Support Cloud image messages and editable draft handoffs (#60061)

## Why

Cloud messages previously accepted only plain text, and switching a draft to Cloud waited for session preparation without processing composer input.

## What changed

- Support ordered text and image parts in Cloud submissions and user-message summaries, preserving attachment placeholders, image detail, and client message IDs.
- Accept inline data URLs and uploaded file IDs, including image-only messages. Reject invalid references and text spans locally, and cap encoded image requests at 3.75 MiB while retaining the existing plain-text limit.
- Keep the composer editable and redraw the TUI while Cloud prepares a conversation. Carry the latest draft, attachments, and Vim mode into the destination, retain drafts on failure, and prevent submission to the outgoing session during the handoff.
- Preserve buffered paste input and pending clipboard reads across the handoff, with connection and failure hints.

## Testing

Add regression coverage for image wire encoding, receipt identity, validation and size boundaries, draft editing during successful and failed handoffs, clipboard completion, Vim state, and attachment recovery after rejected submissions.

GitOrigin-RevId: 0cbcde0ab89ba961780b5f88456057e6b6a19cd5

codex-rs/cloud-client/src/items.rs

@@ -1,6 +1,10 @@
 //! Selected item summaries for live start/completion events, without interpreting executor paths.
 use crate::FileUpdate;
+use crate::ImageDetail;
+use crate::ImageReference;
 use crate::TranscriptUpdate;
+use crate::UserMessageInput;
+use crate::UserMessageTextElement;
 use crate::item_wire as wire;
 use crate::transcript::DecodeFailure;
 use crate::transcript::ITEM_STARTED;
@@ -24,7 +28,7 @@ pub enum ItemSummary {
     UserMessage {
         id: String,
         client_id: Option<String>,
-        text: Vec<String>,
+        content: Vec<UserMessageInput>,
     },
     AgentMessage {
         id: String,
@@ -169,17 +173,62 @@ pub(crate) fn decode(
 pub(crate) fn summary(item: wire::Item) -> Result<Option<ItemSummary>, DecodeFailure> {
     let summary = match item.value {
         Some(wire::Value::UserMessage(value)) => {
-            let mut text = Vec::new();
+            let mut content = Vec::new();
             for input in required(value.content)?.values {
-                let Some(crate::turns_wire::Value::Text(input)) = input.value else {
-                    return Ok(None);
+                let input = match input.value {
+                    Some(crate::turns_wire::Value::Text(input)) => {
+                        let text = required(input.text)?;
+                        let text_elements = input
+                            .text_elements
+                            .map_or_else(Vec::new, |elements| elements.values)
+                            .into_iter()
+                            .map(|element| {
+                                let range = required(element.byte_range)?;
+                                let start = usize::try_from(required(range.start)?)
+                                    .map_err(|_| DecodeFailure::MissingField)?;
+                                let end = usize::try_from(required(range.end)?)
+                                    .map_err(|_| DecodeFailure::MissingField)?;
+                                text.get(start..end).ok_or(DecodeFailure::MissingField)?;
+                                Ok(UserMessageTextElement {
+                                    start,
+                                    end,
+                                    placeholder: element.placeholder,
+                                })
+                            })
+                            .collect::<Result<_, DecodeFailure>>()?;
+                        UserMessageInput::Text {
+                            text,
+                            text_elements,
+                        }
+                    }
+                    Some(crate::turns_wire::Value::Image(input)) => {
+                        let (image, detail) = match input.value {
+                            Some(crate::turns_wire::ImageValue::Url(image)) => {
+                                (ImageReference::Url(required(image.value)?), image.detail)
+                            }
+                            Some(crate::turns_wire::ImageValue::FileId(image)) => {
+                                (ImageReference::FileId(required(image.value)?), image.detail)
+                            }
+                            None => return Ok(None),
+                        };
+                        let detail = match detail {
+                            Some(1) => Some(ImageDetail::Auto),
+                            Some(2) => Some(ImageDetail::Low),
+                            Some(3) => Some(ImageDetail::High),
+                            Some(4) => Some(ImageDetail::Original),
+                            // Detail is display metadata. Keep the receipt identity if it is newer.
+                            None | Some(_) => None,
+                        };
+                        UserMessageInput::Image { image, detail }
+                    }
+                    _ => return Ok(None),
                 };
-                text.push(required(input.text)?);
+                content.push(input);
             }
             ItemSummary::UserMessage {
                 id: required(value.id)?,
                 client_id: value.client_id,
-                text,
+                content,
             }
         }
         Some(wire::Value::AgentMessage(value)) => ItemSummary::AgentMessage {