fix(apply-patch): skip initial new-file hunk header emitted by some AI models

codex-cli/src/utils/agent/apply-patch.ts

@@ -307,6 +307,9 @@ class Parser {
 
   private parse_add_file(): PatchAction {
     const lines: Array<string> = [];
+    // Only skip the first new-file hunk header emitted by some AI models (e.g., Gemini 2.0 Flash)
+    const NEW_FILE_HUNK = /^@@\s*-0,0\s*\+1(?:,[1-9]\d*)?\s*@@(?: .*)?$/;
+    let isFirst = true;
     while (
       !this.is_done([
         PATCH_SUFFIX,
@@ -316,6 +319,12 @@ class Parser {
       ])
     ) {
       const s = this.read_str();
+      // Only skip the initial new-file hunk header once
+      if (isFirst && NEW_FILE_HUNK.test(s.trim())) {
+        isFirst = false;
+        continue;
+      }
+      isFirst = false;
       if (!s.startsWith(HUNK_ADD_LINE_PREFIX)) {
         throw new DiffError(`Invalid Add File Line: ${s}`);
       }

codex-cli/tests/apply-patch.test.ts

@@ -98,6 +98,42 @@ test("process_patch - add file", () => {
   expect(fs.removals).toEqual([]);
 });
 
+test("process_patch - add file skips full hunk headers", () => {
+  const patch = `*** Begin Patch
+*** Add File: z.txt
+@@ -0,0 +1,2 @@ SectionName
++foo
++bar
+*** End Patch`;
+  const fs = createInMemoryFS({});
+  process_patch(patch, fs.openFn, fs.writeFn, fs.removeFn);
+  expect(fs.writes).toEqual({ "z.txt": "foo\nbar" });
+  expect(fs.removals).toEqual([]);
+});
+
+// Regex self-test for new-file hunk headers
+test("NEW_FILE_HUNK regex matches expected patterns and rejects others", () => {
+  const re = /^@@\s*-0,0\s*\+1(?:,[1-9]\d*)?\s*@@(?: .*)?$/;
+  const matching = [
+    "@@ -0,0 +1@@",
+    "@@ -0,0 +1 @@",
+    "@@ -0,0 +1,10 @@",
+    "@@ -0,0 +1,10@@",
+    "@@-0,0+1@@",
+    "@@-0,0 +1@@",
+    "@@ -0,0+1,2@@",
+    "@@ -0,0 +1,2@@ foo",
+    "@@ -0,0 +1,2 @@ foo",
+  ];
+  const nonmatching = [
+    "@@ -1,1 +2,2 @@",
+    "@@ -0,0 +2,2 @@",
+    "@@ -0,0 +1,2 @@@@",
+  ];
+  matching.forEach((s) => expect(re.test(s)).toBe(true));
+  nonmatching.forEach((s) => expect(re.test(s)).toBe(false));
+});
+
 test("process_patch - delete file", () => {
   const patch = `*** Begin Patch
 *** Delete File: c.txt

codex-rs/apply-patch/src/lib.rs

@@ -665,6 +665,21 @@ PATCH"#,
         let contents = fs::read_to_string(path).unwrap();
         assert_eq!(contents, "ab\ncd\n");
     }
+
+    #[test]
+    fn test_add_file_skip_full_hunk_header_with_section() {
+        let dir = tempdir().unwrap();
+        let path = dir.path().join("z.txt");
+        let patch = wrap_patch(&format!(
+            "*** Add File: {}\n@@ -0,0 +1,2 @@ SectionName\n+foo\n+bar",
+            path.display()
+        ));
+        let mut stdout = Vec::new();
+        let mut stderr = Vec::new();
+        apply_patch(&patch, &mut stdout, &mut stderr).unwrap();
+        let contents = fs::read_to_string(&path).unwrap();
+        assert_eq!(contents, "foo\nbar\n");
+    }
 
     #[test]
     fn test_delete_file_hunk_removes_file() {

codex-rs/apply-patch/src/parser.rs

@@ -25,6 +25,7 @@
 use std::path::PathBuf;
 
 use thiserror::Error;
+use regex::Regex;
 
 const BEGIN_PATCH_MARKER: &str = "*** Begin Patch";
 const END_PATCH_MARKER: &str = "*** End Patch";
@@ -113,10 +114,20 @@ fn parse_one_hunk(lines: &[&str], line_number: usize) -> Result<(Hunk, usize), P
     // Be tolerant of case mismatches and extra padding around marker strings.
     let first_line = lines[0].trim();
     if let Some(path) = first_line.strip_prefix(ADD_FILE_MARKER) {
-        // Add File
+        // Add File: only skip the first new-file hunk header emitted by some AI models (e.g., Gemini 2.0 Flash)
         let mut contents = String::new();
         let mut parsed_lines = 1;
+        // Regex to match new-file hunk header with optional section name
+        let new_file_re = Regex::new(r"^@@\s*-0,0\s*\+1(?:,[1-9]\d*)?\s*@@(?: .*)?$").unwrap();
+        let mut first = true;
         for add_line in &lines[1..] {
+            // Only skip the initial new-file hunk header once
+            if first && new_file_re.is_match(add_line.trim()) {
+                parsed_lines += 1;
+                first = false;
+                continue;
+            }
+            first = false;
             if let Some(line_to_add) = add_line.strip_prefix('+') {
                 contents.push_str(line_to_add);
                 contents.push('\n');
@@ -497,3 +508,31 @@ fn test_update_file_chunk() {
         ))
     );
 }
+
+// Regex self-test for new-file hunk headers
+#[test]
+fn test_new_file_hunk_regex() {
+    let re = Regex::new(r"^@@\s*-0,0\s*\+1(?:,[1-9]\d*)?\s*@@(?: .*)?$").unwrap();
+    let matching = vec![
+        "@@ -0,0 +1@@",
+        "@@ -0,0 +1 @@",
+        "@@ -0,0 +1,10 @@",
+        "@@ -0,0 +1,10@@",
+        "@@-0,0+1@@",
+        "@@-0,0 +1@@",
+        "@@ -0,0+1,2@@",
+        "@@ -0,0 +1,2@@ foo",
+        "@@ -0,0 +1,2 @@ foo",
+    ];
+    let nonmatching = vec![
+        "@@ -1,1 +2,2 @@",
+        "@@ -0,0 +2,2 @@",
+        "@@ -0,0 +1,2 @@@@",
+    ];
+    for s in matching {
+        assert!(re.is_match(s), "should match: {}", s);
+    }
+    for s in nonmatching {
+        assert!(!re.is_match(s), "should not match: {}", s);
+    }
+}