From 9e5c5648299cb66fd68fa237b56b98c99e0dbc36 Mon Sep 17 00:00:00 2001
From: Chris Bookholt <bookholt@openai.com>
Date: Wed, 1 Jul 2026 16:18:03 -0700
Subject: [PATCH] git-utils: serialize Git error paths safely

---
 codex-rs/core/src/turn_metadata_tests.rs | 58 ++++++++++++++++++++++++
 codex-rs/git-utils/src/errors.rs         | 33 +++++++++++++-
 2 files changed, 90 insertions(+), 1 deletion(-)

diff --git a/codex-rs/core/src/turn_metadata_tests.rs b/codex-rs/core/src/turn_metadata_tests.rs
index be49f1b72c..39cf2bb8c6 100644
--- a/codex-rs/core/src/turn_metadata_tests.rs
+++ b/codex-rs/core/src/turn_metadata_tests.rs
@@ -69,6 +69,64 @@ fn workspace_metadata_serializes_structured_has_changes_refusal() {
     );
 }
 
+#[cfg(any(unix, windows))]
+#[test]
+fn turn_metadata_keeps_workspace_failure_with_non_utf8_repository_path() {
+    use std::ffi::OsString;
+    #[cfg(unix)]
+    use std::os::unix::ffi::OsStringExt;
+    #[cfg(windows)]
+    use std::os::windows::ffi::OsStringExt;
+
+    let temp_dir = TempDir::new().expect("temp dir");
+    #[cfg(unix)]
+    let path = std::path::PathBuf::from(OsString::from_vec(b"/repo/\xff".to_vec()));
+    #[cfg(windows)]
+    let path = std::path::PathBuf::from(OsString::from_wide(&[
+        b'C' as u16,
+        b':' as u16,
+        b'\\' as u16,
+        0xd800,
+    ]));
+    let lossy = path.to_string_lossy().into_owned();
+    let (has_changes, has_changes_unavailable_reason) =
+        split_has_changes(Err(codex_git_utils::GitReadError::NotRepository { path }));
+    let workspace = WorkspaceGitMetadata {
+        associated_remote_urls: None,
+        latest_git_commit_hash: None,
+        has_changes,
+        has_changes_unavailable_reason,
+    }
+    .into();
+    let state = TurnMetadataState::new(
+        "session-a".to_string(),
+        "thread-a".to_string(),
+        /*forked_from_thread_id*/ None,
+        /*parent_thread_id*/ None,
+        &SessionSource::Exec,
+        /*thread_source*/ None,
+        "turn-a".to_string(),
+        temp_dir.path().abs(),
+        &PermissionProfile::read_only(),
+        WindowsSandboxLevel::Disabled,
+        /*enforce_managed_network*/ false,
+    );
+    *state.enriched_workspaces.write().expect("workspace lock") =
+        Some(BTreeMap::from([("/repo".to_string(), workspace)]));
+
+    let metadata: Value = serde_json::from_str(&test_turn_metadata_header(&state))
+        .expect("full turn metadata remains serializable");
+    assert_eq!(
+        metadata["workspaces"]["/repo"],
+        serde_json::json!({
+            "has_changes_unavailable_reason": {
+                "reason": "notRepository",
+                "path": lossy,
+            }
+        })
+    );
+}
+
 fn test_responses_metadata_json(
     state: &TurnMetadataState,
     window_id: &str,
diff --git a/codex-rs/git-utils/src/errors.rs b/codex-rs/git-utils/src/errors.rs
index d22c57de25..4494c7e897 100644
--- a/codex-rs/git-utils/src/errors.rs
+++ b/codex-rs/git-utils/src/errors.rs
@@ -49,10 +49,15 @@ pub enum GitReadError {
     #[error("no trusted Git executable is available")]
     NoTrustedGit,
     #[error("{path:?} is not a Git repository")]
-    NotRepository { path: PathBuf },
+    NotRepository {
+        #[serde(with = "lossy_path")]
+        path: PathBuf,
+    },
     #[error("Git resolved {reported_root:?} instead of expected root {expected_root:?}")]
     RepositoryRootMismatch {
+        #[serde(with = "lossy_path")]
         expected_root: PathBuf,
+        #[serde(with = "lossy_path")]
         reported_root: PathBuf,
     },
     #[error("Git operation {operation:?} timed out")]
@@ -67,3 +72,29 @@ pub enum GitReadError {
     #[error("executable filter {driver:?} is selected for {path:?}")]
     SelectedExecutableFilter { driver: String, path: String },
 }
+
+/// Path fields in caller-visible diagnostic metadata use a deliberately lossy
+/// UTF-8 wire representation. The Rust API remains `PathBuf`; deserialization
+/// reconstructs the serialized path, not any replaced platform-native bytes.
+mod lossy_path {
+    use std::path::Path;
+    use std::path::PathBuf;
+
+    use serde::Deserialize;
+    use serde::Deserializer;
+    use serde::Serializer;
+
+    pub(super) fn serialize<S>(path: &Path, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: Serializer,
+    {
+        serializer.serialize_str(&path.to_string_lossy())
+    }
+
+    pub(super) fn deserialize<'de, D>(deserializer: D) -> Result<PathBuf, D::Error>
+    where
+        D: Deserializer<'de>,
+    {
+        String::deserialize(deserializer).map(PathBuf::from)
+    }
+}