refactor: Drop once_cell in favor of OnceLock/LazyLock

Author: CosmicHorrorDev

CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## Unreleased
+
+Breaking changes
+
+- Replace `SCOPE_REPO` with the `scope_repo` function exposing the same functionality
+
+### Dependencies
+
+- Replace `once_cell` with `std::sync::{LazyLock, OnceLock}`
+
 ## [Version 5.2.0](https://github.com/trishume/syntect/compare/v5.1.0...v5.2.0) (2024-02-07)
 
 ### Improvements

Cargo.lock

@@ -31,7 +31,6 @@ fnv = { version = "1.0", optional = true }
 serde = "1.0"
 serde_derive = "1.0"
 serde_json = "1.0"
-once_cell = "1.8"
 thiserror = "1.0"
 
 [dev-dependencies]

Cargo.toml

@@ -17,10 +17,10 @@ use std::fs::File;
 use std::io::{BufRead, BufReader};
 use std::path::Path;
 use std::str::FromStr;
+use std::sync::LazyLock;
 use std::time::Instant;
 
 use getopts::Options;
-use once_cell::sync::Lazy;
 use regex::Regex;
 use walkdir::{DirEntry, WalkDir};
 
@@ -36,7 +36,7 @@ pub enum SyntaxTestFileResult {
     Success(usize),
 }
 
-pub static SYNTAX_TEST_HEADER_PATTERN: Lazy<Regex> = Lazy::new(|| {
+pub static SYNTAX_TEST_HEADER_PATTERN: LazyLock<Regex> = LazyLock::new(|| {
     Regex::new(
         r#"(?xm)
         ^(?P<testtoken_start>\s*\S+)
@@ -47,7 +47,7 @@ pub static SYNTAX_TEST_HEADER_PATTERN: Lazy<Regex> = Lazy::new(|| {
     )
     .unwrap()
 });
-pub static SYNTAX_TEST_ASSERTION_PATTERN: Lazy<Regex> = Lazy::new(|| {
+pub static SYNTAX_TEST_ASSERTION_PATTERN: LazyLock<Regex> = LazyLock::new(|| {
     Regex::new(
         r#"(?xm)
     \s*(?:

examples/syntest.rs

@@ -3,8 +3,8 @@ use crate::easy::{HighlightFile, HighlightLines};
 use crate::escape::Escape;
 use crate::highlighting::{Color, FontStyle, Style, Theme};
 use crate::parsing::{
-    BasicScopeStackOp, ParseState, Scope, ScopeStack, ScopeStackOp, SyntaxReference, SyntaxSet,
-    SCOPE_REPO,
+    scope_repo, BasicScopeStackOp, ParseState, Scope, ScopeStack, ScopeStackOp, SyntaxReference,
+    SyntaxSet,
 };
 use crate::util::LinesWithEndings;
 use crate::Error;
@@ -241,7 +241,7 @@ pub enum ClassStyle {
 }
 
 fn scope_to_classes(s: &mut String, scope: Scope, style: ClassStyle) {
-    let repo = SCOPE_REPO.lock().unwrap();
+    let repo = scope_repo().lock().unwrap();
     for i in 0..(scope.len()) {
         let atom = scope.atom_at(i as usize);
         let atom_s = repo.atom_str(atom);
@@ -259,7 +259,7 @@ fn scope_to_classes(s: &mut String, scope: Scope, style: ClassStyle) {
 }
 
 fn scope_to_selector(s: &mut String, scope: Scope, style: ClassStyle) {
-    let repo = SCOPE_REPO.lock().unwrap();
+    let repo = scope_repo().lock().unwrap();
     for i in 0..(scope.len()) {
         let atom = scope.atom_at(i as usize);
         let atom_s = repo.atom_str(atom);

src/html.rs

@@ -1,7 +1,7 @@
-use once_cell::sync::OnceCell;
 use serde::de::{Deserialize, Deserializer};
 use serde::ser::{Serialize, Serializer};
 use std::error::Error;
+use std::sync::OnceLock;
 
 /// An abstraction for regex patterns.
 ///
@@ -11,7 +11,7 @@ use std::error::Error;
 #[derive(Debug)]
 pub struct Regex {
     regex_str: String,
-    regex: OnceCell<regex_impl::Regex>,
+    regex: OnceLock<regex_impl::Regex>,
 }
 
 /// A region contains text positions for capture groups in a match result.
@@ -28,7 +28,7 @@ impl Regex {
     pub fn new(regex_str: String) -> Self {
         Self {
             regex_str,
-            regex: OnceCell::new(),
+            regex: OnceLock::new(),
         }
     }
 
@@ -76,7 +76,7 @@ impl Clone for Regex {
     fn clone(&self) -> Self {
         Regex {
             regex_str: self.regex_str.clone(),
-            regex: OnceCell::new(),
+            regex: OnceLock::new(),
         }
     }
 }

src/parsing/regex.rs

@@ -4,11 +4,10 @@ use std::collections::HashMap;
 use std::fmt;
 use std::mem;
 use std::str::FromStr;
-use std::sync::Mutex;
+use std::sync::{Mutex, OnceLock};
 use std::u16;
 use std::u64;
 
-use once_cell::sync::Lazy;
 use serde::de::{Deserialize, Deserializer, Error, Visitor};
 use serde::ser::{Serialize, Serializer};
 use serde_derive::{Deserialize, Serialize};
@@ -29,11 +28,13 @@ pub const ATOM_LEN_BITS: u16 = 3;
 
 /// The global scope repo, exposed in case you want to minimize locking and unlocking.
 ///
-/// Ths shouldn't be necessary for you to use. See the [`ScopeRepository`] docs.
+/// This shouldn't be necessary for you to use. See the [`ScopeRepository`] docs.
 ///
 /// [`ScopeRepository`]: struct.ScopeRepository.html
-pub static SCOPE_REPO: Lazy<Mutex<ScopeRepository>> =
-    Lazy::new(|| Mutex::new(ScopeRepository::new()));
+pub fn scope_repo() -> &'static Mutex<ScopeRepository> {
+    static SCOPE_REPO: OnceLock<Mutex<ScopeRepository>> = OnceLock::new();
+    SCOPE_REPO.get_or_init(|| Mutex::new(ScopeRepository::new()))
+}
 
 /// A hierarchy of atoms with semi-standardized names used to accord semantic information to a
 /// specific piece of text.
@@ -73,12 +74,11 @@ pub enum ParseScopeError {
 /// The structure used to keep track of the mapping between scope atom numbers and their string
 /// names
 ///
-/// It is only exposed in case you want to lock [`SCOPE_REPO`] and then allocate a bunch of scopes
+/// It is only exposed in case you want to lock [`scope_repo()`] and then allocate a bunch of scopes
 /// at once without thrashing the lock. In general, you should just use [`Scope::new()`].
 ///
 /// Only [`Scope`]s created by the same repository have valid comparison results.
 ///
-/// [`SCOPE_REPO`]: struct.SCOPE_REPO.html
 /// [`Scope::new()`]: struct.Scope.html#method.new
 /// [`Scope`]: struct.Scope.html
 #[derive(Debug)]
@@ -223,7 +223,7 @@ impl Scope {
     ///
     /// Example: `Scope::new("meta.rails.controller")`
     pub fn new(s: &str) -> Result<Scope, ParseScopeError> {
-        let mut repo = SCOPE_REPO.lock().unwrap();
+        let mut repo = scope_repo().lock().unwrap();
         repo.build(s.trim())
     }
 
@@ -268,7 +268,7 @@ impl Scope {
     ///
     /// This requires locking a global repo and shouldn't be done frequently.
     pub fn build_string(self) -> String {
-        let repo = SCOPE_REPO.lock().unwrap();
+        let repo = scope_repo().lock().unwrap();
         repo.to_string(self)
     }
 

src/parsing/scope.rs

@@ -13,10 +13,10 @@ use std::fs::File;
 use std::io::{self, BufRead, BufReader};
 use std::mem;
 use std::path::Path;
+use std::sync::OnceLock;
 
 use super::regex::Regex;
 use crate::parsing::syntax_definition::ContextId;
-use once_cell::sync::OnceCell;
 use serde_derive::{Deserialize, Serialize};
 
 /// A syntax set holds multiple syntaxes that have been linked together.
@@ -34,8 +34,8 @@ pub struct SyntaxSet {
     /// Stores the syntax index for every path that was loaded
     path_syntaxes: Vec<(String, usize)>,
 
-    #[serde(skip_serializing, skip_deserializing, default = "OnceCell::new")]
-    first_line_cache: OnceCell<FirstLineCache>,
+    #[serde(skip_serializing, skip_deserializing, default = "OnceLock::new")]
+    first_line_cache: OnceLock<FirstLineCache>,
     /// Metadata, e.g. indent and commenting information.
     ///
     /// NOTE: if serializing, you should handle metadata manually; that is, you should serialize and
@@ -58,7 +58,7 @@ pub struct SyntaxReference {
     #[serde(serialize_with = "ordered_map")]
     pub variables: HashMap<String, String>,
     #[serde(skip)]
-    pub(crate) lazy_contexts: OnceCell<LazyContexts>,
+    pub(crate) lazy_contexts: OnceLock<LazyContexts>,
     pub(crate) serialized_lazy_contexts: Vec<u8>,
 }
 
@@ -114,7 +114,7 @@ impl Clone for SyntaxSet {
             syntaxes: self.syntaxes.clone(),
             path_syntaxes: self.path_syntaxes.clone(),
             // Will need to be re-initialized
-            first_line_cache: OnceCell::new(),
+            first_line_cache: OnceLock::new(),
             #[cfg(feature = "metadata")]
             metadata: self.metadata.clone(),
         }
@@ -126,7 +126,7 @@ impl Default for SyntaxSet {
         SyntaxSet {
             syntaxes: Vec::new(),
             path_syntaxes: Vec::new(),
-            first_line_cache: OnceCell::new(),
+            first_line_cache: OnceLock::new(),
             #[cfg(feature = "metadata")]
             metadata: Metadata::default(),
         }
@@ -627,7 +627,7 @@ impl SyntaxSetBuilder {
                 first_line_match,
                 hidden,
                 variables,
-                lazy_contexts: OnceCell::new(),
+                lazy_contexts: OnceLock::new(),
                 serialized_lazy_contexts: Vec::new(), // initialized in the last step
             };
             syntaxes.push(syntax);
@@ -711,7 +711,7 @@ impl SyntaxSetBuilder {
         SyntaxSet {
             syntaxes,
             path_syntaxes,
-            first_line_cache: OnceCell::new(),
+            first_line_cache: OnceLock::new(),
             #[cfg(feature = "metadata")]
             metadata,
         }

src/parsing/syntax_set.rs

@@ -95,7 +95,7 @@ impl SyntaxDefinition {
             return Err(ParseSyntaxError::EmptyFile);
         }
         let doc = &docs[0];
-        let mut scope_repo = SCOPE_REPO.lock().unwrap();
+        let mut scope_repo = scope_repo().lock().unwrap();
         SyntaxDefinition::parse_top_level(
             doc,
             scope_repo.deref_mut(),

src/parsing/yaml_load.rs

@@ -1219,7 +1219,7 @@ impl core::marker::Unpin for syntect::parsing::SyntaxSetBuilder
 impl core::panic::unwind_safe::RefUnwindSafe for syntect::parsing::SyntaxSetBuilder
 impl core::panic::unwind_safe::UnwindSafe for syntect::parsing::SyntaxSetBuilder
 pub const syntect::parsing::ATOM_LEN_BITS: u16 = 3u16
-pub static syntect::parsing::SCOPE_REPO: once_cell::sync::Lazy<std::sync::mutex::Mutex<syntect::parsing::ScopeRepository>>
+pub fn syntect::parsing::scope_repo() -> &'static std::sync::mutex::Mutex<syntect::parsing::ScopeRepository>
 pub mod syntect::util
 pub struct syntect::util::LinesWithEndings<'a>
 impl<'a> syntect::util::LinesWithEndings<'a>