Files
codex/codex-rs/code-mode/src/runtime/mod.rs
T
Channing Conger e2f074e16c code-mode: move cell state into library actor (#28599)
A code-mode cell is a single JavaScript execution that can produce
output, call tools, wait for asynchronous work, resume, or be
terminated. This PR extracts the existing per-cell run loop into a
dedicated actor that owns the cell’s lifecycle state. It is primarily an
ownership change rather than a new lifecycle contract: existing behavior
now has one clear implementation boundary.

### Architecture
The session service remains responsible for session-wide concerns:
allocating cell IDs, storing shared values, creating cells, and routing
requests to them.

Once a cell is created, its execution state belongs to its actor.
Callers interact with the actor through a handle. The actor receives two
kinds of input: runtime events and control requests.

A single event loop serializes these inputs and applies the lifecycle
rules. It tracks the current observer—the caller waiting for an
update—along with accumulated output, outstanding callbacks, runtime
state, yield deadlines, and termination progress. Observation,
termination, completion, and cleanup therefore have one consistent
owner.

When the runtime has no immediately runnable work and is waiting only on
timers or tool results, the actor can return accumulated output and
information about outstanding tool calls while keeping the cell
available to resume. On completion or termination, it performs the
appropriate callback cleanup before publishing the final result and
removing the cell from the session.

A small host interface connects the actor to session-owned facilities
such as tool dispatch, notifications, stored values, and final cell
removal, keeping those responsibilities outside the actor itself.

### Why
Previously, cell lifecycle state and coordination lived alongside
session management. The actor boundary makes each cell a self-contained
state machine with a single writer, while the service becomes a registry
and adapter around it.

This makes lifecycle behavior easier to reason about and test in
isolation. It also establishes a clean boundary for later changing where
cells run or how they communicate without recreating their lifecycle
rules.
2026-06-16 19:28:55 -07:00

455 lines
14 KiB
Rust

mod callbacks;
mod globals;
mod module_loader;
mod timers;
mod value;
use std::collections::HashMap;
use std::sync::OnceLock;
use std::sync::mpsc as std_mpsc;
use std::thread;
use codex_code_mode_protocol::CodeModeToolKind;
use codex_code_mode_protocol::EnabledToolMetadata;
use codex_code_mode_protocol::ExecuteRequest;
use codex_code_mode_protocol::FunctionCallOutputContentItem;
use codex_code_mode_protocol::enabled_tool_metadata;
use codex_protocol::ToolName;
use serde_json::Value as JsonValue;
use tokio::sync::mpsc;
const EXIT_SENTINEL: &str = "__codex_code_mode_exit__";
#[derive(Debug)]
pub(crate) enum RuntimeCommand {
ToolResponse { id: String, result: JsonValue },
ToolError { id: String, error_text: String },
TimeoutFired { id: u64 },
ObservePendingFrontier,
Terminate,
}
#[derive(Clone, Copy, Debug, PartialEq)]
pub(crate) enum PendingRuntimeMode {
#[cfg(test)]
Continue,
PauseUntilResumed,
}
#[derive(Debug)]
pub(crate) enum RuntimeControlCommand {
Continue,
Resume,
Terminate,
}
#[derive(Debug)]
pub(crate) enum RuntimeEvent {
Started,
Pending,
ContentItem(FunctionCallOutputContentItem),
YieldRequested,
ToolCall {
id: String,
name: ToolName,
kind: CodeModeToolKind,
input: Option<JsonValue>,
},
Notify {
call_id: String,
text: String,
},
Result {
stored_value_writes: HashMap<String, JsonValue>,
error_text: Option<String>,
},
}
pub(crate) fn spawn_runtime(
stored_values: HashMap<String, JsonValue>,
request: ExecuteRequest,
event_tx: mpsc::UnboundedSender<RuntimeEvent>,
pending_mode: PendingRuntimeMode,
) -> Result<
(
std_mpsc::Sender<RuntimeCommand>,
std_mpsc::Sender<RuntimeControlCommand>,
v8::IsolateHandle,
),
String,
> {
initialize_v8()?;
let (command_tx, command_rx) = std_mpsc::channel();
let (control_tx, control_rx) = std_mpsc::channel();
let runtime_command_tx = command_tx.clone();
let (isolate_handle_tx, isolate_handle_rx) = std_mpsc::sync_channel(1);
let enabled_tools = request
.enabled_tools
.iter()
.map(enabled_tool_metadata)
.collect::<Vec<_>>();
let config = RuntimeConfig {
tool_call_id: request.tool_call_id,
enabled_tools,
source: request.source,
stored_values,
};
thread::spawn(move || {
run_runtime(
config,
event_tx,
command_rx,
control_rx,
pending_mode,
isolate_handle_tx,
runtime_command_tx,
);
});
let isolate_handle = isolate_handle_rx
.recv()
.map_err(|_| "failed to initialize code mode runtime".to_string())?;
Ok((command_tx, control_tx, isolate_handle))
}
#[derive(Clone)]
struct RuntimeConfig {
tool_call_id: String,
enabled_tools: Vec<EnabledToolMetadata>,
source: String,
stored_values: HashMap<String, JsonValue>,
}
pub(super) struct RuntimeState {
event_tx: mpsc::UnboundedSender<RuntimeEvent>,
pending_tool_calls: HashMap<String, v8::Global<v8::PromiseResolver>>,
pending_timeouts: HashMap<u64, timers::ScheduledTimeout>,
stored_values: HashMap<String, JsonValue>,
stored_value_writes: HashMap<String, JsonValue>,
enabled_tools: Vec<EnabledToolMetadata>,
next_tool_call_id: u64,
next_timeout_id: u64,
tool_call_id: String,
runtime_command_tx: std_mpsc::Sender<RuntimeCommand>,
exit_requested: bool,
}
pub(super) enum CompletionState {
Pending,
Completed {
stored_value_writes: HashMap<String, JsonValue>,
error_text: Option<String>,
},
}
fn initialize_v8() -> Result<(), String> {
static PLATFORM: OnceLock<Result<v8::SharedRef<v8::Platform>, String>> = OnceLock::new();
match PLATFORM.get_or_init(|| {
v8::icu::set_common_data_77(deno_core_icudata::ICU_DATA)
.map_err(|error_code| format!("failed to initialize ICU data: {error_code}"))?;
let platform = v8::new_default_platform(0, false).make_shared();
v8::V8::initialize_platform(platform.clone());
v8::V8::initialize();
Ok(platform)
}) {
Ok(_) => Ok(()),
Err(error_text) => Err(error_text.clone()),
}
}
fn run_runtime(
config: RuntimeConfig,
event_tx: mpsc::UnboundedSender<RuntimeEvent>,
command_rx: std_mpsc::Receiver<RuntimeCommand>,
control_rx: std_mpsc::Receiver<RuntimeControlCommand>,
pending_mode: PendingRuntimeMode,
isolate_handle_tx: std_mpsc::SyncSender<v8::IsolateHandle>,
runtime_command_tx: std_mpsc::Sender<RuntimeCommand>,
) {
let isolate = &mut v8::Isolate::new(v8::CreateParams::default());
let isolate_handle = isolate.thread_safe_handle();
if isolate_handle_tx.send(isolate_handle).is_err() {
return;
}
isolate.set_host_import_module_dynamically_callback(module_loader::dynamic_import_callback);
v8::scope!(let scope, isolate);
let context = v8::Context::new(scope, Default::default());
let scope = &mut v8::ContextScope::new(scope, context);
scope.set_slot(RuntimeState {
event_tx: event_tx.clone(),
pending_tool_calls: HashMap::new(),
pending_timeouts: HashMap::new(),
stored_values: config.stored_values,
stored_value_writes: HashMap::new(),
enabled_tools: config.enabled_tools,
next_tool_call_id: 1,
next_timeout_id: 1,
tool_call_id: config.tool_call_id,
runtime_command_tx,
exit_requested: false,
});
if let Err(error_text) = globals::install_globals(scope) {
send_result(&event_tx, HashMap::new(), Some(error_text));
return;
}
let _ = event_tx.send(RuntimeEvent::Started);
let pending_promise = match module_loader::evaluate_main_module(scope, &config.source) {
Ok(pending_promise) => pending_promise,
Err(error_text) => {
capture_scope_send_error(scope, &event_tx, Some(error_text));
return;
}
};
match module_loader::completion_state(scope, pending_promise.as_ref()) {
CompletionState::Completed {
stored_value_writes,
error_text,
} => {
send_result(&event_tx, stored_value_writes, error_text);
return;
}
CompletionState::Pending => {}
}
let mut pending_promise = pending_promise;
while let Some(command) =
next_runtime_command(&event_tx, &command_rx, &control_rx, pending_mode)
{
match command {
RuntimeCommand::Terminate => break,
RuntimeCommand::ToolResponse { id, result } => {
if let Err(error_text) =
module_loader::resolve_tool_response(scope, &id, Ok(result))
{
capture_scope_send_error(scope, &event_tx, Some(error_text));
return;
}
}
RuntimeCommand::ToolError { id, error_text } => {
if let Err(runtime_error) =
module_loader::resolve_tool_response(scope, &id, Err(error_text))
{
capture_scope_send_error(scope, &event_tx, Some(runtime_error));
return;
}
}
RuntimeCommand::TimeoutFired { id } => {
if let Err(runtime_error) = timers::invoke_timeout_callback(scope, id) {
capture_scope_send_error(scope, &event_tx, Some(runtime_error));
return;
}
}
RuntimeCommand::ObservePendingFrontier => {}
}
scope.perform_microtask_checkpoint();
match module_loader::completion_state(scope, pending_promise.as_ref()) {
CompletionState::Completed {
stored_value_writes,
error_text,
} => {
send_result(&event_tx, stored_value_writes, error_text);
return;
}
CompletionState::Pending => {}
}
if let Some(promise) = pending_promise.as_ref() {
let promise = v8::Local::new(scope, promise);
if promise.state() != v8::PromiseState::Pending {
pending_promise = None;
}
}
}
}
fn next_runtime_command(
event_tx: &mpsc::UnboundedSender<RuntimeEvent>,
command_rx: &std_mpsc::Receiver<RuntimeCommand>,
control_rx: &std_mpsc::Receiver<RuntimeControlCommand>,
pending_mode: PendingRuntimeMode,
) -> Option<RuntimeCommand> {
loop {
match command_rx.try_recv() {
Ok(command) => return Some(command),
Err(std_mpsc::TryRecvError::Disconnected) => return None,
Err(std_mpsc::TryRecvError::Empty) => {}
}
let _ = event_tx.send(RuntimeEvent::Pending);
match pending_mode {
#[cfg(test)]
PendingRuntimeMode::Continue => return command_rx.recv().ok(),
PendingRuntimeMode::PauseUntilResumed => match control_rx.recv().ok()? {
RuntimeControlCommand::Continue => return command_rx.recv().ok(),
RuntimeControlCommand::Resume => continue,
RuntimeControlCommand::Terminate => return Some(RuntimeCommand::Terminate),
},
}
}
}
fn capture_scope_send_error(
scope: &mut v8::PinScope<'_, '_>,
event_tx: &mpsc::UnboundedSender<RuntimeEvent>,
error_text: Option<String>,
) {
let stored_value_writes = scope
.get_slot::<RuntimeState>()
.map(|state| state.stored_value_writes.clone())
.unwrap_or_default();
send_result(event_tx, stored_value_writes, error_text);
}
fn send_result(
event_tx: &mpsc::UnboundedSender<RuntimeEvent>,
stored_value_writes: HashMap<String, JsonValue>,
error_text: Option<String>,
) {
let _ = event_tx.send(RuntimeEvent::Result {
stored_value_writes,
error_text,
});
}
#[cfg(test)]
mod tests {
use std::collections::HashMap;
use std::time::Duration;
use pretty_assertions::assert_eq;
use tokio::sync::mpsc;
use super::ExecuteRequest;
use super::PendingRuntimeMode;
use super::RuntimeCommand;
use super::RuntimeControlCommand;
use super::RuntimeEvent;
use super::spawn_runtime;
use crate::FunctionCallOutputContentItem;
fn execute_request(source: &str) -> ExecuteRequest {
ExecuteRequest {
tool_call_id: "call_1".to_string(),
enabled_tools: Vec::new(),
source: source.to_string(),
yield_time_ms: Some(1),
max_output_tokens: None,
}
}
#[tokio::test]
async fn terminate_execution_stops_cpu_bound_module() {
let (event_tx, mut event_rx) = mpsc::unbounded_channel();
let (_runtime_tx, _runtime_control_tx, runtime_terminate_handle) = spawn_runtime(
HashMap::new(),
execute_request("while (true) {}"),
event_tx,
PendingRuntimeMode::Continue,
)
.unwrap();
let started_event = tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.unwrap();
assert!(matches!(started_event, RuntimeEvent::Started));
assert!(runtime_terminate_handle.terminate_execution());
let result_event = tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.unwrap();
let RuntimeEvent::Result { error_text, .. } = result_event else {
panic!("expected runtime result after termination");
};
assert!(error_text.is_some());
assert!(
tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.is_none()
);
}
#[tokio::test]
async fn pending_mode_freezes_runtime_commands_until_resume() {
let (event_tx, mut event_rx) = mpsc::unbounded_channel();
let (runtime_tx, runtime_control_tx, _runtime_terminate_handle) = spawn_runtime(
HashMap::new(),
execute_request(
r#"
await new Promise((resolve) => setTimeout(resolve, 60_000));
text("after");
await new Promise(() => {});
"#,
),
event_tx,
PendingRuntimeMode::PauseUntilResumed,
)
.unwrap();
assert!(matches!(
tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.unwrap(),
RuntimeEvent::Started
));
assert!(matches!(
tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.unwrap(),
RuntimeEvent::Pending
));
runtime_tx
.send(RuntimeCommand::TimeoutFired { id: 1 })
.unwrap();
assert!(
tokio::time::timeout(Duration::from_millis(100), event_rx.recv())
.await
.is_err()
);
runtime_control_tx
.send(RuntimeControlCommand::Resume)
.unwrap();
let content_event = tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.unwrap();
let RuntimeEvent::ContentItem(FunctionCallOutputContentItem::InputText { text }) =
content_event
else {
panic!("expected resumed runtime output");
};
assert_eq!(text, "after");
assert!(matches!(
tokio::time::timeout(Duration::from_secs(1), event_rx.recv())
.await
.unwrap()
.unwrap(),
RuntimeEvent::Pending
));
runtime_control_tx
.send(RuntimeControlCommand::Terminate)
.unwrap();
}
}