Use workspace rand and rust-version in ruff_unicode_names2 (#29261)

Summary
--

This updates `ruff_unicode_names2` and `ruff_unicode_names2_generator`
to use our workspace `rand` dependency and `rust-version`, prompted by
#29257 and #29359. Bumping the Rust version also introduces a handful of
Clippy diagnostic that Clippy was also able to fix automatically. I
originally pinned both of these to stay closer to the upstream version,
but we've already diverged a bit, and the required changes were quite
small anyway, so this seems like the better way to go.

This also allows us to drop the `std` feature on `rand`, a transitive
dependency on `zerocopy`, and the accompanying allowlist entry for its
build script.

While I was here, I also switched to our workspace edition, removed
unused feature-gated code and unnecessary suppressions, enabled
formatting and typo checks, and enabled doctests.

These are the changes from `rand` for reference:

- https://rust-random.github.io/book/update-0.10.html
- https://rust-random.github.io/book/update-0.9.html

Test Plan
--

Existing tests

---------

Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>

.cargo/deny.toml

@@ -73,7 +73,6 @@ allow-build-scripts = [
     "windows_x86_64_gnullvm",
     "windows_x86_64_msvc",
     "wit-bindgen",
-    "zerocopy",
     "zmij",
     "zstd-safe",
     "zstd-sys",

Cargo.lock

@@ -2634,15 +2634,6 @@ dependencies = [
  "zerovec",
 ]
 
-[[package]]
-name = "ppv-lite86"
-version = "0.2.21"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "85eae3c4ed2f50dcfe72643da4befc30deadb458a9b590d720cde2f2b1e97da9"
-dependencies = [
- "zerocopy",
-]
-
 [[package]]
 name = "predicates"
 version = "3.1.4"
@@ -2893,8 +2884,6 @@ version = "0.8.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "34af8d1a0e25924bc5b7c43c079c942339d8f0a8b57c39049bef581b46327404"
 dependencies = [
- "libc",
- "rand_chacha",
  "rand_core 0.6.4",
 ]
 
@@ -2909,24 +2898,11 @@ dependencies = [
  "rand_core 0.10.0",
 ]
 
-[[package]]
-name = "rand_chacha"
-version = "0.3.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e6c10a63a0fa32252be49d21e7709d4d4baf8d231c2dbce1eaa8141b9b127d88"
-dependencies = [
- "ppv-lite86",
- "rand_core 0.6.4",
-]
-
 [[package]]
 name = "rand_core"
 version = "0.6.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ec0be4795e2f6a28069bec0b5ff3e2ac9bafc99e6a9a7dc3547996c5c816922c"
-dependencies = [
- "getrandom 0.2.16",
-]
 
 [[package]]
 name = "rand_core"
@@ -3791,7 +3767,7 @@ name = "ruff_unicode_names2"
 version = "0.0.17"
 dependencies = [
  "phf 0.11.3",
- "rand 0.8.5",
+ "rand 0.10.2",
  "ruff_unicode_names2_generator",
 ]
 
@@ -3800,7 +3776,7 @@ name = "ruff_unicode_names2_generator"
 version = "0.0.17"
 dependencies = [
  "phf_codegen",
- "rand 0.8.5",
+ "rand 0.10.2",
 ]
 
 [[package]]
@@ -5893,26 +5869,6 @@ dependencies = [
  "synstructure",
 ]
 
-[[package]]
-name = "zerocopy"
-version = "0.8.27"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0894878a5fa3edfd6da3f88c4805f4c8558e2b996227a3d864f47fe11e38282c"
-dependencies = [
- "zerocopy-derive",
-]
-
-[[package]]
-name = "zerocopy-derive"
-version = "0.8.27"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "88d2b8d9c68ad2b9e4340d7832716a4d21a22a1154777ad56ea55c51a9cf3831"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn 2.0.119",
-]
-
 [[package]]
 name = "zerofrom"
 version = "0.1.6"

_typos.toml

@@ -2,8 +2,8 @@
 # https://github.com/crate-ci/typos/issues/868
 extend-exclude = [
     "crates/ty_vendored/vendor/**/*",
-    "crates/ruff_unicode_names2/**/*",
-    "crates/ruff_unicode_names2_generator/**/*",
+    "crates/ruff_unicode_names2/data/NameAliases.txt",
+    "crates/ruff_unicode_names2/data/UnicodeData.txt",
     "**/resources/**/*",
     "**/snapshots/**/*",
     "crates/ruff_linter/src/rules/flake8_implicit_str_concat/rules/collection_literal.rs",

crates/ruff_unicode_names2/Cargo.toml

@@ -17,20 +17,10 @@ phf = { workspace = true }
 ruff_unicode_names2_generator = { workspace = true }
 
 [dev-dependencies]
-rand = { version = "0.8.5", default-features = false, features = ["std", "std_rng"] }
-
-[lints.rust]
-# Ruff uses std; retain upstream's optional no_std code without enabling it.
-unexpected_cfgs = { level = "warn", check-cfg = ['cfg(feature, values("no_std"))'] }
-
-[lints.clippy]
-# Preserve the vendored source.
-unnecessary_map_or = "allow"
-question_mark = "allow"
+rand = { workspace = true }
 
 [lib]
 name = "unicode_names2"
 # Upstream unit tests, benchmarks, and macro doctests require nightly Rust.
 test = false
-doctest = false
 bench = false

crates/ruff_unicode_names2/examples/count.rs

@@ -12,7 +12,7 @@ use std::char;
 
 fn main() {
     let number = (0u32..0x10FFFF)
-        .filter(|x| char::from_u32(*x).map_or(false, |c| unicode_names2::name(c).is_some()))
+        .filter(|x| char::from_u32(*x).is_some_and(|c| unicode_names2::name(c).is_some()))
         .count();
 
     println!("there are {} named code points", number)

crates/ruff_unicode_names2/rustfmt.toml

@@ -1,2 +0,0 @@
-# Preserve the vendored source formatting.
-disable_all_formatting = true

crates/ruff_unicode_names2/src/lib.rs

@@ -1,9 +1,7 @@
 //! Convert between characters and their standard names.
 //!
 //! This crate provides two functions for mapping from a `char` to the
-//! name given by the Unicode standard (17.0). There are no runtime
-//! requirements so this is usable with only `core` (this requires
-//! specifying the `no_std` cargo feature). The tables are heavily
+//! name given by the Unicode standard (17.0). The tables are heavily
 //! compressed, but still large (500KB), and still offer efficient
 //! `O(1)` look-ups in both directions (more precisely, `O(length of
 //! name)`).
@@ -47,37 +45,24 @@
 //! }
 //! ```
 
-#![cfg_attr(feature = "no_std", no_std)]
 #![cfg_attr(test, feature(test))]
 #![deny(missing_docs, unsafe_code)]
 
-#[cfg(all(test, feature = "no_std"))]
-#[macro_use]
-extern crate std;
-
 use core::{char, fmt};
 use generated::{
-    MAX_NAME_LENGTH, PHRASEBOOK_OFFSETS1, PHRASEBOOK_OFFSETS2, PHRASEBOOK_OFFSET_SHIFT,
+    MAX_NAME_LENGTH, PHRASEBOOK_OFFSET_SHIFT, PHRASEBOOK_OFFSETS1, PHRASEBOOK_OFFSETS2,
 };
 
-#[allow(dead_code)]
-#[rustfmt::skip]
-#[allow(clippy::all)]
 mod generated {
     include!(concat!(env!("OUT_DIR"), "/generated.rs"));
 }
-#[allow(dead_code)]
-#[rustfmt::skip]
-#[allow(clippy::all)]
 mod generated_phf {
     include!(concat!(env!("OUT_DIR"), "/generated_phf.rs"));
 }
-#[allow(dead_code)]
 mod jamo;
 
 /// A map of unicode aliases to their corresponding values.
 /// Generated in generator
-#[allow(dead_code)]
 static ALIASES: phf::Map<&'static [u8], char> =
     include!(concat!(env!("OUT_DIR"), "/generated_alias.rs"));
 
@@ -363,10 +348,7 @@ pub fn character(search_name: &str) -> Option<char> {
                 _ => return None,
             }
         }
-        let ch = match char::from_u32(v) {
-            Some(ch) => ch,
-            None => return None,
-        };
+        let ch = char::from_u32(v)?;
 
         // check if the resulting code is indeed in the known ranges
         if is_cjk_unified_ideograph(ch) {
@@ -422,7 +404,7 @@ pub fn character(search_name: &str) -> Option<char> {
 mod tests {
     use super::*;
     use rand::{
-        distributions::{Distribution, Standard},
+        distr::{Distribution, StandardUniform},
         prelude::{SeedableRng, StdRng},
     };
     use std::char;
@@ -616,7 +598,7 @@ mod tests {
     fn name_10000_invalid(b: &mut Bencher) {
         // be consistent across runs, but avoid sequential/caching.
         let mut rng = StdRng::seed_from_u64(0x12345678);
-        let chars: Vec<char> = Standard
+        let chars: Vec<char> = StandardUniform
             .sample_iter(&mut rng)
             .take(10000)
             .filter_map(|c| match c {
@@ -655,7 +637,7 @@ mod tests {
         // be consistent across runs, but avoid sequential/caching.
         let mut rng = StdRng::seed_from_u64(0x12345678);
 
-        let names: Vec<_> = Standard
+        let names: Vec<_> = StandardUniform
             .sample_iter(&mut rng)
             .take(10000)
             .filter_map(name)
@@ -669,8 +651,3 @@ mod tests {
         })
     }
 }
-
-#[cfg(all(feature = "no_std", not(test)))]
-mod std {
-    pub use core::{clone, fmt, marker};
-}

crates/ruff_unicode_names2_generator/Cargo.toml

@@ -3,8 +3,8 @@ name = "ruff_unicode_names2_generator"
 version = "0.0.17"
 description = "This is an internal component crate of Ruff"
 authors = { workspace = true }
-edition = "2018"
-rust-version = "1.65.0"
+edition = { workspace = true }
+rust-version = { workspace = true }
 homepage = { workspace = true }
 documentation = { workspace = true }
 repository = { workspace = true }
@@ -12,14 +12,7 @@ license = "MIT OR Apache-2.0"
 
 [dependencies]
 phf_codegen = { workspace = true }
-# Preserve the RNG used by upstream's table generator.
-rand = { version = "0.8.5", default-features = false, features = ["std", "std_rng"] }
-
-[lints.rust]
-unexpected_cfgs = { level = "warn", check-cfg = ['cfg(feature, values("timing"))'] }
-
-[lints.rustdoc]
-bare_urls = "allow"
+rand = { workspace = true }
 
 [lib]
 name = "unicode_names2_generator"

crates/ruff_unicode_names2_generator/rustfmt.toml

@@ -1,2 +0,0 @@
-# Preserve the vendored source formatting.
-disable_all_formatting = true

crates/ruff_unicode_names2_generator/src/formatting.rs

@@ -11,7 +11,7 @@ impl Context {
     where
         F: Fn(&T) -> String,
     {
-        w!(self, "pub static {}: &'static [{}] = &[", name, ty);
+        w!(self, "pub static {}: &[{}] = &[", name, ty);
 
         let mut width = LINE_LIMIT;
         for e in elements.iter() {
@@ -41,7 +41,7 @@ impl Context {
     pub fn write_plain_string(&mut self, name: &str, data: &str) {
         assert!(!data.contains('\\'));
 
-        w!(self, "pub static {}: &'static str = \"", name);
+        w!(self, "pub static {}: &str = \"", name);
 
         for chunk in data.as_bytes().chunks(LINE_LIMIT - 6) {
             w!(self, "\\\n    ");

crates/ruff_unicode_names2_generator/src/lib.rs

@@ -1,10 +1,9 @@
 use formatting::Context;
 use std::{
     char, cmp,
-    collections::{hash_map, HashMap},
+    collections::{HashMap, hash_map},
     fs::{self, File},
-    io::{self, prelude::*, BufWriter},
-    iter::repeat,
+    io::{self, BufWriter, prelude::*},
     path::Path,
 };
 
@@ -156,7 +155,7 @@ fn create_lexicon_and_offsets(
                 // insert the suffixes of this word which saves about
                 // 10KB (we could theoretically insert all substrings,
                 // upto a certain length, but this only saves ~300
-                // bytes or so and is noticably slower).
+                // bytes or so and is noticeably slower).
                 for i in 1..n.len() {
                     if t.insert(n[i..].bytes(), Some(offset + i), true).0 {
                         // once we've found a string that's already
@@ -238,7 +237,7 @@ fn bin_data(dat: &[u32]) -> (Vec<u32>, Vec<u32>, usize) {
 fn write_codepoint_maps(ctxt: &mut Context, codepoint_names: Vec<(char, &str)>) {
     let (lexicon_string, mut lexicon_words) = create_lexicon_and_offsets(codepoint_names.clone());
 
-    let num_escapes = (lexicon_words.len() + 255) / 256;
+    let num_escapes = lexicon_words.len().div_ceil(256);
 
     // we reserve the high bit (end of word) and 127,126... for
     // non-space splits.  The high bit saves about 10KB, and doing the
@@ -293,9 +292,11 @@ fn write_codepoint_maps(ctxt: &mut Context, codepoint_names: Vec<(char, &str)>)
             previous_len = len;
         }
 
-        assert!(word_encodings
+        assert!(
+            word_encodings
                 .insert(word, vec![hi as u32, lo as u32])
-            .is_none());
+                .is_none()
+        );
     }
     // don't forget the last one.
     lexicon_ordered_lengths.push((lexicon_offsets.len(), previous_len));
@@ -306,7 +307,7 @@ fn write_codepoint_maps(ctxt: &mut Context, codepoint_names: Vec<(char, &str)>)
     // this is a map from `char` -> the index in phrasebook. it is
     // currently huge, but it has a lot of 0's, so we compress it
     // using the binning, below.
-    let mut phrasebook_offsets = repeat(0).take(0x10FFFF + 1).collect::<Vec<_>>();
+    let mut phrasebook_offsets: Vec<_> = std::iter::repeat_n(0, 0x10FFFF + 1).collect();
     let mut longest_name = 0;
     for &(cp, name) in codepoint_names.iter() {
         longest_name = cmp::max(name.len(), longest_name);

crates/ruff_unicode_names2_generator/src/phf.rs

@@ -1,10 +1,9 @@
 //! Computes a perfect hash table using [the CHD
 //! algorithm](http://cmph.sourceforge.net/papers/esa09.pdf).
 //!
-//! Strongly inspired by https://github.com/sfackler/rust-phf
+//! Strongly inspired by <https://github.com/sfackler/rust-phf>
 
-use rand::prelude::{Rng, SeedableRng, SliceRandom, StdRng};
-use std::iter::repeat;
+use rand::prelude::{RngExt, SeedableRng, SliceRandom, StdRng};
 
 static NOVAL: char = '\0';
 
@@ -51,7 +50,7 @@ fn try_phf_table(
         .collect();
 
     let table_len = hashes.len();
-    let buckets_len = (table_len + lambda - 1) / lambda;
+    let buckets_len = table_len.div_ceil(lambda);
 
     // group the elements into buckets of lambda (on average, for a
     // good hash) based on the suffix of their hash.
@@ -67,8 +66,8 @@ fn try_phf_table(
     // value for `foo` is "just" `map[displace(hash(foo))]`, where
     // `displace` uses the pair of displacements that we computed
     // (stored in `disps`).
-    let mut map = repeat(NOVAL).take(table_len).collect::<Vec<_>>();
-    let mut disps = repeat((0, 0)).take(buckets_len).collect::<Vec<_>>();
+    let mut map: Vec<_> = std::iter::repeat_n(NOVAL, table_len).collect();
+    let mut disps: Vec<_> = std::iter::repeat_n((0, 0), buckets_len).collect();
 
     // the set of index -> value mappings for the next bucket to be
     // placed; we need it separate because it may not work, so we may
@@ -80,7 +79,7 @@ fn try_phf_table(
     // having to clear the whole the whole array each time (just
     // compare against the generation). A u64 won't overflow.
     let mut generation = 0;
-    let mut try_map = repeat(0u64).take(table_len).collect::<Vec<_>>();
+    let mut try_map: Vec<_> = std::iter::repeat_n(0u64, table_len).collect();
     // the placed (index, codepoint) pairs of the current bucket, to
     // be placed into the main map if the whole bucket fits.
     let mut values_to_add = vec![];
@@ -145,25 +144,12 @@ pub fn create_phf(
     max_tries: usize,
 ) -> (u64, Vec<(u32, u32)>, Vec<char>) {
     let mut rng = StdRng::seed_from_u64(0xf0f0f0f0);
-    #[cfg(feature = "timing")]
-    let start = time::Instant::now();
 
     for i in 0..(max_tries) {
-        #[cfg(feature = "timing")]
-        let my_start = time::Instant::now();
-        #[cfg(feature = "timing")]
-        println!("PHF #{}: starting {:.2}", i, my_start - start);
-        #[cfg(not(feature = "timing"))]
         println!("PHF #{}", i);
 
-        let seed = rng.gen();
+        let seed = rng.random();
         if let Some((disp, map)) = try_phf_table(data, lambda, seed, &mut rng) {
-            #[cfg(feature = "timing")]
-            println!(
-                "PHF took: total {:.2} s, successive {:.2} s",
-                start.elapsed(),
-                my_start.elapsed()
-            );
             return (seed, disp, map);
         }
     }