Update to use Feature, with log line about time and compression ratio

codex-rs/core/tests/responses_headers.rs

@@ -12,6 +12,7 @@ use codex_core::ResponseEvent;
 use codex_core::ResponseItem;
 use codex_core::WireApi;
 use codex_core::models_manager::manager::ModelsManager;
+use codex_core::features::RequestCompressionFeature;
 use codex_otel::otel_manager::OtelManager;
 use codex_protocol::ConversationId;
 use codex_protocol::config_types::ReasoningSummary;
@@ -47,7 +48,7 @@ async fn responses_stream_includes_subagent_header_on_review() {
         env_key_instructions: None,
         experimental_bearer_token: None,
         wire_api: WireApi::Responses,
-        request_compression: RequestCompression::Zstd,
+        request_compression: RequestCompression::None,
         query_params: None,
         http_headers: None,
         env_http_headers: None,
@@ -142,7 +143,7 @@ async fn responses_stream_includes_subagent_header_on_other() {
         env_key_instructions: None,
         experimental_bearer_token: None,
         wire_api: WireApi::Responses,
-        request_compression: RequestCompression::Zstd,
+        request_compression: RequestCompression::None,
         query_params: None,
         http_headers: None,
         env_http_headers: None,
@@ -335,13 +336,13 @@ async fn responses_request_body_is_zstd_encoded() {
     let request_recorder = responses::mount_sse_once(&server, response_body).await;
 
     let provider = ModelProviderInfo {
-        name: "mock".into(),
+        name: "OpenAI".into(),
         base_url: Some(format!("{}/v1", server.uri())),
         env_key: None,
         env_key_instructions: None,
         experimental_bearer_token: None,
         wire_api: WireApi::Responses,
-        request_compression: RequestCompression::Zstd,
+        request_compression: RequestCompression::None,
         query_params: None,
         http_headers: None,
         env_http_headers: None,
@@ -355,6 +356,9 @@ async fn responses_request_body_is_zstd_encoded() {
     let mut config = load_default_config_for_test(&codex_home);
     config.model_provider_id = provider.name.clone();
     config.model_provider = provider.clone();
+    config
+        .features
+        .set_request_compression(RequestCompressionFeature::Zstd);
     let effort = config.model_reasoning_effort;
     let summary = config.model_reasoning_summary;
     let model = ModelsManager::get_model_offline(config.model.as_deref());
@@ -376,9 +380,11 @@ async fn responses_request_body_is_zstd_encoded() {
         session_source.clone(),
     );
 
+    let auth_manager =
+        AuthManager::from_auth_for_testing(CodexAuth::create_dummy_chatgpt_auth_for_testing());
     let client = ModelClient::new(
         Arc::clone(&config),
-        None,
+        Some(auth_manager),
         model_family,
         otel_manager,
         provider,
@@ -387,11 +393,6 @@ async fn responses_request_body_is_zstd_encoded() {
         conversation_id,
         session_source,
     );
-    assert_eq!(
-        client.get_provider().request_compression,
-        RequestCompression::Zstd
-    );
-
     let mut prompt = Prompt::default();
     prompt.input = vec![ResponseItem::Message {
         id: None,
@@ -435,7 +436,7 @@ async fn responses_request_body_is_uncompressed_when_disabled() {
     let request_recorder = responses::mount_sse_once(&server, response_body).await;
 
     let provider = ModelProviderInfo {
-        name: "mock".into(),
+        name: "OpenAI".into(),
         base_url: Some(format!("{}/v1", server.uri())),
         env_key: None,
         env_key_instructions: None,
@@ -476,9 +477,11 @@ async fn responses_request_body_is_uncompressed_when_disabled() {
         session_source.clone(),
     );
 
+    let auth_manager =
+        AuthManager::from_auth_for_testing(CodexAuth::create_dummy_chatgpt_auth_for_testing());
     let client = ModelClient::new(
         Arc::clone(&config),
-        None,
+        Some(auth_manager),
         model_family,
         otel_manager,
         provider,