Persist pending input user events (#10656)

- Persist user-message events for mid-turn injected input by emitting
user message turn items when pending input is recorded.

codex-rs/core/src/codex.rs

@@ -3579,19 +3579,35 @@ pub(crate) async fn run_turn(
         // Note that pending_input would be something like a message the user
         // submitted through the UI while the model was running. Though the UI
         // may support this, the model might not.
-        let pending_input = sess
+        let pending_response_items = sess
             .get_pending_input()
             .await
             .into_iter()
             .map(ResponseItem::from)
             .collect::<Vec<ResponseItem>>();
 
+        if !pending_response_items.is_empty() {
+            for response_item in pending_response_items {
+                if let Some(TurnItem::UserMessage(user_message)) = parse_turn_item(&response_item) {
+                    // todo(aibrahim): move pending input to be UserInput only to keep TextElements. context: https://github.com/openai/codex/pull/10656#discussion_r2765522480
+                    sess.record_user_prompt_and_emit_turn_item(
+                        turn_context.as_ref(),
+                        &user_message.content,
+                        response_item,
+                    )
+                    .await;
+                } else {
+                    sess.record_conversation_items(
+                        &turn_context,
+                        std::slice::from_ref(&response_item),
+                    )
+                    .await;
+                }
+            }
+        }
+
         // Construct the input that we will send to the model.
-        let sampling_request_input: Vec<ResponseItem> = {
-            sess.record_conversation_items(&turn_context, &pending_input)
-                .await;
-            sess.clone_history().await.for_prompt()
-        };
+        let sampling_request_input: Vec<ResponseItem> = { sess.clone_history().await.for_prompt() };
 
         let sampling_request_input_messages = sampling_request_input
             .iter()

codex-rs/core/tests/suite/pending_input.rs

@@ -125,6 +125,11 @@ async fn injected_user_input_triggers_follow_up_request_with_deltas() {
 
     let _ = gate_completed_tx.send(());
 
+    let _ = wait_for_event(&codex, |event| {
+        matches!(event, EventMsg::UserMessage(message) if message.message == "second prompt")
+    })
+    .await;
+
     wait_for_event(&codex, |event| matches!(event, EventMsg::TurnComplete(_))).await;
 
     let requests = server.requests().await;