Merge pull request #2833 from ehuss/static-regex

Add a helper for defining a regex

Author: ehuss

crates/mdbook-core/src/utils/mod.rs

@@ -1,11 +1,9 @@
 //! Various helpers and utilities.
 
 use anyhow::Error;
-use regex::Regex;
 use std::borrow::Cow;
 use std::collections::HashMap;
 use std::fmt::Write;
-use std::sync::LazyLock;
 use tracing::error;
 
 pub mod fs;
@@ -19,10 +17,23 @@ pub use self::string::{
     take_rustdoc_include_lines,
 };
 
+/// Defines a `static` with a [`regex::Regex`].
+#[macro_export]
+macro_rules! static_regex {
+    ($name:ident, $regex:literal) => {
+        static $name: std::sync::LazyLock<regex::Regex> =
+            std::sync::LazyLock::new(|| regex::Regex::new($regex).unwrap());
+    };
+    ($name:ident, bytes, $regex:literal) => {
+        static $name: std::sync::LazyLock<regex::bytes::Regex> =
+            std::sync::LazyLock::new(|| regex::bytes::Regex::new($regex).unwrap());
+    };
+}
+
 /// Replaces multiple consecutive whitespace characters with a single space character.
 pub fn collapse_whitespace(text: &str) -> Cow<'_, str> {
-    static RE: LazyLock<Regex> = LazyLock::new(|| Regex::new(r"\s\s+").unwrap());
-    RE.replace_all(text, " ")
+    static_regex!(WS, r"\s\s+");
+    WS.replace_all(text, " ")
 }
 
 /// Convert the given string to a valid HTML element ID.
@@ -48,7 +59,7 @@ fn id_from_content(content: &str) -> String {
     let mut content = content.to_string();
 
     // Skip any tags or html-encoded stuff
-    static HTML: LazyLock<Regex> = LazyLock::new(|| Regex::new(r"(<.*?>)").unwrap());
+    static_regex!(HTML, r"(<.*?>)");
     content = HTML.replace_all(&content, "").into();
     const REPL_SUB: &[&str] = &["&lt;", "&gt;", "&amp;", "&#39;", "&quot;"];
     for sub in REPL_SUB {

crates/mdbook-core/src/utils/string.rs

@@ -1,7 +1,6 @@
-use regex::Regex;
+use crate::static_regex;
 use std::ops::Bound::{Excluded, Included, Unbounded};
 use std::ops::RangeBounds;
-use std::sync::LazyLock;
 
 /// Take a range of lines from a string.
 pub fn take_lines<R: RangeBounds<usize>>(s: &str, range: R) -> String {
@@ -24,10 +23,8 @@ pub fn take_lines<R: RangeBounds<usize>>(s: &str, range: R) -> String {
     }
 }
 
-static ANCHOR_START: LazyLock<Regex> =
-    LazyLock::new(|| Regex::new(r"ANCHOR:\s*(?P<anchor_name>[\w_-]+)").unwrap());
-static ANCHOR_END: LazyLock<Regex> =
-    LazyLock::new(|| Regex::new(r"ANCHOR_END:\s*(?P<anchor_name>[\w_-]+)").unwrap());
+static_regex!(ANCHOR_START, r"ANCHOR:\s*(?P<anchor_name>[\w_-]+)");
+static_regex!(ANCHOR_END, r"ANCHOR_END:\s*(?P<anchor_name>[\w_-]+)");
 
 /// Take anchored lines from a string.
 /// Lines containing anchor are ignored.

crates/mdbook-driver/src/builtin_preprocessors/index.rs

@@ -1,8 +1,8 @@
 use anyhow::Result;
 use mdbook_core::book::{Book, BookItem};
+use mdbook_core::static_regex;
 use mdbook_preprocessor::{Preprocessor, PreprocessorContext};
-use regex::Regex;
-use std::{path::Path, sync::LazyLock};
+use std::path::Path;
 use tracing::warn;
 
 /// A preprocessor for converting file name `README.md` to `index.md` since
@@ -68,9 +68,9 @@ fn warn_readme_name_conflict<P: AsRef<Path>>(readme_path: P, index_path: P) {
 }
 
 fn is_readme_file<P: AsRef<Path>>(path: P) -> bool {
-    static RE: LazyLock<Regex> = LazyLock::new(|| Regex::new(r"(?i)^readme$").unwrap());
+    static_regex!(README, r"(?i)^readme$");
 
-    RE.is_match(
+    README.is_match(
         path.as_ref()
             .file_stem()
             .and_then(std::ffi::OsStr::to_str)

crates/mdbook-driver/src/builtin_preprocessors/links.rs

@@ -1,15 +1,15 @@
 use anyhow::{Context, Result};
 use mdbook_core::book::{Book, BookItem};
+use mdbook_core::static_regex;
 use mdbook_core::utils::{
     take_anchored_lines, take_lines, take_rustdoc_include_anchored_lines,
     take_rustdoc_include_lines,
 };
 use mdbook_preprocessor::{Preprocessor, PreprocessorContext};
-use regex::{CaptureMatches, Captures, Regex};
+use regex::{CaptureMatches, Captures};
 use std::fs;
 use std::ops::{Bound, Range, RangeBounds, RangeFrom, RangeFull, RangeTo};
 use std::path::{Path, PathBuf};
-use std::sync::LazyLock;
 use tracing::{error, warn};
 
 const ESCAPE_CHAR: char = '\\';
@@ -408,23 +408,19 @@ impl<'a> Iterator for LinkIter<'a> {
 }
 
 fn find_links(contents: &str) -> LinkIter<'_> {
-    // lazily compute following regex
-    // r"\\\{\{#.*\}\}|\{\{#([a-zA-Z0-9]+)\s*([^}]+)\}\}")?;
-    static RE: LazyLock<Regex> = LazyLock::new(|| {
-        Regex::new(
-            r"(?x)              # insignificant whitespace mode
+    static_regex!(
+        LINK,
+        r"(?x)              # insignificant whitespace mode
         \\\{\{\#.*\}\}      # match escaped link
         |                   # or
         \{\{\s*             # link opening parens and whitespace
         \#([a-zA-Z0-9_]+)   # link type
         \s+                 # separating whitespace
         ([^}]+)             # link target path and space separated properties
-        \}\}                # link closing parens",
-        )
-        .unwrap()
-    });
+        \}\}                # link closing parens"
+    );
 
-    LinkIter(RE.captures_iter(contents))
+    LinkIter(LINK.captures_iter(contents))
 }
 
 #[cfg(test)]

crates/mdbook-html/src/html_handlebars/hbs_renderer.rs

@@ -5,18 +5,17 @@ use anyhow::{Context, Result, bail};
 use handlebars::Handlebars;
 use mdbook_core::book::{Book, BookItem, Chapter};
 use mdbook_core::config::{BookConfig, Code, Config, HtmlConfig, Playground, RustEdition};
-use mdbook_core::utils;
 use mdbook_core::utils::fs::get_404_output_file;
+use mdbook_core::{static_regex, utils};
 use mdbook_markdown::render_markdown;
 use mdbook_renderer::{RenderContext, Renderer};
-use regex::{Captures, Regex};
+use regex::Captures;
 use serde_json::json;
 use std::borrow::Cow;
 use std::collections::BTreeMap;
 use std::collections::HashMap;
 use std::fs::{self, File};
 use std::path::{Path, PathBuf};
-use std::sync::LazyLock;
 use tracing::error;
 use tracing::{debug, info, trace, warn};
 
@@ -702,9 +701,10 @@ fn make_data(
 /// Goes through the rendered HTML, making sure all header tags have
 /// an anchor respectively so people can link to sections directly.
 fn build_header_links(html: &str) -> String {
-    static BUILD_HEADER_LINKS: LazyLock<Regex> = LazyLock::new(|| {
-        Regex::new(r#"<h(\d)(?: id="([^"]+)")?(?: class="([^"]+)")?>(.*?)</h\d>"#).unwrap()
-    });
+    static_regex!(
+        BUILD_HEADER_LINKS,
+        r#"<h(\d)(?: id="([^"]+)")?(?: class="([^"]+)")?>(.*?)</h\d>"#
+    );
     static IGNORE_CLASS: &[&str] = &["menu-title", "mdbook-help-title"];
 
     let mut id_counter = HashMap::new();
@@ -758,8 +758,8 @@ fn insert_link_into_header(
 fn convert_fontawesome(html: &str) -> String {
     use font_awesome_as_a_crate as fa;
 
-    let regex = Regex::new(r##"<i([^>]+)class="([^"]+)"([^>]*)></i>"##).unwrap();
-    regex
+    static_regex!(FA_RE, r#"<i([^>]+)class="([^"]+)"([^>]*)></i>"#);
+    FA_RE
         .replace_all(html, |caps: &Captures<'_>| {
             let text = &caps[0];
             let before = &caps[1];
@@ -811,8 +811,7 @@ fn convert_fontawesome(html: &str) -> String {
 // ```
 // This function replaces all commas by spaces in the code block classes
 fn fix_code_blocks(html: &str) -> String {
-    static FIX_CODE_BLOCKS: LazyLock<Regex> =
-        LazyLock::new(|| Regex::new(r##"<code([^>]+)class="([^"]+)"([^>]*)>"##).unwrap());
+    static_regex!(FIX_CODE_BLOCKS, r#"<code([^>]+)class="([^"]+)"([^>]*)>"#);
 
     FIX_CODE_BLOCKS
         .replace_all(html, |caps: &Captures<'_>| {
@@ -825,8 +824,10 @@ fn fix_code_blocks(html: &str) -> String {
         .into_owned()
 }
 
-static CODE_BLOCK_RE: LazyLock<Regex> =
-    LazyLock::new(|| Regex::new(r##"((?s)<code[^>]?class="([^"]+)".*?>(.*?)</code>)"##).unwrap());
+static_regex!(
+    CODE_BLOCK_RE,
+    r#"((?s)<code[^>]?class="([^"]+)".*?>(.*?)</code>)"#
+);
 
 fn add_playground_pre(
     html: &str,
@@ -895,10 +896,8 @@ fn add_playground_pre(
 /// Modifies all `<code>` blocks to convert "hidden" lines and to wrap them in
 /// a `<span class="boring">`.
 fn hide_lines(html: &str, code_config: &Code) -> String {
-    static LANGUAGE_REGEX: LazyLock<Regex> =
-        LazyLock::new(|| Regex::new(r"\blanguage-(\w+)\b").unwrap());
-    static HIDELINES_REGEX: LazyLock<Regex> =
-        LazyLock::new(|| Regex::new(r"\bhidelines=(\S+)").unwrap());
+    static_regex!(LANGUAGE_REGEX, r"\blanguage-(\w+)\b");
+    static_regex!(HIDELINES_REGEX, r"\bhidelines=(\S+)");
 
     CODE_BLOCK_RE
         .replace_all(html, |caps: &Captures<'_>| {
@@ -939,8 +938,7 @@ fn hide_lines(html: &str, code_config: &Code) -> String {
 }
 
 fn hide_lines_rust(content: &str) -> String {
-    static BORING_LINES_REGEX: LazyLock<Regex> =
-        LazyLock::new(|| Regex::new(r"^(\s*)#(.?)(.*)$").unwrap());
+    static_regex!(BORING_LINES_REGEX, r"^(\s*)#(.?)(.*)$");
 
     let mut result = String::with_capacity(content.len());
     let mut lines = content.lines().peekable();

crates/mdbook-html/src/html_handlebars/static_files.rs

@@ -4,12 +4,12 @@ use super::helpers::resources::ResourceHelper;
 use crate::theme::{self, Theme, playground_editor};
 use anyhow::{Context, Result};
 use mdbook_core::config::HtmlConfig;
+use mdbook_core::static_regex;
 use mdbook_core::utils;
 use std::borrow::Cow;
 use std::collections::HashMap;
 use std::fs::{self, File};
 use std::path::{Path, PathBuf};
-use std::sync::LazyLock;
 use tracing::debug;
 
 /// Map static files to their final names and contents.
@@ -191,11 +191,10 @@ impl StaticFiles {
 
     pub(super) fn write_files(self, destination: &Path) -> Result<ResourceHelper> {
         use mdbook_core::utils::fs::write_file;
-        use regex::bytes::{Captures, Regex};
+        use regex::bytes::Captures;
         // The `{{ resource "name" }}` directive in static resources look like
         // handlebars syntax, even if they technically aren't.
-        static RESOURCE: LazyLock<Regex> =
-            LazyLock::new(|| Regex::new(r#"\{\{ resource "([^"]+)" \}\}"#).unwrap());
+        static_regex!(RESOURCE, bytes, r#"\{\{ resource "([^"]+)" \}\}"#);
         fn replace_all<'a>(
             hash_map: &HashMap<String, String>,
             data: &'a [u8],