erased == 0 { + // Degenerate case, nothing to do + } else if start + shifted < half_capacity && start < erased { + // Reinsert everything, as there are few kept indices + self.indices.clear(); + + // Reinsert stable indices + for (i, entry) in enumerate(start_entries) { + self.indices.insert_no_grow(entry.hash.get(), i); + } + + // Reinsert shifted indices + for (i, entry) in (start..).zip(shifted_entries) { + self.indices.insert_no_grow(entry.hash.get(), i); + } + } else if erased + shifted < half_capacity { + // Find each affected index, as there are few to adjust + + // Find erased indices + for (i, entry) in (start..).zip(erased_entries) { + let bucket = self.find_index(entry.hash, i).unwrap(); + unsafe { self.indices.erase(bucket) }; + } + + // Find shifted indices + for ((new, old), entry) in (start..).zip(end..).zip(shifted_entries) { + let bucket = self.find_index(entry.hash, old).unwrap(); + unsafe { bucket.write(new) }; + } + } else { + // Sweep the whole table for adjustments + unsafe { + for bucket in self.indices.iter() { + let i = bucket.read(); + if i >= end { + bucket.write(i - erased); + } else if i >= start { + self.indices.erase(bucket); + } + } + } + } + + debug_assert_eq!(self.indices.len(), start + shifted); + } + + pub(crate) fn entry(&mut self, hash: HashValue, key: K) -> Entry<'_, K, V> + where + K: Eq, + { + match self.find_equivalent(hash, &key) { + // Safety: The entry is created with a live raw bucket, at the same time we have a &mut + // reference to the map, so it can not be modified further. + Some(raw_bucket) => Entry::Occupied(OccupiedEntry { + map: self, + raw_bucket, + key, + }), + None => Entry::Vacant(VacantEntry { + map: self, + hash, + key, + }), + } + } + + /// Remove an entry by shifting all entries that follow it + pub(crate) fn shift_remove_full<Q>(&mut self, hash: HashValue, key: &Q) -> Option<(usize, K, V)> + where + Q: ?Sized + Equivalent<K>, + { + match self.find_equivalent(hash, key) { + Some(raw_bucket) => unsafe { Some(self.shift_remove_bucket(raw_bucket)) }, + None => None, + } + } + + /// Remove an entry by shifting all entries that follow it + pub(crate) fn shift_remove_index(&mut self, index: usize) -> Option<(K, V)> { + let raw_bucket = match self.entries.get(index) { + Some(entry) => self.find_index(entry.hash, index).unwrap(), + None => return None, + }; + unsafe { + let (_, key, value) = self.shift_remove_bucket(raw_bucket); + Some((key, value)) + } + } + + /// Remove an entry by shifting all entries that follow it + /// + /// Safety: The caller must pass a live `raw_bucket`. + #[allow(unused_unsafe)] + unsafe fn shift_remove_bucket(&mut self, raw_bucket: RawBucket) -> (usize, K, V) { + // use Vec::remove, but then we need to update the indices that point + // to all of the other entries that have to move + let index = unsafe { self.indices.remove(raw_bucket) }; + let entry = self.entries.remove(index); + + // correct indices that point to the entries that followed the removed entry. + // use a heuristic between a full sweep vs. a `find()` for every shifted item. + let raw_capacity = self.indices.buckets(); + let shifted_entries = &self.entries[index..]; + if shifted_entries.len() > raw_capacity / 2 { + // shift all indices greater than `index` + unsafe { + for bucket in self.indices.iter() { + let i = bucket.read(); + if i > index { + bucket.write(i - 1); + } + } + } + } else { + // find each following entry to shift its index + for (i, entry) in (index + 1..).zip(shifted_entries) { + let shifted_bucket = self.find_index(entry.hash, i).unwrap(); + unsafe { shifted_bucket.write(i - 1) }; + } + } + + (index, entry.key, entry.value) + } + + /// Remove an entry by swapping it with the last + pub(crate) fn swap_remove_full<Q>(&mut self, hash: HashValue, key: &Q) -> Option<(usize, K, V)> + where + Q: ?Sized + Equivalent<K>, + { + match self.find_equivalent(hash, key) { + Some(raw_bucket) => unsafe { Some(self.swap_remove_bucket(raw_bucket)) }, + None => None, + } + } + + /// Remove an entry by swapping it with the last + pub(crate) fn swap_remove_index(&mut self, index: usize) -> Option<(K, V)> { + let raw_bucket = match self.entries.get(index) { + Some(entry) => self.find_index(entry.hash, index).unwrap(), + None => return None, + }; + unsafe { + let (_, key, value) = self.swap_remove_bucket(raw_bucket); + Some((key, value)) + } + } + + /// Remove an entry by swapping it with the last + /// + /// Safety: The caller must pass a live `raw_bucket`. + #[allow(unused_unsafe)] + unsafe fn swap_remove_bucket(&mut self, raw_bucket: RawBucket) -> (usize, K, V) { + // use swap_remove, but then we need to update the index that points + // to the other entry that has to move + let index = unsafe { self.indices.remove(raw_bucket) }; + let entry = self.entries.swap_remove(index); + + // correct index that points to the entry that had to swap places + if let Some(entry) = self.entries.get(index) { + // was not last element + // examine new element in `index` and find it in indices + let last = self.entries.len(); + let swapped_bucket = self.find_index(entry.hash, last).unwrap(); + unsafe { swapped_bucket.write(index) }; + } + + (index, entry.key, entry.value) + } + + pub(crate) fn reverse(&mut self) { + self.entries.reverse(); + + // No need to save hash indices, can easily calculate what they should + // be, given that this is an in-place reversal. + let len = self.entries.len(); + unsafe { + for raw_bucket in self.indices.iter() { + let i = raw_bucket.read(); + raw_bucket.write(len - i - 1); + } + } + } +} + +/// A view into an occupied entry in a `IndexMap`. +/// It is part of the [`Entry`] enum. +/// +/// [`Entry`]: enum.Entry.html +// SAFETY: The lifetime of the map reference also constrains the raw bucket, +// which is essentially a raw pointer into the map indices. +pub struct OccupiedEntry<'a, K, V> { + map: &'a mut IndexMapCore<K, V>, + raw_bucket: RawBucket, + key: K, +} + +// `hashbrown::raw::Bucket` is only `Send`, not `Sync`. +// SAFETY: `&self` only accesses the bucket to read it. +unsafe impl<K: Sync, V: Sync> Sync for OccupiedEntry<'_, K, V> {} + +// The parent module also adds methods that don't threaten the unsafe encapsulation. +impl<'a, K, V> OccupiedEntry<'a, K, V> { + pub fn key(&self) -> &K { + &self.key + } + + pub fn get(&self) -> &V { + &self.map.entries[self.index()].value + } + + pub fn get_mut(&mut self) -> &mut V { + let index = self.index(); + &mut self.map.entries[index].value + } + + /// Put the new key in the occupied entry's key slot + pub(crate) fn replace_key(self) -> K { + let index = self.index(); + let old_key = &mut self.map.entries[index].key; + replace(old_key, self.key) + } + + /// Return the index of the key-value pair + #[inline] + pub fn index(&self) -> usize { + unsafe { self.raw_bucket.read() } + } + + pub fn into_mut(self) -> &'a mut V { + let index = self.index(); + &mut self.map.entries[index].value + } + + /// Remove and return the key, value pair stored in the map for this entry + /// + /// Like `Vec::swap_remove`, the pair is removed by swapping it with the + /// last element of the map and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove_entry(self) -> (K, V) { + // This is safe because it can only happen once (self is consumed) + // and map.indices have not been modified since entry construction + unsafe { + let (_, key, value) = self.map.swap_remove_bucket(self.raw_bucket); + (key, value) + } + } + + /// Remove and return the key, value pair stored in the map for this entry + /// + /// Like `Vec::remove`, the pair is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove_entry(self) -> (K, V) { + // This is safe because it can only happen once (self is consumed) + // and map.indices have not been modified since entry construction + unsafe { + let (_, key, value) = self.map.shift_remove_bucket(self.raw_bucket); + (key, value) + } + } +} diff --git a/third_party/rust/indexmap/src/mutable_keys.rs b/third_party/rust/indexmap/src/mutable_keys.rs new file mode 100644 index 0000000000..0688441ee9 --- /dev/null +++ b/third_party/rust/indexmap/src/mutable_keys.rs @@ -0,0 +1,75 @@ +use core::hash::{BuildHasher, Hash}; + +use super::{Equivalent, IndexMap}; + +pub struct PrivateMarker {} + +/// Opt-in mutable access to keys. +/// +/// These methods expose `&mut K`, mutable references to the key as it is stored +/// in the map. +/// You are allowed to modify the keys in the hashmap **if the modifcation +/// does not change the key’s hash and equality**. +/// +/// If keys are modified erronously, you can no longer look them up. +/// This is sound (memory safe) but a logical error hazard (just like +/// implementing PartialEq, Eq, or Hash incorrectly would be). +/// +/// `use` this trait to enable its methods for `IndexMap`. +pub trait MutableKeys { + type Key; + type Value; + + /// Return item index, mutable reference to key and value + fn get_full_mut2<Q: ?Sized>( + &mut self, + key: &Q, + ) -> Option<(usize, &mut Self::Key, &mut Self::Value)> + where + Q: Hash + Equivalent<Self::Key>; + + /// Scan through each key-value pair in the map and keep those where the + /// closure `keep` returns `true`. + /// + /// The elements are visited in order, and remaining elements keep their + /// order. + /// + /// Computes in **O(n)** time (average). + fn retain2<F>(&mut self, keep: F) + where + F: FnMut(&mut Self::Key, &mut Self::Value) -> bool; + + /// This method is not useful in itself – it is there to “seal” the trait + /// for external implementation, so that we can add methods without + /// causing breaking changes. + fn __private_marker(&self) -> PrivateMarker; +} + +/// Opt-in mutable access to keys. +/// +/// See [`MutableKeys`](trait.MutableKeys.html) for more information. +impl<K, V, S> MutableKeys for IndexMap<K, V, S> +where + K: Eq + Hash, + S: BuildHasher, +{ + type Key = K; + type Value = V; + fn get_full_mut2<Q: ?Sized>(&mut self, key: &Q) -> Option<(usize, &mut K, &mut V)> + where + Q: Hash + Equivalent<K>, + { + self.get_full_mut2_impl(key) + } + + fn retain2<F>(&mut self, keep: F) + where + F: FnMut(&mut K, &mut V) -> bool, + { + self.retain_mut(keep) + } + + fn __private_marker(&self) -> PrivateMarker { + PrivateMarker {} + } +} diff --git a/third_party/rust/indexmap/src/rayon/map.rs b/third_party/rust/indexmap/src/rayon/map.rs new file mode 100644 index 0000000000..ed4c74e438 --- /dev/null +++ b/third_party/rust/indexmap/src/rayon/map.rs @@ -0,0 +1,496 @@ +//! Parallel iterator types for `IndexMap` with [rayon](https://docs.rs/rayon/1.0/rayon). +//! +//! You will rarely need to interact with this module directly unless you need to name one of the +//! iterator types. +//! +//! Requires crate feature `"rayon"` + +use super::collect; +use rayon::iter::plumbing::{Consumer, ProducerCallback, UnindexedConsumer}; +use rayon::prelude::*; + +use crate::vec::Vec; +use core::cmp::Ordering; +use core::fmt; +use core::hash::{BuildHasher, Hash}; + +use crate::Bucket; +use crate::Entries; +use crate::IndexMap; + +/// Requires crate feature `"rayon"`. +impl<K, V, S> IntoParallelIterator for IndexMap<K, V, S> +where + K: Send, + V: Send, +{ + type Item = (K, V); + type Iter = IntoParIter<K, V>; + + fn into_par_iter(self) -> Self::Iter { + IntoParIter { + entries: self.into_entries(), + } + } +} + +/// A parallel owning iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`into_par_iter`] method on [`IndexMap`] +/// (provided by rayon's `IntoParallelIterator` trait). See its documentation for more. +/// +/// [`into_par_iter`]: ../struct.IndexMap.html#method.into_par_iter +/// [`IndexMap`]: ../struct.IndexMap.html +pub struct IntoParIter<K, V> { + entries: Vec<Bucket<K, V>>, +} + +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for IntoParIter<K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.entries.iter().map(Bucket::refs); + f.debug_list().entries(iter).finish() + } +} + +impl<K: Send, V: Send> ParallelIterator for IntoParIter<K, V> { + type Item = (K, V); + + parallel_iterator_methods!(Bucket::key_value); +} + +impl<K: Send, V: Send> IndexedParallelIterator for IntoParIter<K, V> { + indexed_parallel_iterator_methods!(Bucket::key_value); +} + +/// Requires crate feature `"rayon"`. +impl<'a, K, V, S> IntoParallelIterator for &'a IndexMap<K, V, S> +where + K: Sync, + V: Sync, +{ + type Item = (&'a K, &'a V); + type Iter = ParIter<'a, K, V>; + + fn into_par_iter(self) -> Self::Iter { + ParIter { + entries: self.as_entries(), + } + } +} + +/// A parallel iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`par_iter`] method on [`IndexMap`] +/// (provided by rayon's `IntoParallelRefIterator` trait). See its documentation for more. +/// +/// [`par_iter`]: ../struct.IndexMap.html#method.par_iter +/// [`IndexMap`]: ../struct.IndexMap.html +pub struct ParIter<'a, K, V> { + entries: &'a [Bucket<K, V>], +} + +impl<K, V> Clone for ParIter<'_, K, V> { + fn clone(&self) -> Self { + ParIter { ..*self } + } +} + +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for ParIter<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.entries.iter().map(Bucket::refs); + f.debug_list().entries(iter).finish() + } +} + +impl<'a, K: Sync, V: Sync> ParallelIterator for ParIter<'a, K, V> { + type Item = (&'a K, &'a V); + + parallel_iterator_methods!(Bucket::refs); +} + +impl<K: Sync, V: Sync> IndexedParallelIterator for ParIter<'_, K, V> { + indexed_parallel_iterator_methods!(Bucket::refs); +} + +/// Requires crate feature `"rayon"`. +impl<'a, K, V, S> IntoParallelIterator for &'a mut IndexMap<K, V, S> +where + K: Sync + Send, + V: Send, +{ + type Item = (&'a K, &'a mut V); + type Iter = ParIterMut<'a, K, V>; + + fn into_par_iter(self) -> Self::Iter { + ParIterMut { + entries: self.as_entries_mut(), + } + } +} + +/// A parallel mutable iterator over the entries of a `IndexMap`. +/// +/// This `struct` is created by the [`par_iter_mut`] method on [`IndexMap`] +/// (provided by rayon's `IntoParallelRefMutIterator` trait). See its documentation for more. +/// +/// [`par_iter_mut`]: ../struct.IndexMap.html#method.par_iter_mut +/// [`IndexMap`]: ../struct.IndexMap.html +pub struct ParIterMut<'a, K, V> { + entries: &'a mut [Bucket<K, V>], +} + +impl<'a, K: Sync + Send, V: Send> ParallelIterator for ParIterMut<'a, K, V> { + type Item = (&'a K, &'a mut V); + + parallel_iterator_methods!(Bucket::ref_mut); +} + +impl<K: Sync + Send, V: Send> IndexedParallelIterator for ParIterMut<'_, K, V> { + indexed_parallel_iterator_methods!(Bucket::ref_mut); +} + +/// Parallel iterator methods and other parallel methods. +/// +/// The following methods **require crate feature `"rayon"`**. +/// +/// See also the `IntoParallelIterator` implementations. +impl<K, V, S> IndexMap<K, V, S> +where + K: Sync, + V: Sync, +{ + /// Return a parallel iterator over the keys of the map. + /// + /// While parallel iterators can process items in any order, their relative order + /// in the map is still preserved for operations like `reduce` and `collect`. + pub fn par_keys(&self) -> ParKeys<'_, K, V> { + ParKeys { + entries: self.as_entries(), + } + } + + /// Return a parallel iterator over the values of the map. + /// + /// While parallel iterators can process items in any order, their relative order + /// in the map is still preserved for operations like `reduce` and `collect`. + pub fn par_values(&self) -> ParValues<'_, K, V> { + ParValues { + entries: self.as_entries(), + } + } +} + +impl<K, V, S> IndexMap<K, V, S> +where + K: Hash + Eq + Sync, + V: Sync, + S: BuildHasher, +{ + /// Returns `true` if `self` contains all of the same key-value pairs as `other`, + /// regardless of each map's indexed order, determined in parallel. + pub fn par_eq<V2, S2>(&self, other: &IndexMap<K, V2, S2>) -> bool + where + V: PartialEq<V2>, + V2: Sync, + S2: BuildHasher + Sync, + { + self.len() == other.len() + && self + .par_iter() + .all(move |(key, value)| other.get(key).map_or(false, |v| *value == *v)) + } +} + +/// A parallel iterator over the keys of a `IndexMap`. +/// +/// This `struct` is created by the [`par_keys`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`par_keys`]: ../struct.IndexMap.html#method.par_keys +/// [`IndexMap`]: ../struct.IndexMap.html +pub struct ParKeys<'a, K, V> { + entries: &'a [Bucket<K, V>], +} + +impl<K, V> Clone for ParKeys<'_, K, V> { + fn clone(&self) -> Self { + ParKeys { ..*self } + } +} + +impl<K: fmt::Debug, V> fmt::Debug for ParKeys<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.entries.iter().map(Bucket::key_ref); + f.debug_list().entries(iter).finish() + } +} + +impl<'a, K: Sync, V: Sync> ParallelIterator for ParKeys<'a, K, V> { + type Item = &'a K; + + parallel_iterator_methods!(Bucket::key_ref); +} + +impl<K: Sync, V: Sync> IndexedParallelIterator for ParKeys<'_, K, V> { + indexed_parallel_iterator_methods!(Bucket::key_ref); +} + +/// A parallel iterator over the values of a `IndexMap`. +/// +/// This `struct` is created by the [`par_values`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`par_values`]: ../struct.IndexMap.html#method.par_values +/// [`IndexMap`]: ../struct.IndexMap.html +pub struct ParValues<'a, K, V> { + entries: &'a [Bucket<K, V>], +} + +impl<K, V> Clone for ParValues<'_, K, V> { + fn clone(&self) -> Self { + ParValues { ..*self } + } +} + +impl<K, V: fmt::Debug> fmt::Debug for ParValues<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.entries.iter().map(Bucket::value_ref); + f.debug_list().entries(iter).finish() + } +} + +impl<'a, K: Sync, V: Sync> ParallelIterator for ParValues<'a, K, V> { + type Item = &'a V; + + parallel_iterator_methods!(Bucket::value_ref); +} + +impl<K: Sync, V: Sync> IndexedParallelIterator for ParValues<'_, K, V> { + indexed_parallel_iterator_methods!(Bucket::value_ref); +} + +/// Requires crate feature `"rayon"`. +impl<K, V, S> IndexMap<K, V, S> +where + K: Send, + V: Send, +{ + /// Return a parallel iterator over mutable references to the the values of the map + /// + /// While parallel iterators can process items in any order, their relative order + /// in the map is still preserved for operations like `reduce` and `collect`. + pub fn par_values_mut(&mut self) -> ParValuesMut<'_, K, V> { + ParValuesMut { + entries: self.as_entries_mut(), + } + } +} + +impl<K, V, S> IndexMap<K, V, S> +where + K: Hash + Eq + Send, + V: Send, + S: BuildHasher, +{ + /// Sort the map’s key-value pairs in parallel, by the default ordering of the keys. + pub fn par_sort_keys(&mut self) + where + K: Ord, + { + self.with_entries(|entries| { + entries.par_sort_by(|a, b| K::cmp(&a.key, &b.key)); + }); + } + + /// Sort the map’s key-value pairs in place and in parallel, using the comparison + /// function `compare`. + /// + /// The comparison function receives two key and value pairs to compare (you + /// can sort by keys or values or their combination as needed). + pub fn par_sort_by<F>(&mut self, cmp: F) + where + F: Fn(&K, &V, &K, &V) -> Ordering + Sync, + { + self.with_entries(|entries| { + entries.par_sort_by(move |a, b| cmp(&a.key, &a.value, &b.key, &b.value)); + }); + } + + /// Sort the key-value pairs of the map in parallel and return a by value parallel + /// iterator of the key-value pairs with the result. + pub fn par_sorted_by<F>(self, cmp: F) -> IntoParIter<K, V> + where + F: Fn(&K, &V, &K, &V) -> Ordering + Sync, + { + let mut entries = self.into_entries(); + entries.par_sort_by(move |a, b| cmp(&a.key, &a.value, &b.key, &b.value)); + IntoParIter { entries } + } +} + +/// A parallel mutable iterator over the values of a `IndexMap`. +/// +/// This `struct` is created by the [`par_values_mut`] method on [`IndexMap`]. See its +/// documentation for more. +/// +/// [`par_values_mut`]: ../struct.IndexMap.html#method.par_values_mut +/// [`IndexMap`]: ../struct.IndexMap.html +pub struct ParValuesMut<'a, K, V> { + entries: &'a mut [Bucket<K, V>], +} + +impl<'a, K: Send, V: Send> ParallelIterator for ParValuesMut<'a, K, V> { + type Item = &'a mut V; + + parallel_iterator_methods!(Bucket::value_mut); +} + +impl<K: Send, V: Send> IndexedParallelIterator for ParValuesMut<'_, K, V> { + indexed_parallel_iterator_methods!(Bucket::value_mut); +} + +/// Requires crate feature `"rayon"`. +impl<K, V, S> FromParallelIterator<(K, V)> for IndexMap<K, V, S> +where + K: Eq + Hash + Send, + V: Send, + S: BuildHasher + Default + Send, +{ + fn from_par_iter<I>(iter: I) -> Self + where + I: IntoParallelIterator<Item = (K, V)>, + { + let list = collect(iter); + let len = list.iter().map(Vec::len).sum(); + let mut map = Self::with_capacity_and_hasher(len, S::default()); + for vec in list { + map.extend(vec); + } + map + } +} + +/// Requires crate feature `"rayon"`. +impl<K, V, S> ParallelExtend<(K, V)> for IndexMap<K, V, S> +where + K: Eq + Hash + Send, + V: Send, + S: BuildHasher + Send, +{ + fn par_extend<I>(&mut self, iter: I) + where + I: IntoParallelIterator<Item = (K, V)>, + { + for vec in collect(iter) { + self.extend(vec); + } + } +} + +/// Requires crate feature `"rayon"`. +impl<'a, K: 'a, V: 'a, S> ParallelExtend<(&'a K, &'a V)> for IndexMap<K, V, S> +where + K: Copy + Eq + Hash + Send + Sync, + V: Copy + Send + Sync, + S: BuildHasher + Send, +{ + fn par_extend<I>(&mut self, iter: I) + where + I: IntoParallelIterator<Item = (&'a K, &'a V)>, + { + for vec in collect(iter) { + self.extend(vec); + } + } +} + +#[cfg(test)] +mod tests { + use super::*; + use std::string::String; + + #[test] + fn insert_order() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut map = IndexMap::new(); + + for &elt in &insert { + map.insert(elt, ()); + } + + assert_eq!(map.par_keys().count(), map.len()); + assert_eq!(map.par_keys().count(), insert.len()); + insert.par_iter().zip(map.par_keys()).for_each(|(a, b)| { + assert_eq!(a, b); + }); + (0..insert.len()) + .into_par_iter() + .zip(map.par_keys()) + .for_each(|(i, k)| { + assert_eq!(map.get_index(i).unwrap().0, k); + }); + } + + #[test] + fn partial_eq_and_eq() { + let mut map_a = IndexMap::new(); + map_a.insert(1, "1"); + map_a.insert(2, "2"); + let mut map_b = map_a.clone(); + assert!(map_a.par_eq(&map_b)); + map_b.swap_remove(&1); + assert!(!map_a.par_eq(&map_b)); + map_b.insert(3, "3"); + assert!(!map_a.par_eq(&map_b)); + + let map_c: IndexMap<_, String> = + map_b.into_par_iter().map(|(k, v)| (k, v.into())).collect(); + assert!(!map_a.par_eq(&map_c)); + assert!(!map_c.par_eq(&map_a)); + } + + #[test] + fn extend() { + let mut map = IndexMap::new(); + map.par_extend(vec![(&1, &2), (&3, &4)]); + map.par_extend(vec![(5, 6)]); + assert_eq!( + map.into_par_iter().collect::<Vec<_>>(), + vec![(1, 2), (3, 4), (5, 6)] + ); + } + + #[test] + fn keys() { + let vec = vec![(1, 'a'), (2, 'b'), (3, 'c')]; + let map: IndexMap<_, _> = vec.into_par_iter().collect(); + let keys: Vec<_> = map.par_keys().cloned().collect(); + assert_eq!(keys.len(), 3); + assert!(keys.contains(&1)); + assert!(keys.contains(&2)); + assert!(keys.contains(&3)); + } + + #[test] + fn values() { + let vec = vec![(1, 'a'), (2, 'b'), (3, 'c')]; + let map: IndexMap<_, _> = vec.into_par_iter().collect(); + let values: Vec<_> = map.par_values().cloned().collect(); + assert_eq!(values.len(), 3); + assert!(values.contains(&'a')); + assert!(values.contains(&'b')); + assert!(values.contains(&'c')); + } + + #[test] + fn values_mut() { + let vec = vec![(1, 1), (2, 2), (3, 3)]; + let mut map: IndexMap<_, _> = vec.into_par_iter().collect(); + map.par_values_mut().for_each(|value| *value *= 2); + let values: Vec<_> = map.par_values().cloned().collect(); + assert_eq!(values.len(), 3); + assert!(values.contains(&2)); + assert!(values.contains(&4)); + assert!(values.contains(&6)); + } +} diff --git a/third_party/rust/indexmap/src/rayon/mod.rs b/third_party/rust/indexmap/src/rayon/mod.rs new file mode 100644 index 0000000000..57c810beea --- /dev/null +++ b/third_party/rust/indexmap/src/rayon/mod.rs @@ -0,0 +1,83 @@ +use rayon::prelude::*; + +#[cfg(not(has_std))] +use alloc::collections::LinkedList; + +#[cfg(has_std)] +use std::collections::LinkedList; + +use crate::vec::Vec; + +// generate `ParallelIterator` methods by just forwarding to the underlying +// self.entries and mapping its elements. +macro_rules! parallel_iterator_methods { + // $map_elt is the mapping function from the underlying iterator's element + ($map_elt:expr) => { + fn drive_unindexed<C>(self, consumer: C) -> C::Result + where + C: UnindexedConsumer<Self::Item>, + { + self.entries + .into_par_iter() + .map($map_elt) + .drive_unindexed(consumer) + } + + // NB: This allows indexed collection, e.g. directly into a `Vec`, but the + // underlying iterator must really be indexed. We should remove this if we + // start having tombstones that must be filtered out. + fn opt_len(&self) -> Option<usize> { + Some(self.entries.len()) + } + }; +} + +// generate `IndexedParallelIterator` methods by just forwarding to the underlying +// self.entries and mapping its elements. +macro_rules! indexed_parallel_iterator_methods { + // $map_elt is the mapping function from the underlying iterator's element + ($map_elt:expr) => { + fn drive<C>(self, consumer: C) -> C::Result + where + C: Consumer<Self::Item>, + { + self.entries.into_par_iter().map($map_elt).drive(consumer) + } + + fn len(&self) -> usize { + self.entries.len() + } + + fn with_producer<CB>(self, callback: CB) -> CB::Output + where + CB: ProducerCallback<Self::Item>, + { + self.entries + .into_par_iter() + .map($map_elt) + .with_producer(callback) + } + }; +} + +pub mod map; +pub mod set; + +// This form of intermediate collection is also how Rayon collects `HashMap`. +// Note that the order will also be preserved! +fn collect<I: IntoParallelIterator>(iter: I) -> LinkedList<Vec<I::Item>> { + iter.into_par_iter() + .fold(Vec::new, |mut vec, elem| { + vec.push(elem); + vec + }) + .map(|vec| { + let mut list = LinkedList::new(); + list.push_back(vec); + list + }) + .reduce(LinkedList::new, |mut list1, mut list2| { + list1.append(&mut list2); + list1 + }) +} diff --git a/third_party/rust/indexmap/src/rayon/set.rs b/third_party/rust/indexmap/src/rayon/set.rs new file mode 100644 index 0000000000..f6e08d4837 --- /dev/null +++ b/third_party/rust/indexmap/src/rayon/set.rs @@ -0,0 +1,673 @@ +//! Parallel iterator types for `IndexSet` with [rayon](https://docs.rs/rayon/1.0/rayon). +//! +//! You will rarely need to interact with this module directly unless you need to name one of the +//! iterator types. +//! +//! Requires crate feature `"rayon"`. + +use super::collect; +use rayon::iter::plumbing::{Consumer, ProducerCallback, UnindexedConsumer}; +use rayon::prelude::*; + +use crate::vec::Vec; +use core::cmp::Ordering; +use core::fmt; +use core::hash::{BuildHasher, Hash}; + +use crate::Entries; +use crate::IndexSet; + +type Bucket<T> = crate::Bucket<T, ()>; + +/// Requires crate feature `"rayon"`. +impl<T, S> IntoParallelIterator for IndexSet<T, S> +where + T: Send, +{ + type Item = T; + type Iter = IntoParIter<T>; + + fn into_par_iter(self) -> Self::Iter { + IntoParIter { + entries: self.into_entries(), + } + } +} + +/// A parallel owning iterator over the items of a `IndexSet`. +/// +/// This `struct` is created by the [`into_par_iter`] method on [`IndexSet`] +/// (provided by rayon's `IntoParallelIterator` trait). See its documentation for more. +/// +/// [`IndexSet`]: ../struct.IndexSet.html +/// [`into_par_iter`]: ../struct.IndexSet.html#method.into_par_iter +pub struct IntoParIter<T> { + entries: Vec<Bucket<T>>, +} + +impl<T: fmt::Debug> fmt::Debug for IntoParIter<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.entries.iter().map(Bucket::key_ref); + f.debug_list().entries(iter).finish() + } +} + +impl<T: Send> ParallelIterator for IntoParIter<T> { + type Item = T; + + parallel_iterator_methods!(Bucket::key); +} + +impl<T: Send> IndexedParallelIterator for IntoParIter<T> { + indexed_parallel_iterator_methods!(Bucket::key); +} + +/// Requires crate feature `"rayon"`. +impl<'a, T, S> IntoParallelIterator for &'a IndexSet<T, S> +where + T: Sync, +{ + type Item = &'a T; + type Iter = ParIter<'a, T>; + + fn into_par_iter(self) -> Self::Iter { + ParIter { + entries: self.as_entries(), + } + } +} + +/// A parallel iterator over the items of a `IndexSet`. +/// +/// This `struct` is created by the [`par_iter`] method on [`IndexSet`] +/// (provided by rayon's `IntoParallelRefIterator` trait). See its documentation for more. +/// +/// [`IndexSet`]: ../struct.IndexSet.html +/// [`par_iter`]: ../struct.IndexSet.html#method.par_iter +pub struct ParIter<'a, T> { + entries: &'a [Bucket<T>], +} + +impl<T> Clone for ParIter<'_, T> { + fn clone(&self) -> Self { + ParIter { ..*self } + } +} + +impl<T: fmt::Debug> fmt::Debug for ParIter<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.entries.iter().map(Bucket::key_ref); + f.debug_list().entries(iter).finish() + } +} + +impl<'a, T: Sync> ParallelIterator for ParIter<'a, T> { + type Item = &'a T; + + parallel_iterator_methods!(Bucket::key_ref); +} + +impl<T: Sync> IndexedParallelIterator for ParIter<'_, T> { + indexed_parallel_iterator_methods!(Bucket::key_ref); +} + +/// Parallel iterator methods and other parallel methods. +/// +/// The following methods **require crate feature `"rayon"`**. +/// +/// See also the `IntoParallelIterator` implementations. +impl<T, S> IndexSet<T, S> +where + T: Hash + Eq + Sync, + S: BuildHasher + Sync, +{ + /// Return a parallel iterator over the values that are in `self` but not `other`. + /// + /// While parallel iterators can process items in any order, their relative order + /// in the `self` set is still preserved for operations like `reduce` and `collect`. + pub fn par_difference<'a, S2>( + &'a self, + other: &'a IndexSet<T, S2>, + ) -> ParDifference<'a, T, S, S2> + where + S2: BuildHasher + Sync, + { + ParDifference { + set1: self, + set2: other, + } + } + + /// Return a parallel iterator over the values that are in `self` or `other`, + /// but not in both. + /// + /// While parallel iterators can process items in any order, their relative order + /// in the sets is still preserved for operations like `reduce` and `collect`. + /// Values from `self` are produced in their original order, followed by + /// values from `other` in their original order. + pub fn par_symmetric_difference<'a, S2>( + &'a self, + other: &'a IndexSet<T, S2>, + ) -> ParSymmetricDifference<'a, T, S, S2> + where + S2: BuildHasher + Sync, + { + ParSymmetricDifference { + set1: self, + set2: other, + } + } + + /// Return a parallel iterator over the values that are in both `self` and `other`. + /// + /// While parallel iterators can process items in any order, their relative order + /// in the `self` set is still preserved for operations like `reduce` and `collect`. + pub fn par_intersection<'a, S2>( + &'a self, + other: &'a IndexSet<T, S2>, + ) -> ParIntersection<'a, T, S, S2> + where + S2: BuildHasher + Sync, + { + ParIntersection { + set1: self, + set2: other, + } + } + + /// Return a parallel iterator over all values that are in `self` or `other`. + /// + /// While parallel iterators can process items in any order, their relative order + /// in the sets is still preserved for operations like `reduce` and `collect`. + /// Values from `self` are produced in their original order, followed by + /// values that are unique to `other` in their original order. + pub fn par_union<'a, S2>(&'a self, other: &'a IndexSet<T, S2>) -> ParUnion<'a, T, S, S2> + where + S2: BuildHasher + Sync, + { + ParUnion { + set1: self, + set2: other, + } + } + + /// Returns `true` if `self` contains all of the same values as `other`, + /// regardless of each set's indexed order, determined in parallel. + pub fn par_eq<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher + Sync, + { + self.len() == other.len() && self.par_is_subset(other) + } + + /// Returns `true` if `self` has no elements in common with `other`, + /// determined in parallel. + pub fn par_is_disjoint<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher + Sync, + { + if self.len() <= other.len() { + self.par_iter().all(move |value| !other.contains(value)) + } else { + other.par_iter().all(move |value| !self.contains(value)) + } + } + + /// Returns `true` if all elements of `other` are contained in `self`, + /// determined in parallel. + pub fn par_is_superset<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher + Sync, + { + other.par_is_subset(self) + } + + /// Returns `true` if all elements of `self` are contained in `other`, + /// determined in parallel. + pub fn par_is_subset<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher + Sync, + { + self.len() <= other.len() && self.par_iter().all(move |value| other.contains(value)) + } +} + +/// A parallel iterator producing elements in the difference of `IndexSet`s. +/// +/// This `struct` is created by the [`par_difference`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: ../struct.IndexSet.html +/// [`par_difference`]: ../struct.IndexSet.html#method.par_difference +pub struct ParDifference<'a, T, S1, S2> { + set1: &'a IndexSet<T, S1>, + set2: &'a IndexSet<T, S2>, +} + +impl<T, S1, S2> Clone for ParDifference<'_, T, S1, S2> { + fn clone(&self) -> Self { + ParDifference { ..*self } + } +} + +impl<T, S1, S2> fmt::Debug for ParDifference<'_, T, S1, S2> +where + T: fmt::Debug + Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list() + .entries(self.set1.difference(&self.set2)) + .finish() + } +} + +impl<'a, T, S1, S2> ParallelIterator for ParDifference<'a, T, S1, S2> +where + T: Hash + Eq + Sync, + S1: BuildHasher + Sync, + S2: BuildHasher + Sync, +{ + type Item = &'a T; + + fn drive_unindexed<C>(self, consumer: C) -> C::Result + where + C: UnindexedConsumer<Self::Item>, + { + let Self { set1, set2 } = self; + + set1.par_iter() + .filter(move |&item| !set2.contains(item)) + .drive_unindexed(consumer) + } +} + +/// A parallel iterator producing elements in the intersection of `IndexSet`s. +/// +/// This `struct` is created by the [`par_intersection`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: ../struct.IndexSet.html +/// [`par_intersection`]: ../struct.IndexSet.html#method.par_intersection +pub struct ParIntersection<'a, T, S1, S2> { + set1: &'a IndexSet<T, S1>, + set2: &'a IndexSet<T, S2>, +} + +impl<T, S1, S2> Clone for ParIntersection<'_, T, S1, S2> { + fn clone(&self) -> Self { + ParIntersection { ..*self } + } +} + +impl<T, S1, S2> fmt::Debug for ParIntersection<'_, T, S1, S2> +where + T: fmt::Debug + Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list() + .entries(self.set1.intersection(&self.set2)) + .finish() + } +} + +impl<'a, T, S1, S2> ParallelIterator for ParIntersection<'a, T, S1, S2> +where + T: Hash + Eq + Sync, + S1: BuildHasher + Sync, + S2: BuildHasher + Sync, +{ + type Item = &'a T; + + fn drive_unindexed<C>(self, consumer: C) -> C::Result + where + C: UnindexedConsumer<Self::Item>, + { + let Self { set1, set2 } = self; + + set1.par_iter() + .filter(move |&item| set2.contains(item)) + .drive_unindexed(consumer) + } +} + +/// A parallel iterator producing elements in the symmetric difference of `IndexSet`s. +/// +/// This `struct` is created by the [`par_symmetric_difference`] method on +/// [`IndexSet`]. See its documentation for more. +/// +/// [`IndexSet`]: ../struct.IndexSet.html +/// [`par_symmetric_difference`]: ../struct.IndexSet.html#method.par_symmetric_difference +pub struct ParSymmetricDifference<'a, T, S1, S2> { + set1: &'a IndexSet<T, S1>, + set2: &'a IndexSet<T, S2>, +} + +impl<T, S1, S2> Clone for ParSymmetricDifference<'_, T, S1, S2> { + fn clone(&self) -> Self { + ParSymmetricDifference { ..*self } + } +} + +impl<T, S1, S2> fmt::Debug for ParSymmetricDifference<'_, T, S1, S2> +where + T: fmt::Debug + Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list() + .entries(self.set1.symmetric_difference(&self.set2)) + .finish() + } +} + +impl<'a, T, S1, S2> ParallelIterator for ParSymmetricDifference<'a, T, S1, S2> +where + T: Hash + Eq + Sync, + S1: BuildHasher + Sync, + S2: BuildHasher + Sync, +{ + type Item = &'a T; + + fn drive_unindexed<C>(self, consumer: C) -> C::Result + where + C: UnindexedConsumer<Self::Item>, + { + let Self { set1, set2 } = self; + + set1.par_difference(set2) + .chain(set2.par_difference(set1)) + .drive_unindexed(consumer) + } +} + +/// A parallel iterator producing elements in the union of `IndexSet`s. +/// +/// This `struct` is created by the [`par_union`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: ../struct.IndexSet.html +/// [`par_union`]: ../struct.IndexSet.html#method.par_union +pub struct ParUnion<'a, T, S1, S2> { + set1: &'a IndexSet<T, S1>, + set2: &'a IndexSet<T, S2>, +} + +impl<T, S1, S2> Clone for ParUnion<'_, T, S1, S2> { + fn clone(&self) -> Self { + ParUnion { ..*self } + } +} + +impl<T, S1, S2> fmt::Debug for ParUnion<'_, T, S1, S2> +where + T: fmt::Debug + Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.set1.union(&self.set2)).finish() + } +} + +impl<'a, T, S1, S2> ParallelIterator for ParUnion<'a, T, S1, S2> +where + T: Hash + Eq + Sync, + S1: BuildHasher + Sync, + S2: BuildHasher + Sync, +{ + type Item = &'a T; + + fn drive_unindexed<C>(self, consumer: C) -> C::Result + where + C: UnindexedConsumer<Self::Item>, + { + let Self { set1, set2 } = self; + + set1.par_iter() + .chain(set2.par_difference(set1)) + .drive_unindexed(consumer) + } +} + +/// Parallel sorting methods. +/// +/// The following methods **require crate feature `"rayon"`**. +impl<T, S> IndexSet<T, S> +where + T: Hash + Eq + Send, + S: BuildHasher + Send, +{ + /// Sort the set’s values in parallel by their default ordering. + pub fn par_sort(&mut self) + where + T: Ord, + { + self.with_entries(|entries| { + entries.par_sort_by(|a, b| T::cmp(&a.key, &b.key)); + }); + } + + /// Sort the set’s values in place and in parallel, using the comparison function `compare`. + pub fn par_sort_by<F>(&mut self, cmp: F) + where + F: Fn(&T, &T) -> Ordering + Sync, + { + self.with_entries(|entries| { + entries.par_sort_by(move |a, b| cmp(&a.key, &b.key)); + }); + } + + /// Sort the values of the set in parallel and return a by value parallel iterator of + /// the values with the result. + pub fn par_sorted_by<F>(self, cmp: F) -> IntoParIter<T> + where + F: Fn(&T, &T) -> Ordering + Sync, + { + let mut entries = self.into_entries(); + entries.par_sort_by(move |a, b| cmp(&a.key, &b.key)); + IntoParIter { entries } + } +} + +/// Requires crate feature `"rayon"`. +impl<T, S> FromParallelIterator<T> for IndexSet<T, S> +where + T: Eq + Hash + Send, + S: BuildHasher + Default + Send, +{ + fn from_par_iter<I>(iter: I) -> Self + where + I: IntoParallelIterator<Item = T>, + { + let list = collect(iter); + let len = list.iter().map(Vec::len).sum(); + let mut set = Self::with_capacity_and_hasher(len, S::default()); + for vec in list { + set.extend(vec); + } + set + } +} + +/// Requires crate feature `"rayon"`. +impl<T, S> ParallelExtend<T> for IndexSet<T, S> +where + T: Eq + Hash + Send, + S: BuildHasher + Send, +{ + fn par_extend<I>(&mut self, iter: I) + where + I: IntoParallelIterator<Item = T>, + { + for vec in collect(iter) { + self.extend(vec); + } + } +} + +/// Requires crate feature `"rayon"`. +impl<'a, T: 'a, S> ParallelExtend<&'a T> for IndexSet<T, S> +where + T: Copy + Eq + Hash + Send + Sync, + S: BuildHasher + Send, +{ + fn par_extend<I>(&mut self, iter: I) + where + I: IntoParallelIterator<Item = &'a T>, + { + for vec in collect(iter) { + self.extend(vec); + } + } +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn insert_order() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut set = IndexSet::new(); + + for &elt in &insert { + set.insert(elt); + } + + assert_eq!(set.par_iter().count(), set.len()); + assert_eq!(set.par_iter().count(), insert.len()); + insert.par_iter().zip(&set).for_each(|(a, b)| { + assert_eq!(a, b); + }); + (0..insert.len()) + .into_par_iter() + .zip(&set) + .for_each(|(i, v)| { + assert_eq!(set.get_index(i).unwrap(), v); + }); + } + + #[test] + fn partial_eq_and_eq() { + let mut set_a = IndexSet::new(); + set_a.insert(1); + set_a.insert(2); + let mut set_b = set_a.clone(); + assert!(set_a.par_eq(&set_b)); + set_b.swap_remove(&1); + assert!(!set_a.par_eq(&set_b)); + set_b.insert(3); + assert!(!set_a.par_eq(&set_b)); + + let set_c: IndexSet<_> = set_b.into_par_iter().collect(); + assert!(!set_a.par_eq(&set_c)); + assert!(!set_c.par_eq(&set_a)); + } + + #[test] + fn extend() { + let mut set = IndexSet::new(); + set.par_extend(vec![&1, &2, &3, &4]); + set.par_extend(vec![5, 6]); + assert_eq!( + set.into_par_iter().collect::<Vec<_>>(), + vec![1, 2, 3, 4, 5, 6] + ); + } + + #[test] + fn comparisons() { + let set_a: IndexSet<_> = (0..3).collect(); + let set_b: IndexSet<_> = (3..6).collect(); + let set_c: IndexSet<_> = (0..6).collect(); + let set_d: IndexSet<_> = (3..9).collect(); + + assert!(!set_a.par_is_disjoint(&set_a)); + assert!(set_a.par_is_subset(&set_a)); + assert!(set_a.par_is_superset(&set_a)); + + assert!(set_a.par_is_disjoint(&set_b)); + assert!(set_b.par_is_disjoint(&set_a)); + assert!(!set_a.par_is_subset(&set_b)); + assert!(!set_b.par_is_subset(&set_a)); + assert!(!set_a.par_is_superset(&set_b)); + assert!(!set_b.par_is_superset(&set_a)); + + assert!(!set_a.par_is_disjoint(&set_c)); + assert!(!set_c.par_is_disjoint(&set_a)); + assert!(set_a.par_is_subset(&set_c)); + assert!(!set_c.par_is_subset(&set_a)); + assert!(!set_a.par_is_superset(&set_c)); + assert!(set_c.par_is_superset(&set_a)); + + assert!(!set_c.par_is_disjoint(&set_d)); + assert!(!set_d.par_is_disjoint(&set_c)); + assert!(!set_c.par_is_subset(&set_d)); + assert!(!set_d.par_is_subset(&set_c)); + assert!(!set_c.par_is_superset(&set_d)); + assert!(!set_d.par_is_superset(&set_c)); + } + + #[test] + fn iter_comparisons() { + use std::iter::empty; + + fn check<'a, I1, I2>(iter1: I1, iter2: I2) + where + I1: ParallelIterator<Item = &'a i32>, + I2: Iterator<Item = i32>, + { + let v1: Vec<_> = iter1.cloned().collect(); + let v2: Vec<_> = iter2.collect(); + assert_eq!(v1, v2); + } + + let set_a: IndexSet<_> = (0..3).collect(); + let set_b: IndexSet<_> = (3..6).collect(); + let set_c: IndexSet<_> = (0..6).collect(); + let set_d: IndexSet<_> = (3..9).rev().collect(); + + check(set_a.par_difference(&set_a), empty()); + check(set_a.par_symmetric_difference(&set_a), empty()); + check(set_a.par_intersection(&set_a), 0..3); + check(set_a.par_union(&set_a), 0..3); + + check(set_a.par_difference(&set_b), 0..3); + check(set_b.par_difference(&set_a), 3..6); + check(set_a.par_symmetric_difference(&set_b), 0..6); + check(set_b.par_symmetric_difference(&set_a), (3..6).chain(0..3)); + check(set_a.par_intersection(&set_b), empty()); + check(set_b.par_intersection(&set_a), empty()); + check(set_a.par_union(&set_b), 0..6); + check(set_b.par_union(&set_a), (3..6).chain(0..3)); + + check(set_a.par_difference(&set_c), empty()); + check(set_c.par_difference(&set_a), 3..6); + check(set_a.par_symmetric_difference(&set_c), 3..6); + check(set_c.par_symmetric_difference(&set_a), 3..6); + check(set_a.par_intersection(&set_c), 0..3); + check(set_c.par_intersection(&set_a), 0..3); + check(set_a.par_union(&set_c), 0..6); + check(set_c.par_union(&set_a), 0..6); + + check(set_c.par_difference(&set_d), 0..3); + check(set_d.par_difference(&set_c), (6..9).rev()); + check( + set_c.par_symmetric_difference(&set_d), + (0..3).chain((6..9).rev()), + ); + check( + set_d.par_symmetric_difference(&set_c), + (6..9).rev().chain(0..3), + ); + check(set_c.par_intersection(&set_d), 3..6); + check(set_d.par_intersection(&set_c), (3..6).rev()); + check(set_c.par_union(&set_d), (0..6).chain((6..9).rev())); + check(set_d.par_union(&set_c), (3..9).rev().chain(0..3)); + } +} diff --git a/third_party/rust/indexmap/src/serde.rs b/third_party/rust/indexmap/src/serde.rs new file mode 100644 index 0000000000..853c6b9aa3 --- /dev/null +++ b/third_party/rust/indexmap/src/serde.rs @@ -0,0 +1,163 @@ +use serde::de::value::{MapDeserializer, SeqDeserializer}; +use serde::de::{ + Deserialize, Deserializer, Error, IntoDeserializer, MapAccess, SeqAccess, Visitor, +}; +use serde::ser::{Serialize, SerializeMap, SerializeSeq, Serializer}; + +use core::fmt::{self, Formatter}; +use core::hash::{BuildHasher, Hash}; +use core::marker::PhantomData; + +use crate::IndexMap; + +/// Requires crate feature `"serde-1"` +impl<K, V, S> Serialize for IndexMap<K, V, S> +where + K: Serialize + Hash + Eq, + V: Serialize, + S: BuildHasher, +{ + fn serialize<T>(&self, serializer: T) -> Result<T::Ok, T::Error> + where + T: Serializer, + { + let mut map_serializer = serializer.serialize_map(Some(self.len()))?; + for (key, value) in self { + map_serializer.serialize_entry(key, value)?; + } + map_serializer.end() + } +} + +struct IndexMapVisitor<K, V, S>(PhantomData<(K, V, S)>); + +impl<'de, K, V, S> Visitor<'de> for IndexMapVisitor<K, V, S> +where + K: Deserialize<'de> + Eq + Hash, + V: Deserialize<'de>, + S: Default + BuildHasher, +{ + type Value = IndexMap<K, V, S>; + + fn expecting(&self, formatter: &mut Formatter<'_>) -> fmt::Result { + write!(formatter, "a map") + } + + fn visit_map<A>(self, mut map: A) -> Result<Self::Value, A::Error> + where + A: MapAccess<'de>, + { + let mut values = + IndexMap::with_capacity_and_hasher(map.size_hint().unwrap_or(0), S::default()); + + while let Some((key, value)) = map.next_entry()? { + values.insert(key, value); + } + + Ok(values) + } +} + +/// Requires crate feature `"serde-1"` +impl<'de, K, V, S> Deserialize<'de> for IndexMap<K, V, S> +where + K: Deserialize<'de> + Eq + Hash, + V: Deserialize<'de>, + S: Default + BuildHasher, +{ + fn deserialize<D>(deserializer: D) -> Result<Self, D::Error> + where + D: Deserializer<'de>, + { + deserializer.deserialize_map(IndexMapVisitor(PhantomData)) + } +} + +impl<'de, K, V, S, E> IntoDeserializer<'de, E> for IndexMap<K, V, S> +where + K: IntoDeserializer<'de, E> + Eq + Hash, + V: IntoDeserializer<'de, E>, + S: BuildHasher, + E: Error, +{ + type Deserializer = MapDeserializer<'de, <Self as IntoIterator>::IntoIter, E>; + + fn into_deserializer(self) -> Self::Deserializer { + MapDeserializer::new(self.into_iter()) + } +} + +use crate::IndexSet; + +/// Requires crate feature `"serde-1"` +impl<T, S> Serialize for IndexSet<T, S> +where + T: Serialize + Hash + Eq, + S: BuildHasher, +{ + fn serialize<Se>(&self, serializer: Se) -> Result<Se::Ok, Se::Error> + where + Se: Serializer, + { + let mut set_serializer = serializer.serialize_seq(Some(self.len()))?; + for value in self { + set_serializer.serialize_element(value)?; + } + set_serializer.end() + } +} + +struct IndexSetVisitor<T, S>(PhantomData<(T, S)>); + +impl<'de, T, S> Visitor<'de> for IndexSetVisitor<T, S> +where + T: Deserialize<'de> + Eq + Hash, + S: Default + BuildHasher, +{ + type Value = IndexSet<T, S>; + + fn expecting(&self, formatter: &mut Formatter<'_>) -> fmt::Result { + write!(formatter, "a set") + } + + fn visit_seq<A>(self, mut seq: A) -> Result<Self::Value, A::Error> + where + A: SeqAccess<'de>, + { + let mut values = + IndexSet::with_capacity_and_hasher(seq.size_hint().unwrap_or(0), S::default()); + + while let Some(value) = seq.next_element()? { + values.insert(value); + } + + Ok(values) + } +} + +/// Requires crate feature `"serde-1"` +impl<'de, T, S> Deserialize<'de> for IndexSet<T, S> +where + T: Deserialize<'de> + Eq + Hash, + S: Default + BuildHasher, +{ + fn deserialize<D>(deserializer: D) -> Result<Self, D::Error> + where + D: Deserializer<'de>, + { + deserializer.deserialize_seq(IndexSetVisitor(PhantomData)) + } +} + +impl<'de, T, S, E> IntoDeserializer<'de, E> for IndexSet<T, S> +where + T: IntoDeserializer<'de, E> + Eq + Hash, + S: BuildHasher, + E: Error, +{ + type Deserializer = SeqDeserializer<<Self as IntoIterator>::IntoIter, E>; + + fn into_deserializer(self) -> Self::Deserializer { + SeqDeserializer::new(self.into_iter()) + } +} diff --git a/third_party/rust/indexmap/src/set.rs b/third_party/rust/indexmap/src/set.rs new file mode 100644 index 0000000000..9949389187 --- /dev/null +++ b/third_party/rust/indexmap/src/set.rs @@ -0,0 +1,1607 @@ +//! A hash set implemented using `IndexMap` + +#[cfg(feature = "rayon")] +pub use crate::rayon::set as rayon; + +#[cfg(has_std)] +use std::collections::hash_map::RandomState; + +use crate::vec::{self, Vec}; +use core::cmp::Ordering; +use core::fmt; +use core::hash::{BuildHasher, Hash}; +use core::iter::{Chain, FromIterator}; +use core::ops::{BitAnd, BitOr, BitXor, Index, RangeBounds, Sub}; +use core::slice; + +use super::{Entries, Equivalent, IndexMap}; + +type Bucket<T> = super::Bucket<T, ()>; + +/// A hash set where the iteration order of the values is independent of their +/// hash values. +/// +/// The interface is closely compatible with the standard `HashSet`, but also +/// has additional features. +/// +/// # Order +/// +/// The values have a consistent order that is determined by the sequence of +/// insertion and removal calls on the set. The order does not depend on the +/// values or the hash function at all. Note that insertion order and value +/// are not affected if a re-insertion is attempted once an element is +/// already present. +/// +/// All iterators traverse the set *in order*. Set operation iterators like +/// `union` produce a concatenated order, as do their matching "bitwise" +/// operators. See their documentation for specifics. +/// +/// The insertion order is preserved, with **notable exceptions** like the +/// `.remove()` or `.swap_remove()` methods. Methods such as `.sort_by()` of +/// course result in a new order, depending on the sorting order. +/// +/// # Indices +/// +/// The values are indexed in a compact range without holes in the range +/// `0..self.len()`. For example, the method `.get_full` looks up the index for +/// a value, and the method `.get_index` looks up the value by index. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexSet; +/// +/// // Collects which letters appear in a sentence. +/// let letters: IndexSet<_> = "a short treatise on fungi".chars().collect(); +/// +/// assert!(letters.contains(&'s')); +/// assert!(letters.contains(&'t')); +/// assert!(letters.contains(&'u')); +/// assert!(!letters.contains(&'y')); +/// ``` +#[cfg(has_std)] +pub struct IndexSet<T, S = RandomState> { + map: IndexMap<T, (), S>, +} +#[cfg(not(has_std))] +pub struct IndexSet<T, S> { + map: IndexMap<T, (), S>, +} + +impl<T, S> Clone for IndexSet<T, S> +where + T: Clone, + S: Clone, +{ + fn clone(&self) -> Self { + IndexSet { + map: self.map.clone(), + } + } + + fn clone_from(&mut self, other: &Self) { + self.map.clone_from(&other.map); + } +} + +impl<T, S> Entries for IndexSet<T, S> { + type Entry = Bucket<T>; + + #[inline] + fn into_entries(self) -> Vec<Self::Entry> { + self.map.into_entries() + } + + #[inline] + fn as_entries(&self) -> &[Self::Entry] { + self.map.as_entries() + } + + #[inline] + fn as_entries_mut(&mut self) -> &mut [Self::Entry] { + self.map.as_entries_mut() + } + + fn with_entries<F>(&mut self, f: F) + where + F: FnOnce(&mut [Self::Entry]), + { + self.map.with_entries(f); + } +} + +impl<T, S> fmt::Debug for IndexSet<T, S> +where + T: fmt::Debug, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + if cfg!(not(feature = "test_debug")) { + f.debug_set().entries(self.iter()).finish() + } else { + // Let the inner `IndexMap` print all of its details + f.debug_struct("IndexSet").field("map", &self.map).finish() + } + } +} + +#[cfg(has_std)] +impl<T> IndexSet<T> { + /// Create a new set. (Does not allocate.) + pub fn new() -> Self { + IndexSet { + map: IndexMap::new(), + } + } + + /// Create a new set with capacity for `n` elements. + /// (Does not allocate if `n` is zero.) + /// + /// Computes in **O(n)** time. + pub fn with_capacity(n: usize) -> Self { + IndexSet { + map: IndexMap::with_capacity(n), + } + } +} + +impl<T, S> IndexSet<T, S> { + /// Create a new set with capacity for `n` elements. + /// (Does not allocate if `n` is zero.) + /// + /// Computes in **O(n)** time. + pub fn with_capacity_and_hasher(n: usize, hash_builder: S) -> Self { + IndexSet { + map: IndexMap::with_capacity_and_hasher(n, hash_builder), + } + } + + /// Create a new set with `hash_builder` + pub fn with_hasher(hash_builder: S) -> Self { + IndexSet { + map: IndexMap::with_hasher(hash_builder), + } + } + + /// Computes in **O(1)** time. + pub fn capacity(&self) -> usize { + self.map.capacity() + } + + /// Return a reference to the set's `BuildHasher`. + pub fn hasher(&self) -> &S { + self.map.hasher() + } + + /// Return the number of elements in the set. + /// + /// Computes in **O(1)** time. + pub fn len(&self) -> usize { + self.map.len() + } + + /// Returns true if the set contains no elements. + /// + /// Computes in **O(1)** time. + pub fn is_empty(&self) -> bool { + self.map.is_empty() + } + + /// Return an iterator over the values of the set, in their order + pub fn iter(&self) -> Iter<'_, T> { + Iter { + iter: self.map.keys().iter, + } + } + + /// Remove all elements in the set, while preserving its capacity. + /// + /// Computes in **O(n)** time. + pub fn clear(&mut self) { + self.map.clear(); + } + + /// Clears the `IndexSet` in the given index range, returning those values + /// as a drain iterator. + /// + /// The range may be any type that implements `RangeBounds<usize>`, + /// including all of the `std::ops::Range*` types, or even a tuple pair of + /// `Bound` start and end values. To drain the set entirely, use `RangeFull` + /// like `set.drain(..)`. + /// + /// This shifts down all entries following the drained range to fill the + /// gap, and keeps the allocated memory for reuse. + /// + /// ***Panics*** if the starting point is greater than the end point or if + /// the end point is greater than the length of the set. + pub fn drain<R>(&mut self, range: R) -> Drain<'_, T> + where + R: RangeBounds<usize>, + { + Drain { + iter: self.map.drain(range).iter, + } + } +} + +impl<T, S> IndexSet<T, S> +where + T: Hash + Eq, + S: BuildHasher, +{ + /// Reserve capacity for `additional` more values. + /// + /// Computes in **O(n)** time. + pub fn reserve(&mut self, additional: usize) { + self.map.reserve(additional); + } + + /// Shrink the capacity of the set as much as possible. + /// + /// Computes in **O(n)** time. + pub fn shrink_to_fit(&mut self) { + self.map.shrink_to_fit(); + } + + /// Insert the value into the set. + /// + /// If an equivalent item already exists in the set, it returns + /// `false` leaving the original value in the set and without + /// altering its insertion order. Otherwise, it inserts the new + /// item and returns `true`. + /// + /// Computes in **O(1)** time (amortized average). + pub fn insert(&mut self, value: T) -> bool { + self.map.insert(value, ()).is_none() + } + + /// Insert the value into the set, and get its index. + /// + /// If an equivalent item already exists in the set, it returns + /// the index of the existing item and `false`, leaving the + /// original value in the set and without altering its insertion + /// order. Otherwise, it inserts the new item and returns the index + /// of the inserted item and `true`. + /// + /// Computes in **O(1)** time (amortized average). + pub fn insert_full(&mut self, value: T) -> (usize, bool) { + use super::map::Entry::*; + + match self.map.entry(value) { + Occupied(e) => (e.index(), false), + Vacant(e) => { + let index = e.index(); + e.insert(()); + (index, true) + } + } + } + + /// Return an iterator over the values that are in `self` but not `other`. + /// + /// Values are produced in the same order that they appear in `self`. + pub fn difference<'a, S2>(&'a self, other: &'a IndexSet<T, S2>) -> Difference<'a, T, S2> + where + S2: BuildHasher, + { + Difference { + iter: self.iter(), + other, + } + } + + /// Return an iterator over the values that are in `self` or `other`, + /// but not in both. + /// + /// Values from `self` are produced in their original order, followed by + /// values from `other` in their original order. + pub fn symmetric_difference<'a, S2>( + &'a self, + other: &'a IndexSet<T, S2>, + ) -> SymmetricDifference<'a, T, S, S2> + where + S2: BuildHasher, + { + SymmetricDifference { + iter: self.difference(other).chain(other.difference(self)), + } + } + + /// Return an iterator over the values that are in both `self` and `other`. + /// + /// Values are produced in the same order that they appear in `self`. + pub fn intersection<'a, S2>(&'a self, other: &'a IndexSet<T, S2>) -> Intersection<'a, T, S2> + where + S2: BuildHasher, + { + Intersection { + iter: self.iter(), + other, + } + } + + /// Return an iterator over all values that are in `self` or `other`. + /// + /// Values from `self` are produced in their original order, followed by + /// values that are unique to `other` in their original order. + pub fn union<'a, S2>(&'a self, other: &'a IndexSet<T, S2>) -> Union<'a, T, S> + where + S2: BuildHasher, + { + Union { + iter: self.iter().chain(other.difference(self)), + } + } + + /// Return `true` if an equivalent to `value` exists in the set. + /// + /// Computes in **O(1)** time (average). + pub fn contains<Q: ?Sized>(&self, value: &Q) -> bool + where + Q: Hash + Equivalent<T>, + { + self.map.contains_key(value) + } + + /// Return a reference to the value stored in the set, if it is present, + /// else `None`. + /// + /// Computes in **O(1)** time (average). + pub fn get<Q: ?Sized>(&self, value: &Q) -> Option<&T> + where + Q: Hash + Equivalent<T>, + { + self.map.get_key_value(value).map(|(x, &())| x) + } + + /// Return item index and value + pub fn get_full<Q: ?Sized>(&self, value: &Q) -> Option<(usize, &T)> + where + Q: Hash + Equivalent<T>, + { + self.map.get_full(value).map(|(i, x, &())| (i, x)) + } + + /// Return item index, if it exists in the set + pub fn get_index_of<Q: ?Sized>(&self, value: &Q) -> Option<usize> + where + Q: Hash + Equivalent<T>, + { + self.map.get_index_of(value) + } + + /// Adds a value to the set, replacing the existing value, if any, that is + /// equal to the given one. Returns the replaced value. + /// + /// Computes in **O(1)** time (average). + pub fn replace(&mut self, value: T) -> Option<T> { + use super::map::Entry::*; + + match self.map.entry(value) { + Vacant(e) => { + e.insert(()); + None + } + Occupied(e) => Some(e.replace_key()), + } + } + + /// Remove the value from the set, and return `true` if it was present. + /// + /// **NOTE:** This is equivalent to `.swap_remove(value)`, if you want + /// to preserve the order of the values in the set, use `.shift_remove(value)`. + /// + /// Computes in **O(1)** time (average). + pub fn remove<Q: ?Sized>(&mut self, value: &Q) -> bool + where + Q: Hash + Equivalent<T>, + { + self.swap_remove(value) + } + + /// Remove the value from the set, and return `true` if it was present. + /// + /// Like `Vec::swap_remove`, the value is removed by swapping it with the + /// last element of the set and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Return `false` if `value` was not in the set. + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove<Q: ?Sized>(&mut self, value: &Q) -> bool + where + Q: Hash + Equivalent<T>, + { + self.map.swap_remove(value).is_some() + } + + /// Remove the value from the set, and return `true` if it was present. + /// + /// Like `Vec::remove`, the value is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Return `false` if `value` was not in the set. + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove<Q: ?Sized>(&mut self, value: &Q) -> bool + where + Q: Hash + Equivalent<T>, + { + self.map.shift_remove(value).is_some() + } + + /// Removes and returns the value in the set, if any, that is equal to the + /// given one. + /// + /// **NOTE:** This is equivalent to `.swap_take(value)`, if you need to + /// preserve the order of the values in the set, use `.shift_take(value)` + /// instead. + /// + /// Computes in **O(1)** time (average). + pub fn take<Q: ?Sized>(&mut self, value: &Q) -> Option<T> + where + Q: Hash + Equivalent<T>, + { + self.swap_take(value) + } + + /// Removes and returns the value in the set, if any, that is equal to the + /// given one. + /// + /// Like `Vec::swap_remove`, the value is removed by swapping it with the + /// last element of the set and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Return `None` if `value` was not in the set. + /// + /// Computes in **O(1)** time (average). + pub fn swap_take<Q: ?Sized>(&mut self, value: &Q) -> Option<T> + where + Q: Hash + Equivalent<T>, + { + self.map.swap_remove_entry(value).map(|(x, ())| x) + } + + /// Removes and returns the value in the set, if any, that is equal to the + /// given one. + /// + /// Like `Vec::remove`, the value is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Return `None` if `value` was not in the set. + /// + /// Computes in **O(n)** time (average). + pub fn shift_take<Q: ?Sized>(&mut self, value: &Q) -> Option<T> + where + Q: Hash + Equivalent<T>, + { + self.map.shift_remove_entry(value).map(|(x, ())| x) + } + + /// Remove the value from the set return it and the index it had. + /// + /// Like `Vec::swap_remove`, the value is removed by swapping it with the + /// last element of the set and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Return `None` if `value` was not in the set. + pub fn swap_remove_full<Q: ?Sized>(&mut self, value: &Q) -> Option<(usize, T)> + where + Q: Hash + Equivalent<T>, + { + self.map.swap_remove_full(value).map(|(i, x, ())| (i, x)) + } + + /// Remove the value from the set return it and the index it had. + /// + /// Like `Vec::remove`, the value is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Return `None` if `value` was not in the set. + pub fn shift_remove_full<Q: ?Sized>(&mut self, value: &Q) -> Option<(usize, T)> + where + Q: Hash + Equivalent<T>, + { + self.map.shift_remove_full(value).map(|(i, x, ())| (i, x)) + } + + /// Remove the last value + /// + /// Computes in **O(1)** time (average). + pub fn pop(&mut self) -> Option<T> { + self.map.pop().map(|(x, ())| x) + } + + /// Scan through each value in the set and keep those where the + /// closure `keep` returns `true`. + /// + /// The elements are visited in order, and remaining elements keep their + /// order. + /// + /// Computes in **O(n)** time (average). + pub fn retain<F>(&mut self, mut keep: F) + where + F: FnMut(&T) -> bool, + { + self.map.retain(move |x, &mut ()| keep(x)) + } + + /// Sort the set’s values by their default ordering. + /// + /// See `sort_by` for details. + pub fn sort(&mut self) + where + T: Ord, + { + self.map.sort_keys() + } + + /// Sort the set’s values in place using the comparison function `compare`. + /// + /// Computes in **O(n log n)** time and **O(n)** space. The sort is stable. + pub fn sort_by<F>(&mut self, mut compare: F) + where + F: FnMut(&T, &T) -> Ordering, + { + self.map.sort_by(move |a, _, b, _| compare(a, b)); + } + + /// Sort the values of the set and return a by value iterator of + /// the values with the result. + /// + /// The sort is stable. + pub fn sorted_by<F>(self, mut cmp: F) -> IntoIter<T> + where + F: FnMut(&T, &T) -> Ordering, + { + IntoIter { + iter: self.map.sorted_by(move |a, &(), b, &()| cmp(a, b)).iter, + } + } + + /// Reverses the order of the set’s values in place. + /// + /// Computes in **O(n)** time and **O(1)** space. + pub fn reverse(&mut self) { + self.map.reverse() + } +} + +impl<T, S> IndexSet<T, S> { + /// Get a value by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Computes in **O(1)** time. + pub fn get_index(&self, index: usize) -> Option<&T> { + self.map.get_index(index).map(|(x, &())| x) + } + + /// Remove the key-value pair by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Like `Vec::swap_remove`, the value is removed by swapping it with the + /// last element of the set and popping it off. **This perturbs + /// the postion of what used to be the last element!** + /// + /// Computes in **O(1)** time (average). + pub fn swap_remove_index(&mut self, index: usize) -> Option<T> { + self.map.swap_remove_index(index).map(|(x, ())| x) + } + + /// Remove the key-value pair by index + /// + /// Valid indices are *0 <= index < self.len()* + /// + /// Like `Vec::remove`, the value is removed by shifting all of the + /// elements that follow it, preserving their relative order. + /// **This perturbs the index of all of those elements!** + /// + /// Computes in **O(n)** time (average). + pub fn shift_remove_index(&mut self, index: usize) -> Option<T> { + self.map.shift_remove_index(index).map(|(x, ())| x) + } +} + +/// Access `IndexSet` values at indexed positions. +/// +/// # Examples +/// +/// ``` +/// use indexmap::IndexSet; +/// +/// let mut set = IndexSet::new(); +/// for word in "Lorem ipsum dolor sit amet".split_whitespace() { +/// set.insert(word.to_string()); +/// } +/// assert_eq!(set[0], "Lorem"); +/// assert_eq!(set[1], "ipsum"); +/// set.reverse(); +/// assert_eq!(set[0], "amet"); +/// assert_eq!(set[1], "sit"); +/// set.sort(); +/// assert_eq!(set[0], "Lorem"); +/// assert_eq!(set[1], "amet"); +/// ``` +/// +/// ```should_panic +/// use indexmap::IndexSet; +/// +/// let mut set = IndexSet::new(); +/// set.insert("foo"); +/// println!("{:?}", set[10]); // panics! +/// ``` +impl<T, S> Index<usize> for IndexSet<T, S> { + type Output = T; + + /// Returns a reference to the value at the supplied `index`. + /// + /// ***Panics*** if `index` is out of bounds. + fn index(&self, index: usize) -> &T { + self.get_index(index) + .expect("IndexSet: index out of bounds") + } +} + +/// An owning iterator over the items of a `IndexSet`. +/// +/// This `struct` is created by the [`into_iter`] method on [`IndexSet`] +/// (provided by the `IntoIterator` trait). See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`into_iter`]: struct.IndexSet.html#method.into_iter +pub struct IntoIter<T> { + iter: vec::IntoIter<Bucket<T>>, +} + +impl<T> Iterator for IntoIter<T> { + type Item = T; + + iterator_methods!(Bucket::key); +} + +impl<T> DoubleEndedIterator for IntoIter<T> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::key) + } +} + +impl<T> ExactSizeIterator for IntoIter<T> { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl<T: fmt::Debug> fmt::Debug for IntoIter<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let iter = self.iter.as_slice().iter().map(Bucket::key_ref); + f.debug_list().entries(iter).finish() + } +} + +/// An iterator over the items of a `IndexSet`. +/// +/// This `struct` is created by the [`iter`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`iter`]: struct.IndexSet.html#method.iter +pub struct Iter<'a, T> { + iter: slice::Iter<'a, Bucket<T>>, +} + +impl<'a, T> Iterator for Iter<'a, T> { + type Item = &'a T; + + iterator_methods!(Bucket::key_ref); +} + +impl<T> DoubleEndedIterator for Iter<'_, T> { + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back().map(Bucket::key_ref) + } +} + +impl<T> ExactSizeIterator for Iter<'_, T> { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl<T> Clone for Iter<'_, T> { + fn clone(&self) -> Self { + Iter { + iter: self.iter.clone(), + } + } +} + +impl<T: fmt::Debug> fmt::Debug for Iter<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A draining iterator over the items of a `IndexSet`. +/// +/// This `struct` is created by the [`drain`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`drain`]: struct.IndexSet.html#method.drain +pub struct Drain<'a, T> { + iter: vec::Drain<'a, Bucket<T>>, +} + +impl<T> Iterator for Drain<'_, T> { + type Item = T; + + iterator_methods!(Bucket::key); +} + +impl<T> DoubleEndedIterator for Drain<'_, T> { + double_ended_iterator_methods!(Bucket::key); +} + +impl<'a, T, S> IntoIterator for &'a IndexSet<T, S> { + type Item = &'a T; + type IntoIter = Iter<'a, T>; + + fn into_iter(self) -> Self::IntoIter { + self.iter() + } +} + +impl<T, S> IntoIterator for IndexSet<T, S> { + type Item = T; + type IntoIter = IntoIter<T>; + + fn into_iter(self) -> Self::IntoIter { + IntoIter { + iter: self.map.into_iter().iter, + } + } +} + +impl<T, S> FromIterator<T> for IndexSet<T, S> +where + T: Hash + Eq, + S: BuildHasher + Default, +{ + fn from_iter<I: IntoIterator<Item = T>>(iterable: I) -> Self { + let iter = iterable.into_iter().map(|x| (x, ())); + IndexSet { + map: IndexMap::from_iter(iter), + } + } +} + +impl<T, S> Extend<T> for IndexSet<T, S> +where + T: Hash + Eq, + S: BuildHasher, +{ + fn extend<I: IntoIterator<Item = T>>(&mut self, iterable: I) { + let iter = iterable.into_iter().map(|x| (x, ())); + self.map.extend(iter); + } +} + +impl<'a, T, S> Extend<&'a T> for IndexSet<T, S> +where + T: Hash + Eq + Copy + 'a, + S: BuildHasher, +{ + fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iterable: I) { + let iter = iterable.into_iter().cloned(); // FIXME: use `copied` in Rust 1.36 + self.extend(iter); + } +} + +impl<T, S> Default for IndexSet<T, S> +where + S: Default, +{ + /// Return an empty `IndexSet` + fn default() -> Self { + IndexSet { + map: IndexMap::default(), + } + } +} + +impl<T, S1, S2> PartialEq<IndexSet<T, S2>> for IndexSet<T, S1> +where + T: Hash + Eq, + S1: BuildHasher, + S2: BuildHasher, +{ + fn eq(&self, other: &IndexSet<T, S2>) -> bool { + self.len() == other.len() && self.is_subset(other) + } +} + +impl<T, S> Eq for IndexSet<T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ +} + +impl<T, S> IndexSet<T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + /// Returns `true` if `self` has no elements in common with `other`. + pub fn is_disjoint<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher, + { + if self.len() <= other.len() { + self.iter().all(move |value| !other.contains(value)) + } else { + other.iter().all(move |value| !self.contains(value)) + } + } + + /// Returns `true` if all elements of `self` are contained in `other`. + pub fn is_subset<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher, + { + self.len() <= other.len() && self.iter().all(move |value| other.contains(value)) + } + + /// Returns `true` if all elements of `other` are contained in `self`. + pub fn is_superset<S2>(&self, other: &IndexSet<T, S2>) -> bool + where + S2: BuildHasher, + { + other.is_subset(self) + } +} + +/// A lazy iterator producing elements in the difference of `IndexSet`s. +/// +/// This `struct` is created by the [`difference`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`difference`]: struct.IndexSet.html#method.difference +pub struct Difference<'a, T, S> { + iter: Iter<'a, T>, + other: &'a IndexSet<T, S>, +} + +impl<'a, T, S> Iterator for Difference<'a, T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + type Item = &'a T; + + fn next(&mut self) -> Option<Self::Item> { + while let Some(item) = self.iter.next() { + if !self.other.contains(item) { + return Some(item); + } + } + None + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (0, self.iter.size_hint().1) + } +} + +impl<T, S> DoubleEndedIterator for Difference<'_, T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + fn next_back(&mut self) -> Option<Self::Item> { + while let Some(item) = self.iter.next_back() { + if !self.other.contains(item) { + return Some(item); + } + } + None + } +} + +impl<T, S> Clone for Difference<'_, T, S> { + fn clone(&self) -> Self { + Difference { + iter: self.iter.clone(), + ..*self + } + } +} + +impl<T, S> fmt::Debug for Difference<'_, T, S> +where + T: fmt::Debug + Eq + Hash, + S: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A lazy iterator producing elements in the intersection of `IndexSet`s. +/// +/// This `struct` is created by the [`intersection`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`intersection`]: struct.IndexSet.html#method.intersection +pub struct Intersection<'a, T, S> { + iter: Iter<'a, T>, + other: &'a IndexSet<T, S>, +} + +impl<'a, T, S> Iterator for Intersection<'a, T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + type Item = &'a T; + + fn next(&mut self) -> Option<Self::Item> { + while let Some(item) = self.iter.next() { + if self.other.contains(item) { + return Some(item); + } + } + None + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (0, self.iter.size_hint().1) + } +} + +impl<T, S> DoubleEndedIterator for Intersection<'_, T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + fn next_back(&mut self) -> Option<Self::Item> { + while let Some(item) = self.iter.next_back() { + if self.other.contains(item) { + return Some(item); + } + } + None + } +} + +impl<T, S> Clone for Intersection<'_, T, S> { + fn clone(&self) -> Self { + Intersection { + iter: self.iter.clone(), + ..*self + } + } +} + +impl<T, S> fmt::Debug for Intersection<'_, T, S> +where + T: fmt::Debug + Eq + Hash, + S: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A lazy iterator producing elements in the symmetric difference of `IndexSet`s. +/// +/// This `struct` is created by the [`symmetric_difference`] method on +/// [`IndexSet`]. See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`symmetric_difference`]: struct.IndexSet.html#method.symmetric_difference +pub struct SymmetricDifference<'a, T, S1, S2> { + iter: Chain<Difference<'a, T, S2>, Difference<'a, T, S1>>, +} + +impl<'a, T, S1, S2> Iterator for SymmetricDifference<'a, T, S1, S2> +where + T: Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + type Item = &'a T; + + fn next(&mut self) -> Option<Self::Item> { + self.iter.next() + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } + + fn fold<B, F>(self, init: B, f: F) -> B + where + F: FnMut(B, Self::Item) -> B, + { + self.iter.fold(init, f) + } +} + +impl<T, S1, S2> DoubleEndedIterator for SymmetricDifference<'_, T, S1, S2> +where + T: Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back() + } +} + +impl<T, S1, S2> Clone for SymmetricDifference<'_, T, S1, S2> { + fn clone(&self) -> Self { + SymmetricDifference { + iter: self.iter.clone(), + } + } +} + +impl<T, S1, S2> fmt::Debug for SymmetricDifference<'_, T, S1, S2> +where + T: fmt::Debug + Eq + Hash, + S1: BuildHasher, + S2: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A lazy iterator producing elements in the union of `IndexSet`s. +/// +/// This `struct` is created by the [`union`] method on [`IndexSet`]. +/// See its documentation for more. +/// +/// [`IndexSet`]: struct.IndexSet.html +/// [`union`]: struct.IndexSet.html#method.union +pub struct Union<'a, T, S> { + iter: Chain<Iter<'a, T>, Difference<'a, T, S>>, +} + +impl<'a, T, S> Iterator for Union<'a, T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + type Item = &'a T; + + fn next(&mut self) -> Option<Self::Item> { + self.iter.next() + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } + + fn fold<B, F>(self, init: B, f: F) -> B + where + F: FnMut(B, Self::Item) -> B, + { + self.iter.fold(init, f) + } +} + +impl<T, S> DoubleEndedIterator for Union<'_, T, S> +where + T: Eq + Hash, + S: BuildHasher, +{ + fn next_back(&mut self) -> Option<Self::Item> { + self.iter.next_back() + } +} + +impl<T, S> Clone for Union<'_, T, S> { + fn clone(&self) -> Self { + Union { + iter: self.iter.clone(), + } + } +} + +impl<T, S> fmt::Debug for Union<'_, T, S> +where + T: fmt::Debug + Eq + Hash, + S: BuildHasher, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +impl<T, S1, S2> BitAnd<&IndexSet<T, S2>> for &IndexSet<T, S1> +where + T: Eq + Hash + Clone, + S1: BuildHasher + Default, + S2: BuildHasher, +{ + type Output = IndexSet<T, S1>; + + /// Returns the set intersection, cloned into a new set. + /// + /// Values are collected in the same order that they appear in `self`. + fn bitand(self, other: &IndexSet<T, S2>) -> Self::Output { + self.intersection(other).cloned().collect() + } +} + +impl<T, S1, S2> BitOr<&IndexSet<T, S2>> for &IndexSet<T, S1> +where + T: Eq + Hash + Clone, + S1: BuildHasher + Default, + S2: BuildHasher, +{ + type Output = IndexSet<T, S1>; + + /// Returns the set union, cloned into a new set. + /// + /// Values from `self` are collected in their original order, followed by + /// values that are unique to `other` in their original order. + fn bitor(self, other: &IndexSet<T, S2>) -> Self::Output { + self.union(other).cloned().collect() + } +} + +impl<T, S1, S2> BitXor<&IndexSet<T, S2>> for &IndexSet<T, S1> +where + T: Eq + Hash + Clone, + S1: BuildHasher + Default, + S2: BuildHasher, +{ + type Output = IndexSet<T, S1>; + + /// Returns the set symmetric-difference, cloned into a new set. + /// + /// Values from `self` are collected in their original order, followed by + /// values from `other` in their original order. + fn bitxor(self, other: &IndexSet<T, S2>) -> Self::Output { + self.symmetric_difference(other).cloned().collect() + } +} + +impl<T, S1, S2> Sub<&IndexSet<T, S2>> for &IndexSet<T, S1> +where + T: Eq + Hash + Clone, + S1: BuildHasher + Default, + S2: BuildHasher, +{ + type Output = IndexSet<T, S1>; + + /// Returns the set difference, cloned into a new set. + /// + /// Values are collected in the same order that they appear in `self`. + fn sub(self, other: &IndexSet<T, S2>) -> Self::Output { + self.difference(other).cloned().collect() + } +} + +#[cfg(test)] +mod tests { + use super::*; + use crate::util::enumerate; + use std::string::String; + + #[test] + fn it_works() { + let mut set = IndexSet::new(); + assert_eq!(set.is_empty(), true); + set.insert(1); + set.insert(1); + assert_eq!(set.len(), 1); + assert!(set.get(&1).is_some()); + assert_eq!(set.is_empty(), false); + } + + #[test] + fn new() { + let set = IndexSet::<String>::new(); + println!("{:?}", set); + assert_eq!(set.capacity(), 0); + assert_eq!(set.len(), 0); + assert_eq!(set.is_empty(), true); + } + + #[test] + fn insert() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5]; + let not_present = [1, 3, 6, 9, 10]; + let mut set = IndexSet::with_capacity(insert.len()); + + for (i, &elt) in enumerate(&insert) { + assert_eq!(set.len(), i); + set.insert(elt); + assert_eq!(set.len(), i + 1); + assert_eq!(set.get(&elt), Some(&elt)); + } + println!("{:?}", set); + + for &elt in ¬_present { + assert!(set.get(&elt).is_none()); + } + } + + #[test] + fn insert_full() { + let insert = vec![9, 2, 7, 1, 4, 6, 13]; + let present = vec![1, 6, 2]; + let mut set = IndexSet::with_capacity(insert.len()); + + for (i, &elt) in enumerate(&insert) { + assert_eq!(set.len(), i); + let (index, success) = set.insert_full(elt); + assert!(success); + assert_eq!(Some(index), set.get_full(&elt).map(|x| x.0)); + assert_eq!(set.len(), i + 1); + } + + let len = set.len(); + for &elt in &present { + let (index, success) = set.insert_full(elt); + assert!(!success); + assert_eq!(Some(index), set.get_full(&elt).map(|x| x.0)); + assert_eq!(set.len(), len); + } + } + + #[test] + fn insert_2() { + let mut set = IndexSet::with_capacity(16); + + let mut values = vec![]; + values.extend(0..16); + values.extend(128..267); + + for &i in &values { + let old_set = set.clone(); + set.insert(i); + for value in old_set.iter() { + if set.get(value).is_none() { + println!("old_set: {:?}", old_set); + println!("set: {:?}", set); + panic!("did not find {} in set", value); + } + } + } + + for &i in &values { + assert!(set.get(&i).is_some(), "did not find {}", i); + } + } + + #[test] + fn insert_dup() { + let mut elements = vec![0, 2, 4, 6, 8]; + let mut set: IndexSet<u8> = elements.drain(..).collect(); + { + let (i, v) = set.get_full(&0).unwrap(); + assert_eq!(set.len(), 5); + assert_eq!(i, 0); + assert_eq!(*v, 0); + } + { + let inserted = set.insert(0); + let (i, v) = set.get_full(&0).unwrap(); + assert_eq!(set.len(), 5); + assert_eq!(inserted, false); + assert_eq!(i, 0); + assert_eq!(*v, 0); + } + } + + #[test] + fn insert_order() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut set = IndexSet::new(); + + for &elt in &insert { + set.insert(elt); + } + + assert_eq!(set.iter().count(), set.len()); + assert_eq!(set.iter().count(), insert.len()); + for (a, b) in insert.iter().zip(set.iter()) { + assert_eq!(a, b); + } + for (i, v) in (0..insert.len()).zip(set.iter()) { + assert_eq!(set.get_index(i).unwrap(), v); + } + } + + #[test] + fn grow() { + let insert = [0, 4, 2, 12, 8, 7, 11]; + let not_present = [1, 3, 6, 9, 10]; + let mut set = IndexSet::with_capacity(insert.len()); + + for (i, &elt) in enumerate(&insert) { + assert_eq!(set.len(), i); + set.insert(elt); + assert_eq!(set.len(), i + 1); + assert_eq!(set.get(&elt), Some(&elt)); + } + + println!("{:?}", set); + for &elt in &insert { + set.insert(elt * 10); + } + for &elt in &insert { + set.insert(elt * 100); + } + for (i, &elt) in insert.iter().cycle().enumerate().take(100) { + set.insert(elt * 100 + i as i32); + } + println!("{:?}", set); + for &elt in ¬_present { + assert!(set.get(&elt).is_none()); + } + } + + #[test] + fn reserve() { + let mut set = IndexSet::<usize>::new(); + assert_eq!(set.capacity(), 0); + set.reserve(100); + let capacity = set.capacity(); + assert!(capacity >= 100); + for i in 0..capacity { + assert_eq!(set.len(), i); + set.insert(i); + assert_eq!(set.len(), i + 1); + assert_eq!(set.capacity(), capacity); + assert_eq!(set.get(&i), Some(&i)); + } + set.insert(capacity); + assert_eq!(set.len(), capacity + 1); + assert!(set.capacity() > capacity); + assert_eq!(set.get(&capacity), Some(&capacity)); + } + + #[test] + fn shrink_to_fit() { + let mut set = IndexSet::<usize>::new(); + assert_eq!(set.capacity(), 0); + for i in 0..100 { + assert_eq!(set.len(), i); + set.insert(i); + assert_eq!(set.len(), i + 1); + assert!(set.capacity() >= i + 1); + assert_eq!(set.get(&i), Some(&i)); + set.shrink_to_fit(); + assert_eq!(set.len(), i + 1); + assert_eq!(set.capacity(), i + 1); + assert_eq!(set.get(&i), Some(&i)); + } + } + + #[test] + fn remove() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut set = IndexSet::new(); + + for &elt in &insert { + set.insert(elt); + } + + assert_eq!(set.iter().count(), set.len()); + assert_eq!(set.iter().count(), insert.len()); + for (a, b) in insert.iter().zip(set.iter()) { + assert_eq!(a, b); + } + + let remove_fail = [99, 77]; + let remove = [4, 12, 8, 7]; + + for &value in &remove_fail { + assert!(set.swap_remove_full(&value).is_none()); + } + println!("{:?}", set); + for &value in &remove { + //println!("{:?}", set); + let index = set.get_full(&value).unwrap().0; + assert_eq!(set.swap_remove_full(&value), Some((index, value))); + } + println!("{:?}", set); + + for value in &insert { + assert_eq!(set.get(value).is_some(), !remove.contains(value)); + } + assert_eq!(set.len(), insert.len() - remove.len()); + assert_eq!(set.iter().count(), insert.len() - remove.len()); + } + + #[test] + fn swap_remove_index() { + let insert = [0, 4, 2, 12, 8, 7, 11, 5, 3, 17, 19, 22, 23]; + let mut set = IndexSet::new(); + + for &elt in &insert { + set.insert(elt); + } + + let mut vector = insert.to_vec(); + let remove_sequence = &[3, 3, 10, 4, 5, 4, 3, 0, 1]; + + // check that the same swap remove sequence on vec and set + // have the same result. + for &rm in remove_sequence { + let out_vec = vector.swap_remove(rm); + let out_set = set.swap_remove_index(rm).unwrap(); + assert_eq!(out_vec, out_set); + } + assert_eq!(vector.len(), set.len()); + for (a, b) in vector.iter().zip(set.iter()) { + assert_eq!(a, b); + } + } + + #[test] + fn partial_eq_and_eq() { + let mut set_a = IndexSet::new(); + set_a.insert(1); + set_a.insert(2); + let mut set_b = set_a.clone(); + assert_eq!(set_a, set_b); + set_b.swap_remove(&1); + assert_ne!(set_a, set_b); + + let set_c: IndexSet<_> = set_b.into_iter().collect(); + assert_ne!(set_a, set_c); + assert_ne!(set_c, set_a); + } + + #[test] + fn extend() { + let mut set = IndexSet::new(); + set.extend(vec![&1, &2, &3, &4]); + set.extend(vec![5, 6]); + assert_eq!(set.into_iter().collect::<Vec<_>>(), vec![1, 2, 3, 4, 5, 6]); + } + + #[test] + fn comparisons() { + let set_a: IndexSet<_> = (0..3).collect(); + let set_b: IndexSet<_> = (3..6).collect(); + let set_c: IndexSet<_> = (0..6).collect(); + let set_d: IndexSet<_> = (3..9).collect(); + + assert!(!set_a.is_disjoint(&set_a)); + assert!(set_a.is_subset(&set_a)); + assert!(set_a.is_superset(&set_a)); + + assert!(set_a.is_disjoint(&set_b)); + assert!(set_b.is_disjoint(&set_a)); + assert!(!set_a.is_subset(&set_b)); + assert!(!set_b.is_subset(&set_a)); + assert!(!set_a.is_superset(&set_b)); + assert!(!set_b.is_superset(&set_a)); + + assert!(!set_a.is_disjoint(&set_c)); + assert!(!set_c.is_disjoint(&set_a)); + assert!(set_a.is_subset(&set_c)); + assert!(!set_c.is_subset(&set_a)); + assert!(!set_a.is_superset(&set_c)); + assert!(set_c.is_superset(&set_a)); + + assert!(!set_c.is_disjoint(&set_d)); + assert!(!set_d.is_disjoint(&set_c)); + assert!(!set_c.is_subset(&set_d)); + assert!(!set_d.is_subset(&set_c)); + assert!(!set_c.is_superset(&set_d)); + assert!(!set_d.is_superset(&set_c)); + } + + #[test] + fn iter_comparisons() { + use std::iter::empty; + + fn check<'a, I1, I2>(iter1: I1, iter2: I2) + where + I1: Iterator<Item = &'a i32>, + I2: Iterator<Item = i32>, + { + assert!(iter1.cloned().eq(iter2)); + } + + let set_a: IndexSet<_> = (0..3).collect(); + let set_b: IndexSet<_> = (3..6).collect(); + let set_c: IndexSet<_> = (0..6).collect(); + let set_d: IndexSet<_> = (3..9).rev().collect(); + + check(set_a.difference(&set_a), empty()); + check(set_a.symmetric_difference(&set_a), empty()); + check(set_a.intersection(&set_a), 0..3); + check(set_a.union(&set_a), 0..3); + + check(set_a.difference(&set_b), 0..3); + check(set_b.difference(&set_a), 3..6); + check(set_a.symmetric_difference(&set_b), 0..6); + check(set_b.symmetric_difference(&set_a), (3..6).chain(0..3)); + check(set_a.intersection(&set_b), empty()); + check(set_b.intersection(&set_a), empty()); + check(set_a.union(&set_b), 0..6); + check(set_b.union(&set_a), (3..6).chain(0..3)); + + check(set_a.difference(&set_c), empty()); + check(set_c.difference(&set_a), 3..6); + check(set_a.symmetric_difference(&set_c), 3..6); + check(set_c.symmetric_difference(&set_a), 3..6); + check(set_a.intersection(&set_c), 0..3); + check(set_c.intersection(&set_a), 0..3); + check(set_a.union(&set_c), 0..6); + check(set_c.union(&set_a), 0..6); + + check(set_c.difference(&set_d), 0..3); + check(set_d.difference(&set_c), (6..9).rev()); + check( + set_c.symmetric_difference(&set_d), + (0..3).chain((6..9).rev()), + ); + check(set_d.symmetric_difference(&set_c), (6..9).rev().chain(0..3)); + check(set_c.intersection(&set_d), 3..6); + check(set_d.intersection(&set_c), (3..6).rev()); + check(set_c.union(&set_d), (0..6).chain((6..9).rev())); + check(set_d.union(&set_c), (3..9).rev().chain(0..3)); + } + + #[test] + fn ops() { + let empty = IndexSet::<i32>::new(); + let set_a: IndexSet<_> = (0..3).collect(); + let set_b: IndexSet<_> = (3..6).collect(); + let set_c: IndexSet<_> = (0..6).collect(); + let set_d: IndexSet<_> = (3..9).rev().collect(); + + // FIXME: #[allow(clippy::eq_op)] in Rust 1.31 + #[cfg_attr(feature = "cargo-clippy", allow(renamed_and_removed_lints, eq_op))] + { + assert_eq!(&set_a & &set_a, set_a); + assert_eq!(&set_a | &set_a, set_a); + assert_eq!(&set_a ^ &set_a, empty); + assert_eq!(&set_a - &set_a, empty); + } + + assert_eq!(&set_a & &set_b, empty); + assert_eq!(&set_b & &set_a, empty); + assert_eq!(&set_a | &set_b, set_c); + assert_eq!(&set_b | &set_a, set_c); + assert_eq!(&set_a ^ &set_b, set_c); + assert_eq!(&set_b ^ &set_a, set_c); + assert_eq!(&set_a - &set_b, set_a); + assert_eq!(&set_b - &set_a, set_b); + + assert_eq!(&set_a & &set_c, set_a); + assert_eq!(&set_c & &set_a, set_a); + assert_eq!(&set_a | &set_c, set_c); + assert_eq!(&set_c | &set_a, set_c); + assert_eq!(&set_a ^ &set_c, set_b); + assert_eq!(&set_c ^ &set_a, set_b); + assert_eq!(&set_a - &set_c, empty); + assert_eq!(&set_c - &set_a, set_b); + + assert_eq!(&set_c & &set_d, set_b); + assert_eq!(&set_d & &set_c, set_b); + assert_eq!(&set_c | &set_d, &set_a | &set_d); + assert_eq!(&set_d | &set_c, &set_a | &set_d); + assert_eq!(&set_c ^ &set_d, &set_a | &(&set_d - &set_b)); + assert_eq!(&set_d ^ &set_c, &set_a | &(&set_d - &set_b)); + assert_eq!(&set_c - &set_d, set_a); + assert_eq!(&set_d - &set_c, &set_d - &set_b); + } +} diff --git a/third_party/rust/indexmap/src/util.rs b/third_party/rust/indexmap/src/util.rs new file mode 100644 index 0000000000..5388f470cf --- /dev/null +++ b/third_party/rust/indexmap/src/util.rs @@ -0,0 +1,39 @@ +use core::iter::Enumerate; +use core::ops::{Bound, Range, RangeBounds}; + +pub(crate) fn third<A, B, C>(t: (A, B, C)) -> C { + t.2 +} + +pub(crate) fn enumerate<I>(iterable: I) -> Enumerate<I::IntoIter> +where + I: IntoIterator, +{ + iterable.into_iter().enumerate() +} + +pub(crate) fn simplify_range<R>(range: R, len: usize) -> Range<usize> +where + R: RangeBounds<usize>, +{ + let start = match range.start_bound() { + Bound::Unbounded => 0, + Bound::Included(&i) if i <= len => i, + Bound::Excluded(&i) if i < len => i + 1, + bound => panic!("range start {:?} should be <= length {}", bound, len), + }; + let end = match range.end_bound() { + Bound::Unbounded => len, + Bound::Excluded(&i) if i <= len => i, + Bound::Included(&i) if i < len => i + 1, + bound => panic!("range end {:?} should be <= length {}", bound, len), + }; + if start > end { + panic!( + "range start {:?} should be <= range end {:?}", + range.start_bound(), + range.end_bound() + ); + } + start..end +} diff --git a/third_party/rust/indexmap/tests/equivalent_trait.rs b/third_party/rust/indexmap/tests/equivalent_trait.rs new file mode 100644 index 0000000000..ff5943a3ed --- /dev/null +++ b/third_party/rust/indexmap/tests/equivalent_trait.rs @@ -0,0 +1,53 @@ +use indexmap::indexmap; +use indexmap::Equivalent; + +use std::hash::Hash; + +#[derive(Debug, Hash)] +pub struct Pair<A, B>(pub A, pub B); + +impl<A, B, C, D> PartialEq<(A, B)> for Pair<C, D> +where + C: PartialEq<A>, + D: PartialEq<B>, +{ + fn eq(&self, rhs: &(A, B)) -> bool { + self.0 == rhs.0 && self.1 == rhs.1 + } +} + +impl<A, B, X> Equivalent<X> for Pair<A, B> +where + Pair<A, B>: PartialEq<X>, + A: Hash + Eq, + B: Hash + Eq, +{ + fn equivalent(&self, other: &X) -> bool { + *self == *other + } +} + +#[test] +fn test_lookup() { + let s = String::from; + let map = indexmap! { + (s("a"), s("b")) => 1, + (s("a"), s("x")) => 2, + }; + + assert!(map.contains_key(&Pair("a", "b"))); + assert!(!map.contains_key(&Pair("b", "a"))); +} + +#[test] +fn test_string_str() { + let s = String::from; + let mut map = indexmap! { + s("a") => 1, s("b") => 2, + s("x") => 3, s("y") => 4, + }; + + assert!(map.contains_key("a")); + assert!(!map.contains_key("z")); + assert_eq!(map.swap_remove("b"), Some(2)); +} diff --git a/third_party/rust/indexmap/tests/macros_full_path.rs b/third_party/rust/indexmap/tests/macros_full_path.rs new file mode 100644 index 0000000000..2467d9b4f5 --- /dev/null +++ b/third_party/rust/indexmap/tests/macros_full_path.rs @@ -0,0 +1,19 @@ +#[test] +fn test_create_map() { + let _m = indexmap::indexmap! { + 1 => 2, + 7 => 1, + 2 => 2, + 3 => 3, + }; +} + +#[test] +fn test_create_set() { + let _s = indexmap::indexset! { + 1, + 7, + 2, + 3, + }; +} diff --git a/third_party/rust/indexmap/tests/quick.rs b/third_party/rust/indexmap/tests/quick.rs new file mode 100644 index 0000000000..10e11b5285 --- /dev/null +++ b/third_party/rust/indexmap/tests/quick.rs @@ -0,0 +1,493 @@ +use indexmap::{IndexMap, IndexSet}; +use itertools::Itertools; + +use quickcheck::quickcheck; +use quickcheck::Arbitrary; +use quickcheck::Gen; +use quickcheck::TestResult; + +use rand::Rng; + +use fnv::FnvHasher; +use std::hash::{BuildHasher, BuildHasherDefault}; +type FnvBuilder = BuildHasherDefault<FnvHasher>; +type IndexMapFnv<K, V> = IndexMap<K, V, FnvBuilder>; + +use std::cmp::min; +use std::collections::HashMap; +use std::collections::HashSet; +use std::fmt::Debug; +use std::hash::Hash; +use std::iter::FromIterator; +use std::ops::Bound; +use std::ops::Deref; + +use indexmap::map::Entry as OEntry; +use std::collections::hash_map::Entry as HEntry; + +fn set<'a, T: 'a, I>(iter: I) -> HashSet<T> +where + I: IntoIterator<Item = &'a T>, + T: Copy + Hash + Eq, +{ + iter.into_iter().cloned().collect() +} + +fn indexmap<'a, T: 'a, I>(iter: I) -> IndexMap<T, ()> +where + I: IntoIterator<Item = &'a T>, + T: Copy + Hash + Eq, +{ + IndexMap::from_iter(iter.into_iter().cloned().map(|k| (k, ()))) +} + +quickcheck! { + fn contains(insert: Vec<u32>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + insert.iter().all(|&key| map.get(&key).is_some()) + } + + fn contains_not(insert: Vec<u8>, not: Vec<u8>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + let nots = &set(¬) - &set(&insert); + nots.iter().all(|&key| map.get(&key).is_none()) + } + + fn insert_remove(insert: Vec<u8>, remove: Vec<u8>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + for &key in &remove { + map.swap_remove(&key); + } + let elements = &set(&insert) - &set(&remove); + map.len() == elements.len() && map.iter().count() == elements.len() && + elements.iter().all(|k| map.get(k).is_some()) + } + + fn insertion_order(insert: Vec<u32>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + itertools::assert_equal(insert.iter().unique(), map.keys()); + true + } + + fn pop(insert: Vec<u8>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + let mut pops = Vec::new(); + while let Some((key, _v)) = map.pop() { + pops.push(key); + } + pops.reverse(); + + itertools::assert_equal(insert.iter().unique(), &pops); + true + } + + fn with_cap(cap: usize) -> bool { + let map: IndexMap<u8, u8> = IndexMap::with_capacity(cap); + println!("wish: {}, got: {} (diff: {})", cap, map.capacity(), map.capacity() as isize - cap as isize); + map.capacity() >= cap + } + + fn drain_full(insert: Vec<u8>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + let mut clone = map.clone(); + let drained = clone.drain(..); + for (key, _) in drained { + map.swap_remove(&key); + } + map.is_empty() + } + + fn drain_bounds(insert: Vec<u8>, range: (Bound<usize>, Bound<usize>)) -> TestResult { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + + // First see if `Vec::drain` is happy with this range. + let result = std::panic::catch_unwind(|| { + let mut keys: Vec<u8> = map.keys().cloned().collect(); + keys.drain(range); + keys + }); + + if let Ok(keys) = result { + map.drain(range); + // Check that our `drain` matches the same key order. + assert!(map.keys().eq(&keys)); + // Check that hash lookups all work too. + assert!(keys.iter().all(|key| map.contains_key(key))); + TestResult::passed() + } else { + // If `Vec::drain` panicked, so should we. + TestResult::must_fail(move || { map.drain(range); }) + } + } + + fn shift_remove(insert: Vec<u8>, remove: Vec<u8>) -> bool { + let mut map = IndexMap::new(); + for &key in &insert { + map.insert(key, ()); + } + for &key in &remove { + map.shift_remove(&key); + } + let elements = &set(&insert) - &set(&remove); + + // Check that order is preserved after removals + let mut iter = map.keys(); + for &key in insert.iter().unique() { + if elements.contains(&key) { + assert_eq!(Some(key), iter.next().cloned()); + } + } + + map.len() == elements.len() && map.iter().count() == elements.len() && + elements.iter().all(|k| map.get(k).is_some()) + } + + fn indexing(insert: Vec<u8>) -> bool { + let mut map: IndexMap<_, _> = insert.into_iter().map(|x| (x, x)).collect(); + let set: IndexSet<_> = map.keys().cloned().collect(); + assert_eq!(map.len(), set.len()); + + for (i, &key) in set.iter().enumerate() { + assert_eq!(map.get_index(i), Some((&key, &key))); + assert_eq!(set.get_index(i), Some(&key)); + assert_eq!(map[i], key); + assert_eq!(set[i], key); + + *map.get_index_mut(i).unwrap().1 >>= 1; + map[i] <<= 1; + } + + set.iter().enumerate().all(|(i, &key)| { + let value = key & !1; + map[&key] == value && map[i] == value + }) + } +} + +use crate::Op::*; +#[derive(Copy, Clone, Debug)] +enum Op<K, V> { + Add(K, V), + Remove(K), + AddEntry(K, V), + RemoveEntry(K), +} + +impl<K, V> Arbitrary for Op<K, V> +where + K: Arbitrary, + V: Arbitrary, +{ + fn arbitrary<G: Gen>(g: &mut G) -> Self { + match g.gen::<u32>() % 4 { + 0 => Add(K::arbitrary(g), V::arbitrary(g)), + 1 => AddEntry(K::arbitrary(g), V::arbitrary(g)), + 2 => Remove(K::arbitrary(g)), + _ => RemoveEntry(K::arbitrary(g)), + } + } +} + +fn do_ops<K, V, S>(ops: &[Op<K, V>], a: &mut IndexMap<K, V, S>, b: &mut HashMap<K, V>) +where + K: Hash + Eq + Clone, + V: Clone, + S: BuildHasher, +{ + for op in ops { + match *op { + Add(ref k, ref v) => { + a.insert(k.clone(), v.clone()); + b.insert(k.clone(), v.clone()); + } + AddEntry(ref k, ref v) => { + a.entry(k.clone()).or_insert_with(|| v.clone()); + b.entry(k.clone()).or_insert_with(|| v.clone()); + } + Remove(ref k) => { + a.swap_remove(k); + b.remove(k); + } + RemoveEntry(ref k) => { + if let OEntry::Occupied(ent) = a.entry(k.clone()) { + ent.swap_remove_entry(); + } + if let HEntry::Occupied(ent) = b.entry(k.clone()) { + ent.remove_entry(); + } + } + } + //println!("{:?}", a); + } +} + +fn assert_maps_equivalent<K, V>(a: &IndexMap<K, V>, b: &HashMap<K, V>) -> bool +where + K: Hash + Eq + Debug, + V: Eq + Debug, +{ + assert_eq!(a.len(), b.len()); + assert_eq!(a.iter().next().is_some(), b.iter().next().is_some()); + for key in a.keys() { + assert!(b.contains_key(key), "b does not contain {:?}", key); + } + for key in b.keys() { + assert!(a.get(key).is_some(), "a does not contain {:?}", key); + } + for key in a.keys() { + assert_eq!(a[key], b[key]); + } + true +} + +quickcheck! { + fn operations_i8(ops: Large<Vec<Op<i8, i8>>>) -> bool { + let mut map = IndexMap::new(); + let mut reference = HashMap::new(); + do_ops(&ops, &mut map, &mut reference); + assert_maps_equivalent(&map, &reference) + } + + fn operations_string(ops: Vec<Op<Alpha, i8>>) -> bool { + let mut map = IndexMap::new(); + let mut reference = HashMap::new(); + do_ops(&ops, &mut map, &mut reference); + assert_maps_equivalent(&map, &reference) + } + + fn keys_values(ops: Large<Vec<Op<i8, i8>>>) -> bool { + let mut map = IndexMap::new(); + let mut reference = HashMap::new(); + do_ops(&ops, &mut map, &mut reference); + let mut visit = IndexMap::new(); + for (k, v) in map.keys().zip(map.values()) { + assert_eq!(&map[k], v); + assert!(!visit.contains_key(k)); + visit.insert(*k, *v); + } + assert_eq!(visit.len(), reference.len()); + true + } + + fn keys_values_mut(ops: Large<Vec<Op<i8, i8>>>) -> bool { + let mut map = IndexMap::new(); + let mut reference = HashMap::new(); + do_ops(&ops, &mut map, &mut reference); + let mut visit = IndexMap::new(); + let keys = Vec::from_iter(map.keys().cloned()); + for (k, v) in keys.iter().zip(map.values_mut()) { + assert_eq!(&reference[k], v); + assert!(!visit.contains_key(k)); + visit.insert(*k, *v); + } + assert_eq!(visit.len(), reference.len()); + true + } + + fn equality(ops1: Vec<Op<i8, i8>>, removes: Vec<usize>) -> bool { + let mut map = IndexMap::new(); + let mut reference = HashMap::new(); + do_ops(&ops1, &mut map, &mut reference); + let mut ops2 = ops1.clone(); + for &r in &removes { + if !ops2.is_empty() { + let i = r % ops2.len(); + ops2.remove(i); + } + } + let mut map2 = IndexMapFnv::default(); + let mut reference2 = HashMap::new(); + do_ops(&ops2, &mut map2, &mut reference2); + assert_eq!(map == map2, reference == reference2); + true + } + + fn retain_ordered(keys: Large<Vec<i8>>, remove: Large<Vec<i8>>) -> () { + let mut map = indexmap(keys.iter()); + let initial_map = map.clone(); // deduplicated in-order input + let remove_map = indexmap(remove.iter()); + let keys_s = set(keys.iter()); + let remove_s = set(remove.iter()); + let answer = &keys_s - &remove_s; + map.retain(|k, _| !remove_map.contains_key(k)); + + // check the values + assert_eq!(map.len(), answer.len()); + for key in &answer { + assert!(map.contains_key(key)); + } + // check the order + itertools::assert_equal(map.keys(), initial_map.keys().filter(|&k| !remove_map.contains_key(k))); + } + + fn sort_1(keyvals: Large<Vec<(i8, i8)>>) -> () { + let mut map: IndexMap<_, _> = IndexMap::from_iter(keyvals.to_vec()); + let mut answer = keyvals.0; + answer.sort_by_key(|t| t.0); + + // reverse dedup: Because IndexMap::from_iter keeps the last value for + // identical keys + answer.reverse(); + answer.dedup_by_key(|t| t.0); + answer.reverse(); + + map.sort_by(|k1, _, k2, _| Ord::cmp(k1, k2)); + + // check it contains all the values it should + for &(key, val) in &answer { + assert_eq!(map[&key], val); + } + + // check the order + + let mapv = Vec::from_iter(map); + assert_eq!(answer, mapv); + + } + + fn sort_2(keyvals: Large<Vec<(i8, i8)>>) -> () { + let mut map: IndexMap<_, _> = IndexMap::from_iter(keyvals.to_vec()); + map.sort_by(|_, v1, _, v2| Ord::cmp(v1, v2)); + assert_sorted_by_key(map, |t| t.1); + } + + fn reverse(keyvals: Large<Vec<(i8, i8)>>) -> () { + let mut map: IndexMap<_, _> = IndexMap::from_iter(keyvals.to_vec()); + + fn generate_answer(input: &Vec<(i8, i8)>) -> Vec<(i8, i8)> { + // to mimic what `IndexMap::from_iter` does: + // need to get (A) the unique keys in forward order, and (B) the + // last value of each of those keys. + + // create (A): an iterable that yields the unique keys in ltr order + let mut seen_keys = HashSet::new(); + let unique_keys_forward = input.iter().filter_map(move |(k, _)| { + if seen_keys.contains(k) { None } + else { seen_keys.insert(*k); Some(*k) } + }); + + // create (B): a mapping of keys to the last value seen for that key + // this is the same as reversing the input and taking the first + // value seen for that key! + let mut last_val_per_key = HashMap::new(); + for &(k, v) in input.iter().rev() { + if !last_val_per_key.contains_key(&k) { + last_val_per_key.insert(k, v); + } + } + + // iterate over the keys in (A) in order, and match each one with + // the corresponding last value from (B) + let mut ans: Vec<_> = unique_keys_forward + .map(|k| (k, *last_val_per_key.get(&k).unwrap())) + .collect(); + + // finally, since this test is testing `.reverse()`, reverse the + // answer in-place + ans.reverse(); + + ans + } + + let answer = generate_answer(&keyvals.0); + + // perform the work + map.reverse(); + + // check it contains all the values it should + for &(key, val) in &answer { + assert_eq!(map[&key], val); + } + + // check the order + let mapv = Vec::from_iter(map); + assert_eq!(answer, mapv); + } +} + +fn assert_sorted_by_key<I, Key, X>(iterable: I, key: Key) +where + I: IntoIterator, + I::Item: Ord + Clone + Debug, + Key: Fn(&I::Item) -> X, + X: Ord, +{ + let input = Vec::from_iter(iterable); + let mut sorted = input.clone(); + sorted.sort_by_key(key); + assert_eq!(input, sorted); +} + +#[derive(Clone, Debug, Hash, PartialEq, Eq)] +struct Alpha(String); + +impl Deref for Alpha { + type Target = String; + fn deref(&self) -> &String { + &self.0 + } +} + +const ALPHABET: &[u8] = b"abcdefghijklmnopqrstuvwxyz"; + +impl Arbitrary for Alpha { + fn arbitrary<G: Gen>(g: &mut G) -> Self { + let len = g.next_u32() % g.size() as u32; + let len = min(len, 16); + Alpha( + (0..len) + .map(|_| ALPHABET[g.next_u32() as usize % ALPHABET.len()] as char) + .collect(), + ) + } + + fn shrink(&self) -> Box<dyn Iterator<Item = Self>> { + Box::new((**self).shrink().map(Alpha)) + } +} + +/// quickcheck Arbitrary adaptor -- make a larger vec +#[derive(Clone, Debug)] +struct Large<T>(T); + +impl<T> Deref for Large<T> { + type Target = T; + fn deref(&self) -> &T { + &self.0 + } +} + +impl<T> Arbitrary for Large<Vec<T>> +where + T: Arbitrary, +{ + fn arbitrary<G: Gen>(g: &mut G) -> Self { + let len = g.next_u32() % (g.size() * 10) as u32; + Large((0..len).map(|_| T::arbitrary(g)).collect()) + } + + fn shrink(&self) -> Box<dyn Iterator<Item = Self>> { + Box::new((**self).shrink().map(Large)) + } +} diff --git a/third_party/rust/indexmap/tests/tests.rs b/third_party/rust/indexmap/tests/tests.rs new file mode 100644 index 0000000000..7d522f1c97 --- /dev/null +++ b/third_party/rust/indexmap/tests/tests.rs @@ -0,0 +1,28 @@ +use indexmap::{indexmap, indexset}; + +#[test] +fn test_sort() { + let m = indexmap! { + 1 => 2, + 7 => 1, + 2 => 2, + 3 => 3, + }; + + itertools::assert_equal( + m.sorted_by(|_k1, v1, _k2, v2| v1.cmp(v2)), + vec![(7, 1), (1, 2), (2, 2), (3, 3)], + ); +} + +#[test] +fn test_sort_set() { + let s = indexset! { + 1, + 7, + 2, + 3, + }; + + itertools::assert_equal(s.sorted_by(|v1, v2| v1.cmp(v2)), vec![1, 2, 3, 7]); +} |