From 50a26a6cfa7d2e10257c9215bd865ac5ebbc6a4c Mon Sep 17 00:00:00 2001
From: Oliver Wooding <oliverwooding@icloud.com>
Date: Sat, 14 Mar 2026 13:46:09 +0100
Subject: [PATCH 1/2] fix: avoid unnecessary allocation in preprocess for clean
 inputs

Return Cow<str> instead of String so that typical source files (no BOM,
no CRLF, no shebang) take a zero-allocation fast path. BOM removal also
avoids allocation since it is a simple slice operation.

Closes #19

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---
 crates/gnomon-parser/src/preprocess.rs | 67 +++++++++++++++++++++++---
 1 file changed, 59 insertions(+), 8 deletions(-)

diff --git a/crates/gnomon-parser/src/preprocess.rs b/crates/gnomon-parser/src/preprocess.rs
index 0212151..59d766c 100644
--- a/crates/gnomon-parser/src/preprocess.rs
+++ b/crates/gnomon-parser/src/preprocess.rs
@@ -6,21 +6,44 @@
 /// 1. BOM removal
 /// 2. CRLF normalization
 /// 3. Shebang removal
-pub fn preprocess(input: &str) -> String {
+///
+/// Returns `Cow::Borrowed` when no transformation is needed, avoiding
+/// allocation for typical source files (no BOM, no CRLF, no shebang).
+pub fn preprocess(input: &str) -> std::borrow::Cow<'_, str> {
+    use std::borrow::Cow;
+
+    let has_bom = input.starts_with('\u{FEFF}');
+    let has_crlf = input.contains("\r\n");
+    let after_bom = if has_bom {
+        &input['\u{FEFF}'.len_utf8()..]
+    } else {
+        input
+    };
+    let has_shebang = after_bom.starts_with("#!");
+
+    // Fast path: no transformations needed.
+    if !has_bom && !has_crlf && !has_shebang {
+        return Cow::Borrowed(input);
+    }
+
     // r[impl lexer.input-format.bom-removal]
-    let input = input.strip_prefix('\u{FEFF}').unwrap_or(input);
+    // (already computed as `after_bom`)
 
     // r[impl lexer.input-format.crlf-normalization]
-    let input = input.replace("\r\n", "\n");
+    let normalized: Cow<'_, str> = if has_crlf {
+        Cow::Owned(after_bom.replace("\r\n", "\n"))
+    } else {
+        Cow::Borrowed(after_bom)
+    };
 
     // r[impl lexer.input-format.shebang-removal]
-    if input.starts_with("#!") {
-        match input.find('\n') {
-            Some(pos) => input[pos + 1..].to_string(),
-            None => String::new(),
+    if normalized.starts_with("#!") {
+        match normalized.find('\n') {
+            Some(pos) => Cow::Owned(normalized[pos + 1..].to_string()),
+            None => Cow::Owned(String::new()),
         }
     } else {
-        input
+        normalized
     }
 }
 
@@ -33,6 +56,34 @@ mod tests {
         assert_eq!(preprocess("hello world"), "hello world");
     }
 
+    #[test]
+    fn no_changes_returns_borrowed() {
+        let input = "hello world";
+        assert!(matches!(preprocess(input), std::borrow::Cow::Borrowed(_)));
+    }
+
+    #[test]
+    fn empty_input_returns_borrowed() {
+        assert!(matches!(preprocess(""), std::borrow::Cow::Borrowed(_)));
+    }
+
+    #[test]
+    fn bom_only_returns_borrowed_slice() {
+        // BOM removal is a slice, not an allocation, so still Borrowed.
+        assert!(matches!(
+            preprocess("\u{FEFF}hello"),
+            std::borrow::Cow::Borrowed(_)
+        ));
+    }
+
+    #[test]
+    fn crlf_returns_owned() {
+        assert!(matches!(
+            preprocess("a\r\nb"),
+            std::borrow::Cow::Owned(_)
+        ));
+    }
+
     // r[verify lexer.input-format.bom-removal]
     #[test]
     fn bom_removal() {

From 174e33a3af32dd9d9666bac05fa7f8fc4d3998db Mon Sep 17 00:00:00 2001
From: Oliver Wooding <oliverwooding@icloud.com>
Date: Sat, 14 Mar 2026 13:47:50 +0100
Subject: [PATCH 2/2] formatting

---
 crates/gnomon-parser/src/preprocess.rs | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/crates/gnomon-parser/src/preprocess.rs b/crates/gnomon-parser/src/preprocess.rs
index 59d766c..715f448 100644
--- a/crates/gnomon-parser/src/preprocess.rs
+++ b/crates/gnomon-parser/src/preprocess.rs
@@ -78,10 +78,7 @@ mod tests {
 
     #[test]
     fn crlf_returns_owned() {
-        assert!(matches!(
-            preprocess("a\r\nb"),
-            std::borrow::Cow::Owned(_)
-        ));
+        assert!(matches!(preprocess("a\r\nb"), std::borrow::Cow::Owned(_)));
     }
 
     // r[verify lexer.input-format.bom-removal]