Auto merge of rust-lang#137285 - yotamofek:pr/rustdoc/pulldown-escapi…

…ng, r=GuillaumeGomez

librustdoc: Use `pulldown-cmark-escape` for HTML escaping

Implementation of `@notriddle` 's [suggestion](rust-lang#137274 (comment)).
Somewhat related to rust-lang#137274 , but the two PRs should be complementary.

Local perf results look like a nice improvement! (so would love a perf run on the CI)

Cargo.lock

@@ -4571,6 +4571,7 @@ dependencies = [
  "itertools",
  "minifier",
  "pulldown-cmark 0.9.6",
+ "pulldown-cmark-escape",
  "regex",
  "rinja",
  "rustdoc-json-types",

src/librustdoc/Cargo.toml

@@ -15,6 +15,7 @@ itertools = "0.12"
 indexmap = "2"
 minifier = { version = "0.3.5", default-features = false }
 pulldown-cmark-old = { version = "0.9.6", package = "pulldown-cmark", default-features = false }
+pulldown-cmark-escape = { version = "0.11.0", features = ["simd"] }
 regex = "1"
 rustdoc-json-types = { path = "../rustdoc-json-types" }
 serde_json = "1.0"

src/librustdoc/html/escape.rs

@@ -5,6 +5,7 @@
 
 use std::fmt;
 
+use pulldown_cmark_escape::FmtWriter;
 use unicode_segmentation::UnicodeSegmentation;
 
 /// Wrapper struct which will emit the HTML-escaped version of the contained
@@ -13,31 +14,7 @@ pub(crate) struct Escape<'a>(pub &'a str);
 
 impl fmt::Display for Escape<'_> {
     fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result {
-        // Because the internet is always right, turns out there's not that many
-        // characters to escape: http://stackoverflow.com/questions/7381974
-        let Escape(s) = *self;
-        let pile_o_bits = s;
-        let mut last = 0;
-        for (i, ch) in s.char_indices() {
-            let s = match ch {
-                '>' => "&gt;",
-                '<' => "&lt;",
-                '&' => "&amp;",
-                '\'' => "&#39;",
-                '"' => "&quot;",
-                _ => continue,
-            };
-            fmt.write_str(&pile_o_bits[last..i])?;
-            fmt.write_str(s)?;
-            // NOTE: we only expect single byte characters here - which is fine as long as we
-            // only match single byte characters
-            last = i + 1;
-        }
-
-        if last < s.len() {
-            fmt.write_str(&pile_o_bits[last..])?;
-        }
-        Ok(())
+        pulldown_cmark_escape::escape_html(FmtWriter(fmt), self.0)
     }
 }
 
@@ -51,29 +28,7 @@ pub(crate) struct EscapeBodyText<'a>(pub &'a str);
 
 impl fmt::Display for EscapeBodyText<'_> {
     fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result {
-        // Because the internet is always right, turns out there's not that many
-        // characters to escape: http://stackoverflow.com/questions/7381974
-        let EscapeBodyText(s) = *self;
-        let pile_o_bits = s;
-        let mut last = 0;
-        for (i, ch) in s.char_indices() {
-            let s = match ch {
-                '>' => "&gt;",
-                '<' => "&lt;",
-                '&' => "&amp;",
-                _ => continue,
-            };
-            fmt.write_str(&pile_o_bits[last..i])?;
-            fmt.write_str(s)?;
-            // NOTE: we only expect single byte characters here - which is fine as long as we
-            // only match single byte characters
-            last = i + 1;
-        }
-
-        if last < s.len() {
-            fmt.write_str(&pile_o_bits[last..])?;
-        }
-        Ok(())
+        pulldown_cmark_escape::escape_html_body_text(FmtWriter(fmt), self.0)
     }
 }