re-use Regex instances

Author: tshepang

ci/sembr/src/main.rs

@@ -1,4 +1,5 @@
 use std::path::PathBuf;
+use std::sync::LazyLock;
 use std::{fs, process};
 
 use anyhow::Result;
@@ -15,6 +16,9 @@ struct Cli {
     line_length_limit: usize,
 }
 
+static REGEX_SPLIT: LazyLock<Regex> = LazyLock::new(|| Regex::new(r"(\.|\?|;|!)\s+").unwrap());
+static REGEX_IGNORE: LazyLock<Regex> = LazyLock::new(|| Regex::new(r"(\d\.|\-|\*)\s+").unwrap());
+
 fn main() -> Result<()> {
     let cli = Cli::parse();
     let mut compliant = Vec::new();
@@ -83,8 +87,6 @@ fn comply(content: &str) -> Result<String> {
     let content: Vec<_> = content.lines().map(std::borrow::ToOwned::to_owned).collect();
     let mut new_content = content.clone();
     let mut new_n = 0;
-    let split_re = Regex::new(r"(\.|\?|;|!)\s+")?;
-    let ignore_re = Regex::new(r"(\d\.|\-|\*)\s+").unwrap();
     let mut in_code_block = false;
     for (n, line) in content.into_iter().enumerate() {
         if n != 0 {
@@ -93,13 +95,13 @@ fn comply(content: &str) -> Result<String> {
         if line.trim().starts_with("```") {
             in_code_block = !in_code_block;
         }
-        if ignore(&line, in_code_block, &ignore_re) {
+        if ignore(&line, in_code_block, &REGEX_IGNORE) {
             continue;
         }
-        if split_re.is_match(&line) {
+        if REGEX_SPLIT.is_match(&line) {
             let indent = line.find(|ch: char| !ch.is_whitespace()).unwrap();
             let new_lines: Vec<_> = line
-                .split_inclusive(&split_re)
+                .split_inclusive(&*REGEX_SPLIT)
                 .map(|portion| format!("{:indent$}{}", "", portion.trim()))
                 .collect();
             new_content.splice(new_n..=new_n, new_lines.clone());
@@ -113,8 +115,6 @@ fn reduce_overlong_lines(content: &str, limit: usize) -> Result<String> {
     let content: Vec<_> = content.lines().map(std::borrow::ToOwned::to_owned).collect();
     let mut new_content = content.clone();
     let mut new_n = 0;
-    let ignore_re = Regex::new(r"(\d\.|\-|\*)\s+").unwrap();
-    let split_re = Regex::new(r"(\.|\?|;|!)\s+")?;
     let mut in_code_block = false;
     let mut skip_next = false;
     for (n, line) in content.iter().enumerate() {
@@ -128,13 +128,13 @@ fn reduce_overlong_lines(content: &str, limit: usize) -> Result<String> {
         if line.trim().starts_with("```") {
             in_code_block = !in_code_block;
         }
-        if ignore(line, in_code_block, &ignore_re) || split_re.is_match(line) {
+        if ignore(line, in_code_block, &REGEX_IGNORE) || REGEX_SPLIT.is_match(line) {
             continue;
         }
         let Some(next_line) = content.get(n + 1) else {
             continue;
         };
-        if ignore(next_line, in_code_block, &ignore_re) || line.trim().ends_with('.') {
+        if ignore(next_line, in_code_block, &REGEX_IGNORE) || line.trim().ends_with('.') {
             continue;
         }
         if line.len() + next_line.len() < limit {