chore: enforce argument-comment lint across all targets

This commit is contained in:
Michael Bolin
2026-03-27 17:34:47 -07:00
parent f4f6eca871
commit f05597bbc0
307 changed files with 7720 additions and 4706 deletions
+24 -24
View File
@@ -77,24 +77,24 @@ async fn responses_stream_includes_subagent_header_on_review() {
conversation_id,
model.as_str(),
model_info.slug.as_str(),
None,
/*account_id*/ None,
Some("test@test.com".to_string()),
Some(auth_mode),
"test_originator".to_string(),
false,
/*log_user_prompts*/ false,
"test".to_string(),
session_source.clone(),
);
let client = ModelClient::new(
None,
/*auth_manager*/ None,
conversation_id,
provider.clone(),
session_source,
config.model_verbosity,
false,
false,
None,
/*enable_request_compression*/ false,
/*include_timing_metrics*/ false,
/*beta_features_header*/ None,
);
let mut client_session = client.new_session();
@@ -116,8 +116,8 @@ async fn responses_stream_includes_subagent_header_on_review() {
&session_telemetry,
effort,
summary.unwrap_or(model_info.default_reasoning_summary),
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("stream failed");
@@ -190,24 +190,24 @@ async fn responses_stream_includes_subagent_header_on_other() {
conversation_id,
model.as_str(),
model_info.slug.as_str(),
None,
/*account_id*/ None,
Some("test@test.com".to_string()),
Some(auth_mode),
"test_originator".to_string(),
false,
/*log_user_prompts*/ false,
"test".to_string(),
session_source.clone(),
);
let client = ModelClient::new(
None,
/*auth_manager*/ None,
conversation_id,
provider.clone(),
session_source,
config.model_verbosity,
false,
false,
None,
/*enable_request_compression*/ false,
/*include_timing_metrics*/ false,
/*beta_features_header*/ None,
);
let mut client_session = client.new_session();
@@ -229,8 +229,8 @@ async fn responses_stream_includes_subagent_header_on_other() {
&session_telemetry,
effort,
summary.unwrap_or(model_info.default_reasoning_summary),
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("stream failed");
@@ -302,24 +302,24 @@ async fn responses_respects_model_info_overrides_from_config() {
conversation_id,
model.as_str(),
model_info.slug.as_str(),
None,
/*account_id*/ None,
Some("test@test.com".to_string()),
auth_mode,
"test_originator".to_string(),
false,
/*log_user_prompts*/ false,
"test".to_string(),
session_source.clone(),
);
let client = ModelClient::new(
None,
/*auth_manager*/ None,
conversation_id,
provider.clone(),
session_source,
config.model_verbosity,
false,
false,
None,
/*enable_request_compression*/ false,
/*include_timing_metrics*/ false,
/*beta_features_header*/ None,
);
let mut client_session = client.new_session();
@@ -341,8 +341,8 @@ async fn responses_respects_model_info_overrides_from_config() {
&session_telemetry,
effort,
summary.unwrap_or(model_info.default_reasoning_summary),
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("stream failed");
+1 -1
View File
@@ -268,7 +268,7 @@ async fn report_agent_job_result_rejects_wrong_thread() -> Result<()> {
.expect("job_id from csv");
let job = db.get_agent_job(job_id.as_str()).await?.expect("job");
let items = db
.list_agent_job_items(job.id.as_str(), None, Some(10))
.list_agent_job_items(job.id.as_str(), /*status*/ None, Some(10))
.await?;
let item = items.first().expect("item");
let wrong_thread_id = "00000000-0000-0000-0000-000000000000";
+16 -5
View File
@@ -273,7 +273,10 @@ async fn websocket_v2_first_turn_uses_updated_fast_tier_after_startup_prewarm()
});
let test = builder.build_with_websocket_server(&server).await?;
let warmup = server.wait_for_request(0, 0).await.body_json();
let warmup = server
.wait_for_request(/*connection_index*/ 0, /*request_index*/ 0)
.await
.body_json();
assert_eq!(warmup["type"].as_str(), Some("response.create"));
assert_eq!(warmup["generate"].as_bool(), Some(false));
assert_eq!(warmup.get("service_tier"), None);
@@ -326,12 +329,16 @@ async fn websocket_v2_first_turn_drops_fast_tier_after_startup_prewarm() -> Resu
});
let test = builder.build_with_websocket_server(&server).await?;
let warmup = server.wait_for_request(0, 0).await.body_json();
let warmup = server
.wait_for_request(/*connection_index*/ 0, /*request_index*/ 0)
.await
.body_json();
assert_eq!(warmup["type"].as_str(), Some("response.create"));
assert_eq!(warmup["generate"].as_bool(), Some(false));
assert_eq!(warmup["service_tier"].as_str(), Some("priority"));
test.submit_turn_with_service_tier("hello", None).await?;
test.submit_turn_with_service_tier("hello", /*service_tier*/ None)
.await?;
assert_eq!(server.handshakes().len(), 1);
let connection = server.single_connection();
@@ -382,14 +389,18 @@ async fn websocket_v2_next_turn_uses_updated_service_tier() -> Result<()> {
});
let test = builder.build_with_websocket_server(&server).await?;
let warmup = server.wait_for_request(0, 0).await.body_json();
let warmup = server
.wait_for_request(/*connection_index*/ 0, /*request_index*/ 0)
.await
.body_json();
assert_eq!(warmup["type"].as_str(), Some("response.create"));
assert_eq!(warmup["generate"].as_bool(), Some(false));
assert_eq!(warmup.get("service_tier"), None);
test.submit_turn_with_service_tier("first", Some(ServiceTier::Fast))
.await?;
test.submit_turn_with_service_tier("second", None).await?;
test.submit_turn_with_service_tier("second", /*service_tier*/ None)
.await?;
assert_eq!(server.handshakes().len(), 1);
let connection = server.single_connection();
+52 -13
View File
@@ -131,7 +131,12 @@ impl ActionKind {
"from pathlib import Path; path = Path({path_str:?}); content = {content:?}; path.write_text(content, encoding='utf-8'); print(path.read_text(encoding='utf-8'), end='')",
);
let command = format!("python3 -c {script:?}");
let event = shell_event(call_id, &command, 5_000, sandbox_permissions)?;
let event = shell_event(
call_id,
&command,
/*timeout_ms*/ 5_000,
sandbox_permissions,
)?;
Ok((event, Some(command)))
}
ActionKind::FetchUrl {
@@ -153,7 +158,12 @@ impl ActionKind {
);
let command = format!("python3 -c \"{script}\"");
let event = shell_event(call_id, &command, 5_000, sandbox_permissions)?;
let event = shell_event(
call_id,
&command,
/*timeout_ms*/ 5_000,
sandbox_permissions,
)?;
Ok((event, Some(command)))
}
ActionKind::FetchUrlNoProxy {
@@ -175,11 +185,21 @@ impl ActionKind {
);
let command = format!("python3 -c \"{script}\"");
let event = shell_event(call_id, &command, 5_000, sandbox_permissions)?;
let event = shell_event(
call_id,
&command,
/*timeout_ms*/ 5_000,
sandbox_permissions,
)?;
Ok((event, Some(command)))
}
ActionKind::RunCommand { command } => {
let event = shell_event(call_id, command, 2_000, sandbox_permissions)?;
let event = shell_event(
call_id,
command,
/*timeout_ms*/ 2_000,
sandbox_permissions,
)?;
Ok((event, Some(command.to_string())))
}
ActionKind::RunUnifiedExecCommand {
@@ -206,7 +226,12 @@ impl ActionKind {
let _ = fs::remove_file(&path);
let patch = build_add_file_patch(&patch_path, content);
let command = shell_apply_patch_command(&patch);
let event = shell_event(call_id, &command, 5_000, sandbox_permissions)?;
let event = shell_event(
call_id,
&command,
/*timeout_ms*/ 5_000,
sandbox_permissions,
)?;
Ok((event, Some(command)))
}
}
@@ -233,7 +258,13 @@ fn shell_event(
timeout_ms: u64,
sandbox_permissions: SandboxPermissions,
) -> Result<Value> {
shell_event_with_prefix_rule(call_id, command, timeout_ms, sandbox_permissions, None)
shell_event_with_prefix_rule(
call_id,
command,
timeout_ms,
sandbox_permissions,
/*prefix_rule*/ None,
)
}
fn shell_event_with_prefix_rule(
@@ -2259,7 +2290,12 @@ async fn matched_prefix_rule_runs_unsandboxed_under_zsh_fork() -> Result<()> {
.await?;
let call_id = "zsh-fork-prefix-rule-unsandboxed";
let event = shell_event(call_id, &command, 1_000, SandboxPermissions::UseDefault)?;
let event = shell_event(
call_id,
&command,
/*timeout_ms*/ 1_000,
SandboxPermissions::UseDefault,
)?;
let _ = mount_sse_once(
&server,
sse(vec![
@@ -2318,7 +2354,7 @@ async fn invalid_requested_prefix_rule_falls_back_for_compound_command() -> Resu
let event = shell_event_with_prefix_rule(
call_id,
command,
1_000,
/*timeout_ms*/ 1_000,
SandboxPermissions::RequireEscalated,
Some(vec!["touch".to_string()]),
)?;
@@ -2369,7 +2405,7 @@ async fn approving_fallback_rule_for_compound_command_works() -> Result<()> {
let event = shell_event_with_prefix_rule(
call_id,
command,
1_000,
/*timeout_ms*/ 1_000,
SandboxPermissions::RequireEscalated,
Some(vec!["touch".to_string()]),
)?;
@@ -2416,7 +2452,7 @@ async fn approving_fallback_rule_for_compound_command_works() -> Result<()> {
let event = shell_event_with_prefix_rule(
call_id,
command,
1_000,
/*timeout_ms*/ 1_000,
SandboxPermissions::RequireEscalated,
Some(vec!["touch".to_string()]),
)?;
@@ -2498,7 +2534,10 @@ allow_local_binding = true
config.permissions.sandbox_policy = Constrained::allow_any(sandbox_policy_for_config);
let layers = config
.config_layer_stack
.get_layers(ConfigLayerStackOrdering::LowestPrecedenceFirst, true)
.get_layers(
ConfigLayerStackOrdering::LowestPrecedenceFirst,
/*include_disabled*/ true,
)
.into_iter()
.cloned()
.collect();
@@ -2542,7 +2581,7 @@ allow_local_binding = true
let first_event = shell_event(
call_id_first,
&fetch_command,
30_000,
/*timeout_ms*/ 30_000,
SandboxPermissions::UseDefault,
)?;
@@ -2682,7 +2721,7 @@ allow_local_binding = true
let second_event = shell_event(
call_id_second,
&fetch_command,
30_000,
/*timeout_ms*/ 30_000,
SandboxPermissions::UseDefault,
)?;
+1 -1
View File
@@ -960,7 +960,7 @@ impl RefreshTokenTestContext {
let auth_manager = AuthManager::shared(
codex_home.path().to_path_buf(),
false,
/*enable_codex_api_key_env*/ false,
AuthCredentialsStoreMode::File,
);
+17 -12
View File
@@ -718,7 +718,8 @@ async fn chatgpt_auth_sends_correct_request() {
)
.await;
let mut model_provider = built_in_model_providers(/* openai_base_url */ None)["openai"].clone();
let mut model_provider =
built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone();
model_provider.base_url = Some(format!("{}/api/codex", server.uri()));
model_provider.supports_websockets = false;
let mut builder = test_codex()
@@ -796,7 +797,7 @@ async fn prefers_apikey_when_config_prefers_apikey_even_with_chatgpt_tokens() {
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
supports_websockets: false,
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
// Init session
@@ -1825,24 +1826,24 @@ async fn azure_responses_request_includes_store_and_reasoning_ids() {
conversation_id,
model.as_str(),
model_info.slug.as_str(),
None,
/*account_id*/ None,
Some("test@test.com".to_string()),
auth_manager.auth_mode().map(TelemetryAuthMode::from),
"test_originator".to_string(),
false,
/*log_user_prompts*/ false,
"test".to_string(),
SessionSource::Exec,
);
let client = ModelClient::new(
None,
/*auth_manager*/ None,
conversation_id,
provider.clone(),
SessionSource::Exec,
config.model_verbosity,
false,
false,
None,
/*enable_request_compression*/ false,
/*include_timing_metrics*/ false,
/*beta_features_header*/ None,
);
let mut client_session = client.new_session();
@@ -1917,8 +1918,8 @@ async fn azure_responses_request_includes_store_and_reasoning_ids() {
&session_telemetry,
effort,
summary.unwrap_or(ReasoningSummary::Auto),
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("responses stream to start");
@@ -1957,7 +1958,10 @@ async fn token_count_includes_rate_limits_snapshot() {
skip_if_no_network!();
let server = MockServer::start().await;
let sse_body = sse(vec![ev_completed_with_tokens("resp_rate", 123)]);
let sse_body = sse(vec![ev_completed_with_tokens(
"resp_rate",
/*total_tokens*/ 123,
)]);
let response = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
@@ -1976,7 +1980,8 @@ async fn token_count_includes_rate_limits_snapshot() {
.mount(&server)
.await;
let mut provider = built_in_model_providers(/* openai_base_url */ None)["openai"].clone();
let mut provider =
built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone();
provider.base_url = Some(format!("{}/v1", server.uri()));
provider.supports_websockets = false;
+59 -45
View File
@@ -139,7 +139,7 @@ async fn responses_websocket_streams_without_feature_flag_when_provider_supports
]]])
.await;
let harness = websocket_harness_with_options(&server, false).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ false).await;
let mut client_session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
@@ -292,7 +292,7 @@ async fn responses_websocket_request_prewarm_reuses_connection() {
]])
.await;
let harness = websocket_harness_with_options(&server, true).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ true).await;
let mut client_session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
client_session
@@ -302,8 +302,8 @@ async fn responses_websocket_request_prewarm_reuses_connection() {
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket prewarm failed");
@@ -383,8 +383,8 @@ async fn responses_websocket_preconnect_is_reused_even_with_header_changes() {
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket stream failed");
@@ -411,7 +411,7 @@ async fn responses_websocket_request_prewarm_is_reused_even_with_header_changes(
]])
.await;
let harness = websocket_harness_with_options(&server, true).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ true).await;
let mut client_session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
client_session
@@ -421,8 +421,8 @@ async fn responses_websocket_request_prewarm_is_reused_even_with_header_changes(
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket prewarm failed");
@@ -433,8 +433,8 @@ async fn responses_websocket_request_prewarm_is_reused_even_with_header_changes(
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket stream failed");
@@ -476,7 +476,7 @@ async fn responses_websocket_prewarm_uses_v2_when_provider_supports_websockets()
]]])
.await;
let harness = websocket_harness_with_options(&server, false).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ false).await;
let mut client_session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
client_session
@@ -486,8 +486,8 @@ async fn responses_websocket_prewarm_uses_v2_when_provider_supports_websockets()
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket prewarm failed");
@@ -533,7 +533,7 @@ async fn responses_websocket_preconnect_runs_when_only_v2_feature_enabled() {
]]])
.await;
let harness = websocket_harness_with_options(&server, true).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ true).await;
let mut client_session = harness.client.new_session();
client_session
.preconnect_websocket(&harness.session_telemetry, &harness.model_info)
@@ -576,7 +576,7 @@ async fn responses_websocket_v2_requests_use_v2_when_provider_supports_websocket
]])
.await;
let harness = websocket_harness_with_options(&server, true).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ true).await;
let mut client_session = harness.client.new_session();
let prompt_one = prompt_with_input(vec![message_item("hello")]);
let prompt_two = prompt_with_input(vec![
@@ -625,7 +625,7 @@ async fn responses_websocket_v2_incremental_requests_are_reused_across_turns() {
]])
.await;
let harness = websocket_harness_with_options(&server, false).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ false).await;
let prompt_one = prompt_with_input(vec![message_item("hello")]);
let prompt_two = prompt_with_input(vec![
message_item("hello"),
@@ -669,7 +669,7 @@ async fn responses_websocket_v2_wins_when_both_features_enabled() {
]])
.await;
let harness = websocket_harness_with_options(&server, false).await;
let harness = websocket_harness_with_options(&server, /*runtime_metrics_enabled*/ false).await;
let mut client_session = harness.client.new_session();
let prompt_one = prompt_with_input(vec![message_item("hello")]);
let prompt_two = prompt_with_input(vec![
@@ -757,7 +757,8 @@ async fn responses_websocket_includes_timing_metrics_header_when_runtime_metrics
]]])
.await;
let harness = websocket_harness_with_runtime_metrics(&server, true).await;
let harness =
websocket_harness_with_runtime_metrics(&server, /*runtime_metrics_enabled*/ true).await;
harness.session_telemetry.reset_runtime_metrics();
let mut client_session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
@@ -795,7 +796,8 @@ async fn responses_websocket_omits_timing_metrics_header_when_runtime_metrics_di
]]])
.await;
let harness = websocket_harness_with_runtime_metrics(&server, false).await;
let harness =
websocket_harness_with_runtime_metrics(&server, /*runtime_metrics_enabled*/ false).await;
let mut client_session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
@@ -833,8 +835,8 @@ async fn responses_websocket_emits_reasoning_included_event() {
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket stream failed");
@@ -906,8 +908,8 @@ async fn responses_websocket_emits_rate_limit_events() {
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket stream failed");
@@ -1213,7 +1215,7 @@ async fn responses_websocket_forwards_turn_metadata_on_initial_and_incremental_c
&mut client_session,
&harness,
&prompt_one,
None,
/*service_tier*/ None,
Some(first_turn_metadata),
)
.await;
@@ -1221,7 +1223,7 @@ async fn responses_websocket_forwards_turn_metadata_on_initial_and_incremental_c
&mut client_session,
&harness,
&prompt_two,
None,
/*service_tier*/ None,
Some(enriched_turn_metadata),
)
.await;
@@ -1381,7 +1383,7 @@ async fn responses_websocket_v2_creates_with_previous_response_id_on_prefix() {
]])
.await;
let harness = websocket_harness_with_v2(&server, true).await;
let harness = websocket_harness_with_v2(&server, /*runtime_metrics_enabled*/ true).await;
let mut session = harness.client.new_session();
let prompt_one = prompt_with_input(vec![message_item("hello")]);
let prompt_two = prompt_with_input(vec![
@@ -1420,7 +1422,7 @@ async fn responses_websocket_v2_creates_without_previous_response_id_when_non_in
]])
.await;
let harness = websocket_harness_with_v2(&server, true).await;
let harness = websocket_harness_with_v2(&server, /*runtime_metrics_enabled*/ true).await;
let mut session = harness.client.new_session();
let prompt_one =
prompt_with_input_and_instructions(vec![message_item("hello")], "base instructions one");
@@ -1467,7 +1469,7 @@ async fn responses_websocket_v2_after_error_uses_full_create_without_previous_re
])
.await;
let harness = websocket_harness_with_v2(&server, true).await;
let harness = websocket_harness_with_v2(&server, /*runtime_metrics_enabled*/ true).await;
let mut session = harness.client.new_session();
let prompt_one = prompt_with_input(vec![message_item("hello")]);
let prompt_two = prompt_with_input(vec![message_item("hello"), message_item("second")]);
@@ -1486,8 +1488,8 @@ async fn responses_websocket_v2_after_error_uses_full_create_without_previous_re
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket stream failed");
@@ -1559,7 +1561,7 @@ async fn responses_websocket_v2_surfaces_terminal_error_without_close_handshake(
}])
.await;
let harness = websocket_harness_with_v2(&server, true).await;
let harness = websocket_harness_with_v2(&server, /*runtime_metrics_enabled*/ true).await;
let mut session = harness.client.new_session();
let prompt_one = prompt_with_input(vec![message_item("hello")]);
let prompt_two = prompt_with_input(vec![message_item("hello"), message_item("second")]);
@@ -1573,8 +1575,8 @@ async fn responses_websocket_v2_surfaces_terminal_error_without_close_handshake(
&harness.session_telemetry,
harness.effort,
harness.summary,
None,
None,
/*service_tier*/ None,
/*turn_metadata_header*/ None,
)
.await
.expect("websocket stream failed");
@@ -1605,7 +1607,7 @@ async fn responses_websocket_v2_sets_openai_beta_header() {
]]])
.await;
let harness = websocket_harness_with_v2(&server, true).await;
let harness = websocket_harness_with_v2(&server, /*runtime_metrics_enabled*/ true).await;
let mut session = harness.client.new_session();
let prompt = prompt_with_input(vec![message_item("hello")]);
@@ -1659,7 +1661,7 @@ fn prompt_with_input_and_instructions(input: Vec<ResponseItem>, instructions: &s
}
fn websocket_provider(server: &WebSocketTestServer) -> ModelProviderInfo {
websocket_provider_with_connect_timeout(server, None)
websocket_provider_with_connect_timeout(server, /*websocket_connect_timeout_ms*/ None)
}
fn websocket_provider_with_connect_timeout(
@@ -1686,7 +1688,7 @@ fn websocket_provider_with_connect_timeout(
}
async fn websocket_harness(server: &WebSocketTestServer) -> WebsocketTestHarness {
websocket_harness_with_runtime_metrics(server, false).await
websocket_harness_with_runtime_metrics(server, /*runtime_metrics_enabled*/ false).await
}
async fn websocket_harness_with_runtime_metrics(
@@ -1739,11 +1741,11 @@ async fn websocket_harness_with_provider_options(
conversation_id,
MODEL,
model_info.slug.as_str(),
None,
/*account_id*/ None,
Some("test@test.com".to_string()),
auth_manager.auth_mode().map(TelemetryAuthMode::from),
"test_originator".to_string(),
false,
/*log_user_prompts*/ false,
"test".to_string(),
SessionSource::Exec,
)
@@ -1751,14 +1753,14 @@ async fn websocket_harness_with_provider_options(
let effort = None;
let summary = ReasoningSummary::Auto;
let client = ModelClient::new(
None,
/*auth_manager*/ None,
conversation_id,
provider.clone(),
SessionSource::Exec,
config.model_verbosity,
false,
/*enable_request_compression*/ false,
runtime_metrics_enabled,
None,
/*beta_features_header*/ None,
);
WebsocketTestHarness {
@@ -1777,7 +1779,13 @@ async fn stream_until_complete(
harness: &WebsocketTestHarness,
prompt: &Prompt,
) {
stream_until_complete_with_service_tier(client_session, harness, prompt, None).await;
stream_until_complete_with_service_tier(
client_session,
harness,
prompt,
/*service_tier*/ None,
)
.await;
}
async fn stream_until_complete_with_service_tier(
@@ -1786,8 +1794,14 @@ async fn stream_until_complete_with_service_tier(
prompt: &Prompt,
service_tier: Option<ServiceTier>,
) {
stream_until_complete_with_turn_metadata(client_session, harness, prompt, service_tier, None)
.await;
stream_until_complete_with_turn_metadata(
client_session,
harness,
prompt,
service_tier,
/*turn_metadata_header*/ None,
)
.await;
}
async fn stream_until_complete_with_turn_metadata(
+85 -75
View File
@@ -252,7 +252,7 @@ async fn code_mode_can_return_exec_command_output() -> Result<()> {
r#"
text(JSON.stringify(await tools.exec_command({ cmd: "printf code_mode_exec_marker" })));
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -264,9 +264,9 @@ text(JSON.stringify(await tools.exec_command({ cmd: "printf code_mode_exec_marke
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
let parsed: Value = serde_json::from_str(text_item(&items, 1))?;
let parsed: Value = serde_json::from_str(text_item(&items, /*index*/ 1))?;
assert!(
parsed
.get("chunk_id")
@@ -378,7 +378,7 @@ const result = await tools.update_plan({
});
text(JSON.stringify(result));
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -483,7 +483,7 @@ text(JSON.stringify(results));
.expect("parallel code mode run should send a completion request");
let items = custom_tool_output_items(&req, "call-1");
assert_eq!(items.len(), 2);
assert_eq!(text_item(&items, 1), "[\"ok\",\"ok\"]");
assert_eq!(text_item(&items, /*index*/ 1), "[\"ok\",\"ok\"]");
Ok(())
}
@@ -503,7 +503,7 @@ text(JSON.stringify(await tools.exec_command({
max_output_tokens: 100
})));
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -515,7 +515,7 @@ text(JSON.stringify(await tools.exec_command({
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
let expected_pattern = r#"(?sx)
\A
@@ -524,7 +524,7 @@ Total\ output\ lines:\ 1\n
.*…\d+\ tokens\ truncated….*
\z
"#;
assert_regex_match(expected_pattern, text_item(&items, 1));
assert_regex_match(expected_pattern, text_item(&items, /*index*/ 1));
Ok(())
}
@@ -542,7 +542,7 @@ text("before crash");
text("still before crash");
throw new Error("boom");
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -554,10 +554,10 @@ throw new Error("boom");
r"(?s)\A",
r"Script failed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert_eq!(text_item(&items, 1), "before crash");
assert_eq!(text_item(&items, 2), "still before crash");
assert_eq!(text_item(&items, /*index*/ 1), "before crash");
assert_eq!(text_item(&items, /*index*/ 2), "still before crash");
assert_regex_match(
r#"(?sx)
\A
@@ -566,7 +566,7 @@ Error:\ boom\n
(?:\s+at\ .+\n?)+
\z
"#,
text_item(&items, 3),
text_item(&items, /*index*/ 3),
);
Ok(())
@@ -589,7 +589,7 @@ try {
text(`caught:${error?.message ?? String(error)}`);
}
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -666,10 +666,10 @@ text("phase 3");
r"(?s)\A",
r"Script running with cell ID \d+\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&first_items, 0),
text_item(&first_items, /*index*/ 0),
);
assert_eq!(text_item(&first_items, 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, 0));
assert_eq!(text_item(&first_items, /*index*/ 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, /*index*/ 0));
responses::mount_sse_once(
&server,
@@ -707,13 +707,13 @@ text("phase 3");
r"(?s)\A",
r"Script running with cell ID \d+\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&second_items, 0),
text_item(&second_items, /*index*/ 0),
);
assert_eq!(
extract_running_cell_id(text_item(&second_items, 0)),
extract_running_cell_id(text_item(&second_items, /*index*/ 0)),
cell_id
);
assert_eq!(text_item(&second_items, 1), "phase 2");
assert_eq!(text_item(&second_items, /*index*/ 1), "phase 2");
responses::mount_sse_once(
&server,
@@ -751,9 +751,9 @@ text("phase 3");
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&third_items, 0),
text_item(&third_items, /*index*/ 0),
);
assert_eq!(text_item(&third_items, 1), "phase 3");
assert_eq!(text_item(&third_items, /*index*/ 1), "phase 3");
Ok(())
}
@@ -806,10 +806,10 @@ while (true) {}
r"(?s)\A",
r"Script running with cell ID \d+\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&first_items, 0),
text_item(&first_items, /*index*/ 0),
);
assert_eq!(text_item(&first_items, 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, 0));
assert_eq!(text_item(&first_items, /*index*/ 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, /*index*/ 0));
responses::mount_sse_once(
&server,
@@ -846,7 +846,7 @@ while (true) {}
r"(?s)\A",
r"Script terminated\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&second_items, 0),
text_item(&second_items, /*index*/ 0),
);
Ok(())
@@ -907,8 +907,8 @@ text("session b done");
let first_request = first_completion.single_request();
let first_items = custom_tool_output_items(&first_request, "call-1");
assert_eq!(first_items.len(), 2);
let session_a_id = extract_running_cell_id(text_item(&first_items, 0));
assert_eq!(text_item(&first_items, 1), "session a start");
let session_a_id = extract_running_cell_id(text_item(&first_items, /*index*/ 0));
assert_eq!(text_item(&first_items, /*index*/ 1), "session a start");
responses::mount_sse_once(
&server,
@@ -933,8 +933,8 @@ text("session b done");
let second_request = second_completion.single_request();
let second_items = custom_tool_output_items(&second_request, "call-2");
assert_eq!(second_items.len(), 2);
let session_b_id = extract_running_cell_id(text_item(&second_items, 0));
assert_eq!(text_item(&second_items, 1), "session b start");
let session_b_id = extract_running_cell_id(text_item(&second_items, /*index*/ 0));
assert_eq!(text_item(&second_items, /*index*/ 1), "session b start");
assert_ne!(session_a_id, session_b_id);
fs::write(&session_a_gate, "ready")?;
@@ -973,9 +973,9 @@ text("session b done");
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&third_items, 0),
text_item(&third_items, /*index*/ 0),
);
assert_eq!(text_item(&third_items, 1), "session a done");
assert_eq!(text_item(&third_items, /*index*/ 1), "session a done");
fs::write(&session_b_gate, "ready")?;
responses::mount_sse_once(
@@ -1013,9 +1013,9 @@ text("session b done");
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&fourth_items, 0),
text_item(&fourth_items, /*index*/ 0),
);
assert_eq!(text_item(&fourth_items, 1), "session b done");
assert_eq!(text_item(&fourth_items, /*index*/ 1), "session b done");
Ok(())
}
@@ -1065,8 +1065,8 @@ text("phase 2");
let first_request = first_completion.single_request();
let first_items = custom_tool_output_items(&first_request, "call-1");
assert_eq!(first_items.len(), 2);
let cell_id = extract_running_cell_id(text_item(&first_items, 0));
assert_eq!(text_item(&first_items, 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, /*index*/ 0));
assert_eq!(text_item(&first_items, /*index*/ 1), "phase 1");
responses::mount_sse_once(
&server,
@@ -1103,7 +1103,7 @@ text("phase 2");
r"(?s)\A",
r"Script terminated\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&second_items, 0),
text_item(&second_items, /*index*/ 0),
);
responses::mount_sse_once(
@@ -1140,9 +1140,9 @@ text("after terminate");
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&third_items, 0),
text_item(&third_items, /*index*/ 0),
);
assert_eq!(text_item(&third_items, 1), "after terminate");
assert_eq!(text_item(&third_items, /*index*/ 1), "after terminate");
Ok(())
}
@@ -1197,10 +1197,10 @@ async fn code_mode_wait_returns_error_for_unknown_session() -> Result<()> {
r"(?s)\A",
r"Script failed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert_eq!(
text_item(&items, 1),
text_item(&items, /*index*/ 1),
"Script error:\nexec cell 999999 not found"
);
@@ -1268,8 +1268,8 @@ text("session b done");
let first_request = first_completion.single_request();
let first_items = custom_tool_output_items(&first_request, "call-1");
assert_eq!(first_items.len(), 2);
let session_a_id = extract_running_cell_id(text_item(&first_items, 0));
assert_eq!(text_item(&first_items, 1), "session a start");
let session_a_id = extract_running_cell_id(text_item(&first_items, /*index*/ 0));
assert_eq!(text_item(&first_items, /*index*/ 1), "session a start");
responses::mount_sse_once(
&server,
@@ -1294,8 +1294,8 @@ text("session b done");
let second_request = second_completion.single_request();
let second_items = custom_tool_output_items(&second_request, "call-2");
assert_eq!(second_items.len(), 2);
let session_b_id = extract_running_cell_id(text_item(&second_items, 0));
assert_eq!(text_item(&second_items, 1), "session b start");
let session_b_id = extract_running_cell_id(text_item(&second_items, /*index*/ 0));
assert_eq!(text_item(&second_items, /*index*/ 1), "session b start");
fs::write(&session_a_gate, "ready")?;
responses::mount_sse_once(
@@ -1333,10 +1333,10 @@ text("session b done");
r"(?s)\A",
r"Script running with cell ID \d+\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&third_items, 0),
text_item(&third_items, /*index*/ 0),
);
assert_eq!(
extract_running_cell_id(text_item(&third_items, 0)),
extract_running_cell_id(text_item(&third_items, /*index*/ 0)),
session_b_id
);
@@ -1384,7 +1384,7 @@ text("session b done");
r"(?s)\A",
r"Script terminated\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&fourth_items, 0),
text_item(&fourth_items, /*index*/ 0),
);
}
2 => {
@@ -1393,9 +1393,9 @@ text("session b done");
r"(?s)\A",
r"Script (?:completed|terminated)\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&fourth_items, 0),
text_item(&fourth_items, /*index*/ 0),
);
assert_eq!(text_item(&fourth_items, 1), "session a done");
assert_eq!(text_item(&fourth_items, /*index*/ 1), "session a done");
}
other => panic!("unexpected number of content items: {other}"),
}
@@ -1455,9 +1455,9 @@ text("after yield");
r"(?s)\A",
r"Script running with cell ID \d+\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&first_items, 0),
text_item(&first_items, /*index*/ 0),
);
assert_eq!(text_item(&first_items, 1), "before yield");
assert_eq!(text_item(&first_items, /*index*/ 1), "before yield");
responses::mount_sse_once(
&server,
@@ -1541,8 +1541,8 @@ text("token one token two token three token four token five token six token seve
let first_request = first_completion.single_request();
let first_items = custom_tool_output_items(&first_request, "call-1");
assert_eq!(first_items.len(), 2);
assert_eq!(text_item(&first_items, 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, 0));
assert_eq!(text_item(&first_items, /*index*/ 1), "phase 1");
let cell_id = extract_running_cell_id(text_item(&first_items, /*index*/ 0));
fs::write(&completion_gate, "ready")?;
responses::mount_sse_once(
@@ -1581,7 +1581,7 @@ text("token one token two token three token four token five token six token seve
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&second_items, 0),
text_item(&second_items, /*index*/ 0),
);
let expected_pattern = r#"(?sx)
\A
@@ -1590,7 +1590,7 @@ Total\ output\ lines:\ 1\n
.*…\d+\ tokens\ truncated….*
\z
"#;
assert_regex_match(expected_pattern, text_item(&second_items, 1));
assert_regex_match(expected_pattern, text_item(&second_items, /*index*/ 1));
Ok(())
}
@@ -1606,7 +1606,7 @@ async fn code_mode_can_output_serialized_text_via_global_helper() -> Result<()>
r#"
text({ json: true });
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -1639,7 +1639,7 @@ notify("code_mode_notify_marker");
await tools.test_sync_tool({});
text("done");
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -1677,7 +1677,7 @@ text("before");
exit();
text("after");
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -1695,9 +1695,9 @@ text("after");
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert_eq!(text_item(&items, 1), "before");
assert_eq!(text_item(&items, /*index*/ 1), "before");
assert_eq!(output, "before");
Ok(())
@@ -1716,7 +1716,7 @@ const circular = {};
circular.self = circular;
text(circular);
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -1736,10 +1736,10 @@ text(circular);
r"(?s)\A",
r"Script failed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert!(text_item(&items, 1).contains("Script error:"));
assert!(text_item(&items, 1).contains("Converting circular structure to JSON"));
assert!(text_item(&items, /*index*/ 1).contains("Script error:"));
assert!(text_item(&items, /*index*/ 1).contains("Converting circular structure to JSON"));
Ok(())
}
@@ -1756,7 +1756,7 @@ async fn code_mode_can_output_images_via_global_helper() -> Result<()> {
image("https://example.com/image.jpg");
image("data:image/png;base64,AAA");
"#,
false,
/*include_apply_patch*/ false,
)
.await?;
@@ -1774,7 +1774,7 @@ image("data:image/png;base64,AAA");
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert_eq!(
items[1],
@@ -1857,7 +1857,7 @@ image(out);
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert_eq!(
@@ -1889,8 +1889,13 @@ async fn code_mode_can_apply_patch_via_nested_tool() -> Result<()> {
);
let code = format!("text(await tools.apply_patch({patch:?}));\n");
let (test, second_mock) =
run_code_mode_turn(&server, "use exec to run apply_patch", &code, true).await?;
let (test, second_mock) = run_code_mode_turn(
&server,
"use exec to run apply_patch",
&code,
/*include_apply_patch*/ true,
)
.await?;
let req = second_mock.single_request();
let items = custom_tool_output_items(&req, "call-1");
@@ -1908,9 +1913,9 @@ async fn code_mode_can_apply_patch_via_nested_tool() -> Result<()> {
r"(?s)\A",
r"Script completed\nWall time \d+\.\d seconds\nOutput:\n\z"
),
text_item(&items, 0),
text_item(&items, /*index*/ 0),
);
assert_eq!(text_item(&items, 1), "{}");
assert_eq!(text_item(&items, /*index*/ 1), "{}");
let file_path = test.cwd_path().join(file_name);
assert_eq!(fs::read_to_string(&file_path)?, "hello from code_mode\n");
@@ -2183,8 +2188,13 @@ const tool = ALL_TOOLS.find(({ name }) => name === "view_image");
text(JSON.stringify(tool));
"#;
let (_test, second_mock) =
run_code_mode_turn(&server, "use exec to inspect ALL_TOOLS", code, false).await?;
let (_test, second_mock) = run_code_mode_turn(
&server,
"use exec to inspect ALL_TOOLS",
code,
/*include_apply_patch*/ false,
)
.await?;
let req = second_mock.single_request();
let (output, success) = custom_tool_output_body_and_success(&req, "call-1");
@@ -2273,7 +2283,7 @@ async fn code_mode_can_call_hidden_dynamic_tools() -> Result<()> {
}),
defer_loading: true,
}],
false,
/*persist_extended_history*/ false,
)
.await?;
let mut test = base_test;
+59 -56
View File
@@ -94,7 +94,8 @@ fn json_fragment(text: &str) -> String {
}
fn non_openai_model_provider(server: &MockServer) -> ModelProviderInfo {
let mut provider = built_in_model_providers(/* openai_base_url */ None)["openai"].clone();
let mut provider =
built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone();
provider.name = "OpenAI (test)".into();
provider.base_url = Some(format!("{}/v1", server.uri()));
provider.supports_websockets = false;
@@ -413,11 +414,11 @@ async fn manual_compact_uses_custom_prompt() {
let server = start_mock_server().await;
let first_turn = sse(vec![
ev_assistant_message("m0", FIRST_REPLY),
ev_completed_with_tokens("r0", 80),
ev_completed_with_tokens("r0", /*total_tokens*/ 80),
]);
let compact_turn = sse(vec![
ev_assistant_message("m1", SUMMARY_TEXT),
ev_completed_with_tokens("r1", 100),
ev_completed_with_tokens("r1", /*total_tokens*/ 100),
]);
let request_log = mount_sse_sequence(&server, vec![first_turn, compact_turn]).await;
@@ -508,7 +509,7 @@ async fn manual_compact_emits_api_and_local_token_usage_events() {
// history.
let sse_compact = sse(vec![
ev_assistant_message("m1", SUMMARY_TEXT),
ev_completed_with_tokens("r1", 0),
ev_completed_with_tokens("r1", /*total_tokens*/ 0),
]);
mount_sse_once(&server, sse_compact).await;
@@ -1211,21 +1212,21 @@ async fn auto_compact_runs_after_token_limit_hit() {
let sse1 = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 70_000),
ev_completed_with_tokens("r1", /*total_tokens*/ 70_000),
]);
let sse2 = sse(vec![
ev_assistant_message("m2", "SECOND_REPLY"),
ev_completed_with_tokens("r2", 330_000),
ev_completed_with_tokens("r2", /*total_tokens*/ 330_000),
]);
let sse3 = sse(vec![
ev_assistant_message("m3", AUTO_SUMMARY_TEXT),
ev_completed_with_tokens("r3", 200),
ev_completed_with_tokens("r3", /*total_tokens*/ 200),
]);
let sse4 = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 120),
ev_completed_with_tokens("r4", /*total_tokens*/ 120),
]);
let prefixed_auto_summary = AUTO_SUMMARY_TEXT;
@@ -1404,19 +1405,19 @@ async fn auto_compact_emits_context_compaction_items() {
let sse1 = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 70_000),
ev_completed_with_tokens("r1", /*total_tokens*/ 70_000),
]);
let sse2 = sse(vec![
ev_assistant_message("m2", "SECOND_REPLY"),
ev_completed_with_tokens("r2", 330_000),
ev_completed_with_tokens("r2", /*total_tokens*/ 330_000),
]);
let sse3 = sse(vec![
ev_assistant_message("m3", AUTO_SUMMARY_TEXT),
ev_completed_with_tokens("r3", 200),
ev_completed_with_tokens("r3", /*total_tokens*/ 200),
]);
let sse4 = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 120),
ev_completed_with_tokens("r4", /*total_tokens*/ 120),
]);
mount_sse_sequence(&server, vec![sse1, sse2, sse3, sse4]).await;
@@ -1487,19 +1488,19 @@ async fn auto_compact_starts_after_turn_started() {
let sse1 = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 70_000),
ev_completed_with_tokens("r1", /*total_tokens*/ 70_000),
]);
let sse2 = sse(vec![
ev_assistant_message("m2", "SECOND_REPLY"),
ev_completed_with_tokens("r2", 330_000),
ev_completed_with_tokens("r2", /*total_tokens*/ 330_000),
]);
let sse3 = sse(vec![
ev_assistant_message("m3", AUTO_SUMMARY_TEXT),
ev_completed_with_tokens("r3", 200),
ev_completed_with_tokens("r3", /*total_tokens*/ 200),
]);
let sse4 = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 120),
ev_completed_with_tokens("r4", /*total_tokens*/ 120),
]);
mount_sse_sequence(&server, vec![sse1, sse2, sse3, sse4]).await;
@@ -1704,8 +1705,8 @@ async fn pre_sampling_compact_runs_on_switch_to_smaller_context_model() {
&server,
ModelsResponse {
models: vec![
model_info_with_context_window(previous_model, 273_000),
model_info_with_context_window(next_model, 125_000),
model_info_with_context_window(previous_model, /*context_window*/ 273_000),
model_info_with_context_window(next_model, /*context_window*/ 125_000),
],
},
)
@@ -1716,15 +1717,15 @@ async fn pre_sampling_compact_runs_on_switch_to_smaller_context_model() {
vec![
sse(vec![
ev_assistant_message("m1", "before switch"),
ev_completed_with_tokens("r1", 120_000),
ev_completed_with_tokens("r1", /*total_tokens*/ 120_000),
]),
sse(vec![
ev_assistant_message("m2", "PRE_SAMPLING_SUMMARY"),
ev_completed_with_tokens("r2", 10),
ev_completed_with_tokens("r2", /*total_tokens*/ 10),
]),
sse(vec![
ev_assistant_message("m3", "after switch"),
ev_completed_with_tokens("r3", 100),
ev_completed_with_tokens("r3", /*total_tokens*/ 100),
]),
],
)
@@ -1830,8 +1831,8 @@ async fn pre_sampling_compact_runs_after_resume_and_switch_to_smaller_model() {
&server,
ModelsResponse {
models: vec![
model_info_with_context_window(previous_model, 273_000),
model_info_with_context_window(next_model, 125_000),
model_info_with_context_window(previous_model, /*context_window*/ 273_000),
model_info_with_context_window(next_model, /*context_window*/ 125_000),
],
},
)
@@ -1842,15 +1843,15 @@ async fn pre_sampling_compact_runs_after_resume_and_switch_to_smaller_model() {
vec![
sse(vec![
ev_assistant_message("m1", "before resume"),
ev_completed_with_tokens("r1", 120_000),
ev_completed_with_tokens("r1", /*total_tokens*/ 120_000),
]),
sse(vec![
ev_assistant_message("m2", "PRE_SAMPLING_SUMMARY"),
ev_completed_with_tokens("r2", 10),
ev_completed_with_tokens("r2", /*total_tokens*/ 10),
]),
sse(vec![
ev_assistant_message("m3", "after resume"),
ev_completed_with_tokens("r3", 100),
ev_completed_with_tokens("r3", /*total_tokens*/ 100),
]),
],
)
@@ -1971,22 +1972,22 @@ async fn auto_compact_persists_rollout_entries() {
let sse1 = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 70_000),
ev_completed_with_tokens("r1", /*total_tokens*/ 70_000),
]);
let sse2 = sse(vec![
ev_assistant_message("m2", "SECOND_REPLY"),
ev_completed_with_tokens("r2", 330_000),
ev_completed_with_tokens("r2", /*total_tokens*/ 330_000),
]);
let auto_summary_payload = auto_summary(AUTO_SUMMARY_TEXT);
let sse3 = sse(vec![
ev_assistant_message("m3", &auto_summary_payload),
ev_completed_with_tokens("r3", 200),
ev_completed_with_tokens("r3", /*total_tokens*/ 200),
]);
let sse4 = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 120),
ev_completed_with_tokens("r4", /*total_tokens*/ 120),
]);
let first_matcher = |req: &wiremock::Request| {
@@ -2492,29 +2493,29 @@ async fn auto_compact_allows_multiple_attempts_when_interleaved_with_other_turn_
let sse1 = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 500),
ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]);
let first_summary_payload = auto_summary(FIRST_AUTO_SUMMARY);
let sse2 = sse(vec![
ev_assistant_message("m2", &first_summary_payload),
ev_completed_with_tokens("r2", 50),
ev_completed_with_tokens("r2", /*total_tokens*/ 50),
]);
let sse3 = sse(vec![
ev_function_call(DUMMY_CALL_ID, DUMMY_FUNCTION_NAME, "{}"),
ev_completed_with_tokens("r3", 150),
ev_completed_with_tokens("r3", /*total_tokens*/ 150),
]);
let sse4 = sse(vec![
ev_assistant_message("m4", SECOND_LARGE_REPLY),
ev_completed_with_tokens("r4", 450),
ev_completed_with_tokens("r4", /*total_tokens*/ 450),
]);
let second_summary_payload = auto_summary(SECOND_AUTO_SUMMARY);
let sse5 = sse(vec![
ev_assistant_message("m5", &second_summary_payload),
ev_completed_with_tokens("r5", 60),
ev_completed_with_tokens("r5", /*total_tokens*/ 60),
]);
let sse6 = sse(vec![
ev_assistant_message("m6", FINAL_REPLY),
ev_completed_with_tokens("r6", 120),
ev_completed_with_tokens("r6", /*total_tokens*/ 120),
]);
let follow_up_user = "FOLLOW_UP_AUTO_COMPACT";
let final_user = "FINAL_AUTO_COMPACT";
@@ -2612,11 +2613,11 @@ async fn snapshot_request_shape_mid_turn_continuation_compaction() {
let auto_summary_payload = auto_summary(AUTO_SUMMARY_TEXT);
let auto_compact_turn = sse(vec![
ev_assistant_message("m2", &auto_summary_payload),
ev_completed_with_tokens("r3", 10),
ev_completed_with_tokens("r3", /*total_tokens*/ 10),
]);
let post_auto_compact_turn = sse(vec![
ev_assistant_message("m3", FINAL_REPLY),
ev_completed_with_tokens("r4", 10),
ev_completed_with_tokens("r4", /*total_tokens*/ 10),
]);
// Mount responses in order and keep mocks only for the ones we assert on.
@@ -2716,9 +2717,11 @@ async fn auto_compact_clamps_config_limit_to_context_window() {
let auto_summary_payload = auto_summary(AUTO_SUMMARY_TEXT);
let auto_compact_turn = sse(vec![
ev_assistant_message("m2", &auto_summary_payload),
ev_completed_with_tokens("r2", 10),
ev_completed_with_tokens("r2", /*total_tokens*/ 10),
]);
let post_auto_compact_turn = sse(vec![ev_completed_with_tokens("r3", 10)]);
let post_auto_compact_turn = sse(vec![ev_completed_with_tokens(
"r3", /*total_tokens*/ 10,
)]);
let first_turn_mock = mount_sse_once(&server, first_turn).await;
let auto_compact_mock = mount_sse_once(&server, auto_compact_turn).await;
@@ -2772,15 +2775,15 @@ async fn auto_compact_counts_encrypted_reasoning_before_last_user() {
let first_turn = sse(vec![
ev_reasoning_item("pre-reasoning", &["pre"], &[&pre_last_reasoning_content]),
ev_completed_with_tokens("r1", 10),
ev_completed_with_tokens("r1", /*total_tokens*/ 10),
]);
let second_turn = sse(vec![
ev_reasoning_item("post-reasoning", &["post"], &[&post_last_reasoning_content]),
ev_completed_with_tokens("r2", 80),
ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]);
let third_turn = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 1),
ev_completed_with_tokens("r4", /*total_tokens*/ 1),
]);
let request_log = mount_sse_sequence(
@@ -2898,15 +2901,15 @@ async fn auto_compact_runs_when_reasoning_header_clears_between_turns() {
let first_turn = sse(vec![
ev_reasoning_item("pre-reasoning", &["pre"], &[&pre_last_reasoning_content]),
ev_completed_with_tokens("r1", 10),
ev_completed_with_tokens("r1", /*total_tokens*/ 10),
]);
let second_turn = sse(vec![
ev_reasoning_item("post-reasoning", &["post"], &[&post_last_reasoning_content]),
ev_completed_with_tokens("r2", 80),
ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]);
let third_turn = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 1),
ev_completed_with_tokens("r4", /*total_tokens*/ 1),
]);
let responses = vec![
@@ -2975,19 +2978,19 @@ async fn snapshot_request_shape_pre_turn_compaction_including_incoming_user_mess
let sse1 = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 60),
ev_completed_with_tokens("r1", /*total_tokens*/ 60),
]);
let sse2 = sse(vec![
ev_assistant_message("m2", "SECOND_REPLY"),
ev_completed_with_tokens("r2", 500),
ev_completed_with_tokens("r2", /*total_tokens*/ 500),
]);
let sse3 = sse(vec![
ev_assistant_message("m3", "PRE_TURN_SUMMARY"),
ev_completed_with_tokens("r3", 100),
ev_completed_with_tokens("r3", /*total_tokens*/ 100),
]);
let sse4 = sse(vec![
ev_assistant_message("m4", FINAL_REPLY),
ev_completed_with_tokens("r4", 80),
ev_completed_with_tokens("r4", /*total_tokens*/ 80),
]);
let request_log = mount_sse_sequence(&server, vec![sse1, sse2, sse3, sse4]).await;
@@ -3100,15 +3103,15 @@ async fn snapshot_request_shape_pre_turn_compaction_strips_incoming_model_switch
vec![
sse(vec![
ev_assistant_message("m1", "BEFORE_SWITCH_REPLY"),
ev_completed_with_tokens("r1", 500),
ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
sse(vec![
ev_assistant_message("m2", "PRETURN_SWITCH_SUMMARY"),
ev_completed_with_tokens("r2", 100),
ev_completed_with_tokens("r2", /*total_tokens*/ 100),
]),
sse(vec![
ev_assistant_message("m3", "AFTER_SWITCH_REPLY"),
ev_completed_with_tokens("r3", 100),
ev_completed_with_tokens("r3", /*total_tokens*/ 100),
]),
],
)
@@ -3222,7 +3225,7 @@ async fn snapshot_request_shape_pre_turn_compaction_context_window_exceeded() {
let first_turn = sse(vec![
ev_assistant_message("m1", FIRST_REPLY),
ev_completed_with_tokens("r1", 500),
ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]);
let mut responses = vec![first_turn];
responses.extend(
@@ -3309,11 +3312,11 @@ async fn snapshot_request_shape_manual_compact_without_previous_user_messages()
let compact_turn = sse(vec![
ev_assistant_message("m1", "MANUAL_EMPTY_SUMMARY"),
ev_completed_with_tokens("r1", 90),
ev_completed_with_tokens("r1", /*total_tokens*/ 90),
]);
let follow_up_turn = sse(vec![
ev_assistant_message("m2", FINAL_REPLY),
ev_completed_with_tokens("r2", 80),
ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]);
let request_log = mount_sse_sequence(&server, vec![compact_turn, follow_up_turn]).await;
+36 -30
View File
@@ -353,7 +353,7 @@ async fn remote_compact_runs_automatically() -> Result<()> {
harness.server(),
sse(vec![
responses::ev_shell_command_call("m1", "echo 'hi'"),
responses::ev_completed_with_tokens("resp-1", 100000000), // over token limit
responses::ev_completed_with_tokens("resp-1", /*total_tokens*/ 100000000), // over token limit
]),
)
.await;
@@ -548,7 +548,10 @@ async fn auto_remote_compact_trims_function_call_history_to_fit_context_window()
vec![
sse(vec![
responses::ev_shell_command_call(retained_call_id, retained_command),
responses::ev_completed_with_tokens("retained-call-response", 100),
responses::ev_completed_with_tokens(
"retained-call-response",
/*total_tokens*/ 100,
),
]),
sse(vec![
responses::ev_assistant_message("retained-assistant", "retained complete"),
@@ -556,11 +559,14 @@ async fn auto_remote_compact_trims_function_call_history_to_fit_context_window()
]),
sse(vec![
responses::ev_shell_command_call(trimmed_call_id, trimmed_command),
responses::ev_completed_with_tokens("trimmed-call-response", 100),
responses::ev_completed_with_tokens(
"trimmed-call-response",
/*total_tokens*/ 100,
),
]),
sse(vec![responses::ev_completed_with_tokens(
"trimmed-final-response",
500_000,
/*total_tokens*/ 500_000,
)]),
],
)
@@ -672,7 +678,7 @@ async fn auto_remote_compact_failure_stops_agent_loop() -> Result<()> {
harness.server(),
sse(vec![
responses::ev_assistant_message("initial-assistant", "initial turn complete"),
responses::ev_completed_with_tokens("initial-response", 500_000),
responses::ev_completed_with_tokens("initial-response", /*total_tokens*/ 500_000),
]),
)
.await;
@@ -1468,11 +1474,11 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_restates_realtime_sta
vec![
responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_FIRST_REPLY"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_SECOND_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
],
)
@@ -1601,11 +1607,11 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_restates_realtime_end
vec![
responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_FIRST_REPLY"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_SECOND_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
],
)
@@ -1686,11 +1692,11 @@ async fn snapshot_request_shape_remote_manual_compact_restates_realtime_start()
vec![
responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_FIRST_REPLY"),
responses::ev_completed_with_tokens("r1", 60),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 60),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_SECOND_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
],
)
@@ -1776,15 +1782,15 @@ async fn snapshot_request_shape_remote_mid_turn_compaction_does_not_restate_real
vec![
responses::sse(vec![
responses::ev_assistant_message("setup", "REMOTE_SETUP_REPLY"),
responses::ev_completed_with_tokens("setup-response", 60),
responses::ev_completed_with_tokens("setup-response", /*total_tokens*/ 60),
]),
responses::sse(vec![
responses::ev_function_call("call-remote-mid-turn", DUMMY_FUNCTION_NAME, "{}"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_MID_TURN_FINAL_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
],
)
@@ -1880,11 +1886,11 @@ async fn snapshot_request_shape_remote_compact_resume_restates_realtime_end() ->
vec![
responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_FIRST_REPLY"),
responses::ev_completed_with_tokens("r1", 60),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 60),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_AFTER_RESUME_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
],
)
@@ -1984,15 +1990,15 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_including_incoming_us
vec![
responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_FIRST_REPLY"),
responses::ev_completed_with_tokens("r1", 60),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 60),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_SECOND_REPLY"),
responses::ev_completed_with_tokens("r2", 500),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 500),
]),
responses::sse(vec![
responses::ev_assistant_message("m3", "REMOTE_FINAL_REPLY"),
responses::ev_completed_with_tokens("r3", 80),
responses::ev_completed_with_tokens("r3", /*total_tokens*/ 80),
]),
],
)
@@ -2088,7 +2094,7 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_strips_incoming_model
harness.server(),
responses::sse(vec![
responses::ev_assistant_message("m1", "BEFORE_SWITCH_REPLY"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
)
.await;
@@ -2096,7 +2102,7 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_strips_incoming_model
harness.server(),
responses::sse(vec![
responses::ev_assistant_message("m2", "AFTER_SWITCH_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
)
.await;
@@ -2224,7 +2230,7 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_context_window_exceed
harness.server(),
vec![responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_FIRST_REPLY"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
])],
)
.await;
@@ -2243,7 +2249,7 @@ async fn snapshot_request_shape_remote_pre_turn_compaction_context_window_exceed
harness.server(),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_POST_COMPACT_SHOULD_NOT_RUN"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
)
.await;
@@ -2325,11 +2331,11 @@ async fn snapshot_request_shape_remote_mid_turn_continuation_compaction() -> Res
vec![
responses::sse(vec![
responses::ev_function_call("call-remote-mid-turn", DUMMY_FUNCTION_NAME, "{}"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_MID_TURN_FINAL_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
],
)
@@ -2394,7 +2400,7 @@ async fn snapshot_request_shape_remote_mid_turn_compaction_summary_only_reinject
harness.server(),
responses::sse(vec![
responses::ev_function_call("call-remote-summary-only", DUMMY_FUNCTION_NAME, "{}"),
responses::ev_completed_with_tokens("r1", 500),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 500),
]),
)
.await;
@@ -2402,7 +2408,7 @@ async fn snapshot_request_shape_remote_mid_turn_compaction_summary_only_reinject
harness.server(),
responses::sse(vec![
responses::ev_assistant_message("m2", "REMOTE_SUMMARY_ONLY_FINAL_REPLY"),
responses::ev_completed_with_tokens("r2", 80),
responses::ev_completed_with_tokens("r2", /*total_tokens*/ 80),
]),
)
.await;
@@ -2477,7 +2483,7 @@ async fn snapshot_request_shape_remote_mid_turn_compaction_multi_summary_reinjec
harness.server(),
responses::sse(vec![
responses::ev_assistant_message("setup", "REMOTE_SETUP_REPLY"),
responses::ev_completed_with_tokens("setup-response", 60),
responses::ev_completed_with_tokens("setup-response", /*total_tokens*/ 60),
]),
)
.await;
@@ -2485,7 +2491,7 @@ async fn snapshot_request_shape_remote_mid_turn_compaction_multi_summary_reinjec
harness.server(),
responses::sse(vec![
responses::ev_shell_command_call("call-remote-multi-summary", "echo multi-summary"),
responses::ev_completed_with_tokens("r1", 1_000),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 1_000),
]),
)
.await;
@@ -2580,7 +2586,7 @@ async fn snapshot_request_shape_remote_manual_compact_without_previous_user_mess
harness.server(),
responses::sse(vec![
responses::ev_assistant_message("m1", "REMOTE_MANUAL_EMPTY_FOLLOW_UP_REPLY"),
responses::ev_completed_with_tokens("r1", 80),
responses::ev_completed_with_tokens("r1", /*total_tokens*/ 80),
]),
)
.await;
@@ -180,7 +180,7 @@ async fn compact_resume_and_fork_preserve_model_history_view() {
"compact+resume test expects resumed path {resumed_path:?} to exist",
);
let forked = fork_thread(&manager, &config, resumed_path, 2).await;
let forked = fork_thread(&manager, &config, resumed_path, /*nth_user_message*/ 2).await;
user_turn(&forked, "AFTER_FORK").await;
// 3. Capture the requests to the model and validate the history slices.
@@ -316,7 +316,7 @@ async fn compact_resume_after_second_compaction_preserves_history() -> Result<()
let request_log = mount_second_compact_sequence(&server).await;
// 2. Drive the conversation through compact -> resume -> fork -> compact -> resume.
let (_home, config, manager, base) = start_test_conversation(&server, None).await;
let (_home, config, manager, base) = start_test_conversation(&server, /*model*/ None).await;
user_turn(&base, "hello world").await;
compact_conversation(&base).await;
@@ -335,7 +335,7 @@ async fn compact_resume_after_second_compaction_preserves_history() -> Result<()
"second compact test expects resumed path {resumed_path:?} to exist",
);
let forked = fork_thread(&manager, &config, resumed_path, 3).await;
let forked = fork_thread(&manager, &config, resumed_path, /*nth_user_message*/ 3).await;
user_turn(&forked, "AFTER_FORK").await;
compact_conversation(&forked).await;
@@ -468,7 +468,7 @@ async fn snapshot_rollback_past_compaction_replays_append_only_history() -> Resu
let request_log = mount_sse_sequence(&server, vec![sse1, sse2, sse3, sse4]).await;
let (_home, _config, _manager, base) = start_test_conversation(&server, None).await;
let (_home, _config, _manager, base) = start_test_conversation(&server, /*model*/ None).await;
user_turn(&base, "hello world").await;
compact_conversation(&base).await;
@@ -845,10 +845,15 @@ async fn resume_conversation(
let auth_manager = codex_core::test_support::auth_manager_from_auth(
codex_core::CodexAuth::from_api_key("dummy"),
);
Box::pin(manager.resume_thread_from_rollout(config.clone(), path, auth_manager, None))
.await
.expect("resume conversation")
.thread
Box::pin(manager.resume_thread_from_rollout(
config.clone(),
path,
auth_manager,
/*parent_trace*/ None,
))
.await
.expect("resume conversation")
.thread
}
#[cfg(test)]
@@ -858,8 +863,14 @@ async fn fork_thread(
path: std::path::PathBuf,
nth_user_message: usize,
) -> Arc<CodexThread> {
Box::pin(manager.fork_thread(nth_user_message, config.clone(), path, false, None))
.await
.expect("fork conversation")
.thread
Box::pin(manager.fork_thread(
nth_user_message,
config.clone(),
path,
/*persist_extended_history*/ false,
/*parent_trace*/ None,
))
.await
.expect("fork conversation")
.thread
}
+4 -3
View File
@@ -30,7 +30,8 @@ fn skip_test() -> bool {
#[expect(clippy::expect_used)]
async fn run_test_cmd(tmp: TempDir, cmd: Vec<&str>) -> Result<ExecToolCallOutput> {
let sandbox_type = get_platform_sandbox(false).expect("should be able to get sandbox type");
let sandbox_type = get_platform_sandbox(/*windows_sandbox_enabled*/ false)
.expect("should be able to get sandbox type");
assert_eq!(sandbox_type, SandboxType::MacosSeatbelt);
let params = ExecParams {
@@ -56,8 +57,8 @@ async fn run_test_cmd(tmp: TempDir, cmd: Vec<&str>) -> Result<ExecToolCallOutput
NetworkSandboxPolicy::from(&policy),
tmp.path(),
&None,
false,
None,
/*use_legacy_landlock*/ false,
/*stdout_stream*/ None,
)
.await
}
+8 -4
View File
@@ -1321,7 +1321,8 @@ async fn pre_tool_use_does_not_fire_for_non_shell_tools() -> Result<()> {
let mut builder = test_codex()
.with_pre_build_hook(|home| {
if let Err(error) = write_pre_tool_use_hook(home, None, "json_deny", "should not fire")
if let Err(error) =
write_pre_tool_use_hook(home, /*matcher*/ None, "json_deny", "should not fire")
{
panic!("failed to write pre tool use hook test fixture: {error}");
}
@@ -1775,9 +1776,12 @@ async fn post_tool_use_does_not_fire_for_non_shell_tools() -> Result<()> {
let mut builder = test_codex()
.with_pre_build_hook(|home| {
if let Err(error) =
write_post_tool_use_hook(home, None, "decision_block", "should not fire")
{
if let Err(error) = write_post_tool_use_hook(
home,
/*matcher*/ None,
"decision_block",
"should not fire",
) {
panic!("failed to write post tool use hook test fixture: {error}");
}
})
+1 -1
View File
@@ -148,7 +148,7 @@ async fn copy_paste_local_image_persists_rollout_request_shape() -> anyhow::Resu
role: "user".to_string(),
content: vec![
ContentItem::InputText {
text: codex_protocol::models::local_image_open_tag_text(1),
text: codex_protocol::models::local_image_open_tag_text(/*label_number*/ 1),
},
ContentItem::InputImage { image_url },
ContentItem::InputText {
+15 -6
View File
@@ -253,7 +253,9 @@ async fn web_search_pollution_moves_selected_thread_into_removed_phase2_inputs()
.resume(&server, home.clone(), rollout_path.clone())
.await?;
let first_phase2_request = wait_for_request(&responses, 1).await.remove(0);
let first_phase2_request = wait_for_request(&responses, /*expected_count*/ 1)
.await
.remove(0);
let first_phase2_prompt = phase2_prompt_text(&first_phase2_request);
assert!(
first_phase2_prompt.contains("- selected inputs this run: 1"),
@@ -295,7 +297,9 @@ async fn web_search_pollution_moves_selected_thread_into_removed_phase2_inputs()
let selection = {
let deadline = Instant::now() + Duration::from_secs(10);
loop {
let selection = db.get_phase2_input_selection(1, 30).await?;
let selection = db
.get_phase2_input_selection(/*n*/ 1, /*max_unused_days*/ 30)
.await?;
if selection.selected.is_empty()
&& selection.retained_thread_ids.is_empty()
&& selection.removed.len() == 1
@@ -340,7 +344,7 @@ async fn build_test_codex(server: &wiremock::MockServer, home: Arc<TempDir>) ->
async fn init_state_db(home: &Arc<TempDir>) -> Result<Arc<codex_state::StateRuntime>> {
let db =
codex_state::StateRuntime::init(home.path().to_path_buf(), "test-provider".into()).await?;
db.mark_backfill_complete(None).await?;
db.mark_backfill_complete(/*last_watermark*/ None).await?;
Ok(db)
}
@@ -379,7 +383,7 @@ async fn seed_stage1_output(
}
async fn wait_for_single_request(mock: &ResponseMock) -> ResponsesRequest {
wait_for_request(mock, 1).await.remove(0)
wait_for_request(mock, /*expected_count*/ 1).await.remove(0)
}
async fn wait_for_request(mock: &ResponseMock, expected_count: usize) -> Vec<ResponsesRequest> {
@@ -412,7 +416,9 @@ async fn wait_for_phase2_success(
) -> Result<()> {
let deadline = Instant::now() + Duration::from_secs(10);
loop {
let selection = db.get_phase2_input_selection(1, 30).await?;
let selection = db
.get_phase2_input_selection(/*n*/ 1, /*max_unused_days*/ 30)
.await?;
if selection.selected.len() == 1
&& selection.selected[0].thread_id == expected_thread_id
&& selection.retained_thread_ids == vec![expected_thread_id]
@@ -439,7 +445,10 @@ async fn seed_stage1_output_for_existing_thread(
) -> Result<()> {
let owner = ThreadId::new();
let claim = db
.try_claim_stage1_job(thread_id, owner, updated_at, 3_600, 64)
.try_claim_stage1_job(
thread_id, owner, updated_at, /*lease_seconds*/ 3_600,
/*max_running_jobs*/ 64,
)
.await?;
let ownership_token = match claim {
codex_state::Stage1JobClaimOutcome::Claimed { ownership_token } => ownership_token,
@@ -16,7 +16,7 @@ async fn offline_model_info_without_tool_output_override() {
let manager = ModelsManager::new(
config.codex_home.clone(),
auth_manager,
None,
/*model_catalog*/ None,
CollaborationModesConfig::default(),
);
@@ -24,7 +24,7 @@ async fn offline_model_info_without_tool_output_override() {
assert_eq!(
model_info.truncation_policy,
TruncationPolicyConfig::bytes(10_000)
TruncationPolicyConfig::bytes(/*limit*/ 10_000)
);
}
@@ -39,7 +39,7 @@ async fn offline_model_info_with_tool_output_override() {
let manager = ModelsManager::new(
config.codex_home.clone(),
auth_manager,
None,
/*model_catalog*/ None,
CollaborationModesConfig::default(),
);
@@ -47,6 +47,6 @@ async fn offline_model_info_with_tool_output_override() {
assert_eq!(
model_info.truncation_policy,
TruncationPolicyConfig::tokens(123)
TruncationPolicyConfig::tokens(/*limit*/ 123)
);
}
+8 -8
View File
@@ -92,7 +92,7 @@ fn test_model_info(
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -308,7 +308,7 @@ async fn service_tier_change_is_applied_on_next_http_turn() -> Result<()> {
test.submit_turn_with_service_tier("fast turn", Some(ServiceTier::Fast))
.await?;
test.submit_turn_with_service_tier("standard turn", None)
test.submit_turn_with_service_tier("standard turn", /*service_tier*/ None)
.await?;
let requests = resp_mock.requests();
@@ -498,7 +498,7 @@ async fn generated_image_is_replayed_for_image_capable_models() -> Result<()> {
sse(vec![
ev_response_created("resp-1"),
ev_image_generation_call("ig_123", "completed", "lobster", "Zm9v"),
ev_completed_with_tokens("resp-1", 10),
ev_completed_with_tokens("resp-1", /*total_tokens*/ 10),
]),
sse_completed("resp-2"),
],
@@ -630,7 +630,7 @@ async fn model_change_from_generated_image_to_text_preserves_prior_generated_ima
sse(vec![
ev_response_created("resp-1"),
ev_image_generation_call("ig_123", "completed", "lobster", "Zm9v"),
ev_completed_with_tokens("resp-1", 10),
ev_completed_with_tokens("resp-1", /*total_tokens*/ 10),
]),
sse_completed("resp-2"),
],
@@ -764,7 +764,7 @@ async fn thread_rollback_after_generated_image_drops_entire_image_turn_history()
sse(vec![
ev_response_created("resp-1"),
ev_image_generation_call("ig_rollback", "completed", "lobster", "Zm9v"),
ev_completed_with_tokens("resp-1", 10),
ev_completed_with_tokens("resp-1", /*total_tokens*/ 10),
]),
sse_completed("resp-2"),
],
@@ -908,7 +908,7 @@ async fn model_switch_to_smaller_model_updates_token_context_window() -> Result<
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(large_context_window),
@@ -935,11 +935,11 @@ async fn model_switch_to_smaller_model_updates_token_context_window() -> Result<
vec![
sse(vec![
ev_response_created("resp-1"),
ev_completed_with_tokens("resp-1", 100),
ev_completed_with_tokens("resp-1", /*total_tokens*/ 100),
]),
sse(vec![
ev_response_created("resp-2"),
ev_completed_with_tokens("resp-2", 120),
ev_completed_with_tokens("resp-2", /*total_tokens*/ 120),
]),
],
)
@@ -44,7 +44,7 @@ const DIFFERENT_VERSION_MODEL: &str = "codex-test-different-version";
async fn renews_cache_ttl_on_matching_models_etag() -> Result<()> {
let server = MockServer::start().await;
let remote_model = test_remote_model(REMOTE_MODEL, 1);
let remote_model = test_remote_model(REMOTE_MODEL, /*priority*/ 1);
let models_mock = responses::mount_models_once_with_etag(
&server,
ModelsResponse {
@@ -138,11 +138,11 @@ async fn renews_cache_ttl_on_matching_models_etag() -> Result<()> {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn uses_cache_when_version_matches() -> Result<()> {
let server = MockServer::start().await;
let cached_model = test_remote_model(VERSIONED_MODEL, 1);
let cached_model = test_remote_model(VERSIONED_MODEL, /*priority*/ 1);
let models_mock = responses::mount_models_once(
&server,
ModelsResponse {
models: vec![test_remote_model("remote", 2)],
models: vec![test_remote_model("remote", /*priority*/ 2)],
},
)
.await;
@@ -185,11 +185,11 @@ async fn uses_cache_when_version_matches() -> Result<()> {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn refreshes_when_cache_version_missing() -> Result<()> {
let server = MockServer::start().await;
let cached_model = test_remote_model(MISSING_VERSION_MODEL, 1);
let cached_model = test_remote_model(MISSING_VERSION_MODEL, /*priority*/ 1);
let models_mock = responses::mount_models_once(
&server,
ModelsResponse {
models: vec![test_remote_model("remote-missing", 2)],
models: vec![test_remote_model("remote-missing", /*priority*/ 2)],
},
)
.await;
@@ -232,9 +232,9 @@ async fn refreshes_when_cache_version_missing() -> Result<()> {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn refreshes_when_cache_version_differs() -> Result<()> {
let server = MockServer::start().await;
let cached_model = test_remote_model(DIFFERENT_VERSION_MODEL, 1);
let cached_model = test_remote_model(DIFFERENT_VERSION_MODEL, /*priority*/ 1);
let models_response = ModelsResponse {
models: vec![test_remote_model("remote-different", 2)],
models: vec![test_remote_model("remote-different", /*priority*/ 2)],
};
let mut models_mocks = Vec::new();
for _ in 0..3 {
@@ -344,7 +344,7 @@ fn test_remote_model(slug: &str, priority: i32) -> ModelInfo {
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -503,8 +503,8 @@ async fn permissions_message_includes_writable_roots() -> Result<()> {
test.config.approvals_reviewer,
&Policy::empty(),
test.config.cwd.as_path(),
false,
false,
/*exec_permission_approvals_enabled*/ false,
/*request_permissions_tool_enabled*/ false,
)
.into_text();
// Normalize line endings to handle Windows vs Unix differences
+2 -2
View File
@@ -661,7 +661,7 @@ async fn remote_model_friendly_personality_instructions_with_feature() -> anyhow
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(128_000),
@@ -777,7 +777,7 @@ async fn user_turn_personality_remote_model_template_includes_update_message() -
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(128_000),
@@ -172,7 +172,11 @@ async fn conversation_start_audio_text_close_round_trip() -> Result<()> {
let mut builder = test_codex();
let test = builder.build_with_websocket_server(&server).await?;
assert!(server.wait_for_handshakes(1, Duration::from_secs(2)).await);
assert!(
server
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
test.codex
.submit(Op::RealtimeConversationStart(ConversationStartParams {
@@ -310,7 +314,11 @@ async fn conversation_start_uses_openai_env_key_fallback_with_chatgpt_auth() ->
let mut builder = test_codex().with_auth(CodexAuth::create_dummy_chatgpt_auth_for_testing());
let test = builder.build_with_websocket_server(&server).await?;
assert!(server.wait_for_handshakes(1, Duration::from_secs(2)).await);
assert!(
server
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
test.codex
.submit(Op::RealtimeConversationStart(ConversationStartParams {
@@ -365,7 +373,11 @@ async fn conversation_transport_close_emits_closed_event() -> Result<()> {
let mut builder = test_codex();
let test = builder.build_with_websocket_server(&server).await?;
assert!(server.wait_for_handshakes(1, Duration::from_secs(2)).await);
assert!(
server
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
test.codex
.submit(Op::RealtimeConversationStart(ConversationStartParams {
@@ -572,7 +584,11 @@ async fn conversation_second_start_replaces_runtime() -> Result<()> {
.await;
let mut builder = test_codex();
let test = builder.build_with_websocket_server(&server).await?;
assert!(server.wait_for_handshakes(1, Duration::from_secs(2)).await);
assert!(
server
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
test.codex
.submit(Op::RealtimeConversationStart(ConversationStartParams {
@@ -672,7 +688,7 @@ async fn conversation_uses_experimental_realtime_ws_base_url_override() -> Resul
let test = builder.build_with_websocket_server(&startup_server).await?;
assert!(
startup_server
.wait_for_handshakes(1, Duration::from_secs(2))
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
@@ -724,7 +740,11 @@ async fn conversation_uses_experimental_realtime_ws_backend_prompt_override() ->
config.experimental_realtime_ws_backend_prompt = Some("prompt from config".to_string());
});
let test = builder.build_with_websocket_server(&server).await?;
assert!(server.wait_for_handshakes(1, Duration::from_secs(2)).await);
assert!(
server
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
test.codex
.submit(Op::RealtimeConversationStart(ConversationStartParams {
@@ -784,7 +804,7 @@ async fn conversation_uses_experimental_realtime_ws_startup_context_override() -
fs::write(test.workspace_path("README.md"), "workspace marker")?;
assert!(
startup_server
.wait_for_handshakes(1, Duration::from_secs(2))
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
@@ -844,7 +864,7 @@ async fn conversation_disables_realtime_startup_context_with_empty_override() ->
fs::write(test.workspace_path("README.md"), "workspace marker")?;
assert!(
startup_server
.wait_for_handshakes(1, Duration::from_secs(2))
.wait_for_handshakes(/*expected*/ 1, Duration::from_secs(2))
.await
);
@@ -1238,7 +1258,9 @@ async fn conversation_handoff_persists_across_item_done_until_turn_complete() ->
})
.await;
let first_append = realtime_server.wait_for_request(0, 1).await;
let first_append = realtime_server
.wait_for_request(/*connection_index*/ 0, /*request_index*/ 1)
.await;
assert_eq!(
first_append.body_json()["type"].as_str(),
Some("conversation.handoff.append")
@@ -1262,7 +1284,9 @@ async fn conversation_handoff_persists_across_item_done_until_turn_complete() ->
let _ = gate_second_message_tx.send(());
let second_append = realtime_server.wait_for_request(0, 2).await;
let second_append = realtime_server
.wait_for_request(/*connection_index*/ 0, /*request_index*/ 2)
.await;
assert_eq!(
second_append.body_json()["type"].as_str(),
Some("conversation.handoff.append")
@@ -1777,7 +1801,9 @@ async fn delegated_turn_user_role_echo_does_not_redelegate_and_still_forwards_au
"delegate now"
);
let mirrored_request = realtime_server.wait_for_request(0, 1).await;
let mirrored_request = realtime_server
.wait_for_request(/*connection_index*/ 0, /*request_index*/ 1)
.await;
let mirrored_request_body = mirrored_request.body_json();
eprintln!(
"[realtime test +{}ms] saw mirrored request type={:?} handoff_id={:?} text={:?}",
+37 -27
View File
@@ -62,14 +62,14 @@ async fn remote_models_get_model_info_uses_longest_matching_prefix() -> Result<(
let generic = test_remote_model_with_policy(
"gpt-5.3",
ModelVisibility::List,
1_000,
TruncationPolicyConfig::bytes(10_000),
/*priority*/ 1_000,
TruncationPolicyConfig::bytes(/*limit*/ 10_000),
);
let specific = test_remote_model_with_policy(
"gpt-5.3-codex",
ModelVisibility::List,
1_000,
TruncationPolicyConfig::bytes(10_000),
/*priority*/ 1_000,
TruncationPolicyConfig::bytes(/*limit*/ 10_000),
);
let specific = ModelInfo {
display_name: "GPT 5.3 Codex".to_string(),
@@ -95,7 +95,7 @@ async fn remote_models_get_model_info_uses_longest_matching_prefix() -> Result<(
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -124,8 +124,8 @@ async fn remote_models_long_model_slug_is_sent_with_high_reasoning() -> Result<(
let mut remote_model = test_remote_model_with_policy(
prefix_model,
ModelVisibility::List,
1_000,
TruncationPolicyConfig::bytes(10_000),
/*priority*/ 1_000,
TruncationPolicyConfig::bytes(/*limit*/ 10_000),
);
remote_model.default_reasoning_level = Some(ReasoningEffort::High);
remote_model.supported_reasoning_levels = vec![
@@ -304,7 +304,7 @@ async fn remote_models_remote_model_uses_unified_exec() -> Result<()> {
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -434,8 +434,8 @@ async fn remote_models_truncation_policy_without_override_preserves_remote() ->
let remote_model = test_remote_model_with_policy(
slug,
ModelVisibility::List,
1,
TruncationPolicyConfig::bytes(12_000),
/*priority*/ 1,
TruncationPolicyConfig::bytes(/*limit*/ 12_000),
);
mount_models_once(
&server,
@@ -458,7 +458,7 @@ async fn remote_models_truncation_policy_without_override_preserves_remote() ->
let model_info = models_manager.get_model_info(slug, &test.config).await;
assert_eq!(
model_info.truncation_policy,
TruncationPolicyConfig::bytes(12_000)
TruncationPolicyConfig::bytes(/*limit*/ 12_000)
);
Ok(())
@@ -478,8 +478,8 @@ async fn remote_models_truncation_policy_with_tool_output_override() -> Result<(
let remote_model = test_remote_model_with_policy(
slug,
ModelVisibility::List,
1,
TruncationPolicyConfig::bytes(10_000),
/*priority*/ 1,
TruncationPolicyConfig::bytes(/*limit*/ 10_000),
);
mount_models_once(
&server,
@@ -503,7 +503,7 @@ async fn remote_models_truncation_policy_with_tool_output_override() -> Result<(
let model_info = models_manager.get_model_info(slug, &test.config).await;
assert_eq!(
model_info.truncation_policy,
TruncationPolicyConfig::bytes(200)
TruncationPolicyConfig::bytes(/*limit*/ 200)
);
Ok(())
@@ -548,7 +548,7 @@ async fn remote_models_apply_remote_base_instructions() -> Result<()> {
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -642,7 +642,8 @@ async fn remote_models_do_not_append_removed_builtin_presets() -> Result<()> {
skip_if_sandbox!(Ok(()));
let server = MockServer::start().await;
let remote_model = test_remote_model("remote-alpha", ModelVisibility::List, 0);
let remote_model =
test_remote_model("remote-alpha", ModelVisibility::List, /*priority*/ 0);
let models_mock = mount_models_once(
&server,
ModelsResponse {
@@ -656,7 +657,7 @@ async fn remote_models_do_not_append_removed_builtin_presets() -> Result<()> {
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -699,7 +700,11 @@ async fn remote_models_merge_adds_new_high_priority_first() -> Result<()> {
skip_if_sandbox!(Ok(()));
let server = MockServer::start().await;
let remote_model = test_remote_model("remote-top", ModelVisibility::List, -10_000);
let remote_model = test_remote_model(
"remote-top",
ModelVisibility::List,
/*priority*/ -10_000,
);
let models_mock = mount_models_once(
&server,
ModelsResponse {
@@ -713,7 +718,7 @@ async fn remote_models_merge_adds_new_high_priority_first() -> Result<()> {
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -744,7 +749,7 @@ async fn remote_models_merge_replaces_overlapping_model() -> Result<()> {
let server = MockServer::start().await;
let slug = bundled_model_slug();
let mut remote_model = test_remote_model(&slug, ModelVisibility::List, 0);
let mut remote_model = test_remote_model(&slug, ModelVisibility::List, /*priority*/ 0);
remote_model.display_name = "Overridden".to_string();
remote_model.description = Some("Overridden description".to_string());
let models_mock = mount_models_once(
@@ -760,7 +765,7 @@ async fn remote_models_merge_replaces_overlapping_model() -> Result<()> {
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -804,7 +809,7 @@ async fn remote_models_merge_preserves_bundled_models_on_empty_response() -> Res
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -830,7 +835,8 @@ async fn remote_models_request_times_out_after_5s() -> Result<()> {
skip_if_sandbox!(Ok(()));
let server = MockServer::start().await;
let remote_model = test_remote_model("remote-timeout", ModelVisibility::List, 0);
let remote_model =
test_remote_model("remote-timeout", ModelVisibility::List, /*priority*/ 0);
let models_mock = mount_models_once_with_delay(
&server,
ModelsResponse {
@@ -845,7 +851,7 @@ async fn remote_models_request_times_out_after_5s() -> Result<()> {
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -897,7 +903,11 @@ async fn remote_models_hide_picker_only_models() -> Result<()> {
skip_if_sandbox!(Ok(()));
let server = MockServer::start().await;
let remote_model = test_remote_model("codex-auto-balanced", ModelVisibility::Hide, 0);
let remote_model = test_remote_model(
"codex-auto-balanced",
ModelVisibility::Hide,
/*priority*/ 0,
);
let models_mock = mount_models_once(
&server,
ModelsResponse {
@@ -911,7 +921,7 @@ async fn remote_models_hide_picker_only_models() -> Result<()> {
let auth = CodexAuth::create_dummy_chatgpt_auth_for_testing();
let provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers(/* openai_base_url */ None)["openai"].clone()
..built_in_model_providers(/* openai_base_url */ /*openai_base_url*/ None)["openai"].clone()
};
let manager = codex_core::test_support::models_manager_with_provider(
codex_home.path().to_path_buf(),
@@ -982,7 +992,7 @@ fn test_remote_model(slug: &str, visibility: ModelVisibility, priority: i32) ->
slug,
visibility,
priority,
TruncationPolicyConfig::bytes(10_000),
TruncationPolicyConfig::bytes(/*limit*/ 10_000),
)
}
+7 -1
View File
@@ -98,7 +98,13 @@ async fn emits_warning_when_resumed_model_differs() {
thread: conversation,
..
} = thread_manager
.resume_thread_with_history(config, initial_history, auth_manager, false, None)
.resume_thread_with_history(
config,
initial_history,
auth_manager,
/*persist_extended_history*/ false,
/*parent_trace*/ None,
)
.await
.expect("resume conversation");
+18 -9
View File
@@ -70,7 +70,8 @@ async fn review_op_emits_lifecycle_and_review_output() {
]"#;
let review_json_escaped = serde_json::to_string(&review_json).unwrap();
let sse_raw = sse_template.replace("__REVIEW__", &review_json_escaped);
let (server, _request_log) = start_responses_server_with_sse(&sse_raw, 1).await;
let (server, _request_log) =
start_responses_server_with_sse(&sse_raw, /*expected_requests*/ 1).await;
let codex_home = Arc::new(TempDir::new().unwrap());
let codex = new_conversation_for_server(&server, codex_home.clone(), |_| {}).await;
@@ -192,7 +193,8 @@ async fn review_op_with_plain_text_emits_review_fallback() {
}},
{"type":"response.completed", "response": {"id": "__ID__"}}
]"#;
let (server, _request_log) = start_responses_server_with_sse(sse_raw, 1).await;
let (server, _request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 1).await;
let codex_home = Arc::new(TempDir::new().unwrap());
let codex = new_conversation_for_server(&server, codex_home.clone(), |_| {}).await;
@@ -253,7 +255,8 @@ async fn review_filters_agent_message_related_events() {
}},
{"type":"response.completed", "response": {"id": "__ID__"}}
]"#;
let (server, _request_log) = start_responses_server_with_sse(sse_raw, 1).await;
let (server, _request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 1).await;
let codex_home = Arc::new(TempDir::new().unwrap());
let codex = new_conversation_for_server(&server, codex_home.clone(), |_| {}).await;
@@ -335,7 +338,8 @@ async fn review_does_not_emit_agent_message_on_structured_output() {
]"#;
let review_json_escaped = serde_json::to_string(&review_json).unwrap();
let sse_raw = sse_template.replace("__REVIEW__", &review_json_escaped);
let (server, _request_log) = start_responses_server_with_sse(&sse_raw, 1).await;
let (server, _request_log) =
start_responses_server_with_sse(&sse_raw, /*expected_requests*/ 1).await;
let codex_home = Arc::new(TempDir::new().unwrap());
let codex = new_conversation_for_server(&server, codex_home.clone(), |_| {}).await;
@@ -390,7 +394,8 @@ async fn review_uses_custom_review_model_from_config() {
let sse_raw = r#"[
{"type":"response.completed", "response": {"id": "__ID__"}}
]"#;
let (server, request_log) = start_responses_server_with_sse(sse_raw, 1).await;
let (server, request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 1).await;
let codex_home = Arc::new(TempDir::new().unwrap());
// Choose a review model different from the main model; ensure it is used.
let codex = new_conversation_for_server(&server, codex_home.clone(), |cfg| {
@@ -444,7 +449,8 @@ async fn review_uses_session_model_when_review_model_unset() {
let sse_raw = r#"[
{"type":"response.completed", "response": {"id": "__ID__"}}
]"#;
let (server, request_log) = start_responses_server_with_sse(sse_raw, 1).await;
let (server, request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 1).await;
let codex_home = Arc::new(TempDir::new().unwrap());
let codex = new_conversation_for_server(&server, codex_home.clone(), |cfg| {
cfg.model = Some("gpt-4.1".to_string());
@@ -498,7 +504,8 @@ async fn review_input_isolated_from_parent_history() {
let sse_raw = r#"[
{"type":"response.completed", "response": {"id": "__ID__"}}
]"#;
let (server, request_log) = start_responses_server_with_sse(sse_raw, 1).await;
let (server, request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 1).await;
// Seed a parent session history via resume file with both user + assistant items.
let codex_home = Arc::new(TempDir::new().unwrap());
@@ -681,7 +688,8 @@ async fn review_history_surfaces_in_parent_session() {
}},
{"type":"response.completed", "response": {"id": "__ID__"}}
]"#;
let (server, request_log) = start_responses_server_with_sse(sse_raw, 2).await;
let (server, request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 2).await;
let codex_home = Arc::new(TempDir::new().unwrap());
let codex = new_conversation_for_server(&server, codex_home.clone(), |_| {}).await;
@@ -773,7 +781,8 @@ async fn review_uses_overridden_cwd_for_base_branch_merge_base() {
skip_if_no_network!();
let sse_raw = r#"[{"type":"response.completed", "response": {"id": "__ID__"}}]"#;
let (server, request_log) = start_responses_server_with_sse(sse_raw, 1).await;
let (server, request_log) =
start_responses_server_with_sse(sse_raw, /*expected_requests*/ 1).await;
let initial_cwd = TempDir::new().unwrap();
+1 -1
View File
@@ -415,7 +415,7 @@ async fn stdio_image_responses_are_sanitized_for_text_only_model() -> anyhow::Re
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -60,7 +60,10 @@ async fn upsert_thread_metadata(codex_home: &Path, thread_id: ThreadId, rollout_
let runtime = StateRuntime::init(codex_home.to_path_buf(), "test-provider".to_string())
.await
.unwrap();
runtime.mark_backfill_complete(None).await.unwrap();
runtime
.mark_backfill_complete(/*last_watermark*/ None)
.await
.unwrap();
let mut builder = ThreadMetadataBuilder::new(
thread_id,
rollout_path,
@@ -168,14 +171,14 @@ async fn find_locates_rollout_file_written_by_recorder() -> std::io::Result<()>
&config,
RolloutRecorderParams::new(
thread_id,
None,
/*forked_from_id*/ None,
SessionSource::Exec,
BaseInstructions::default(),
Vec::new(),
EventPersistenceMode::Limited,
),
None,
None,
/*state_db_ctx*/ None,
/*state_builder*/ None,
)
.await?;
recorder.persist().await?;
+3 -3
View File
@@ -192,7 +192,7 @@ assert os.read(master, 4) == b"ping""#
&policy,
sandbox_cwd.as_path(),
StdioPolicy::RedirectForShellTool,
None,
/*network*/ None,
HashMap::new(),
)
.await
@@ -245,7 +245,7 @@ async fn java_home_finds_runtime_under_seatbelt() {
&policy,
sandbox_cwd.as_path(),
StdioPolicy::RedirectForShellTool,
None,
/*network*/ None,
env,
)
.await
@@ -302,7 +302,7 @@ async fn touch(path: &Path, policy: &SandboxPolicy) -> bool {
policy,
sandbox_cwd.as_path(),
StdioPolicy::RedirectForShellTool,
None,
/*network*/ None,
HashMap::new(),
)
.await
+15 -3
View File
@@ -112,7 +112,13 @@ async fn shell_command_works() -> anyhow::Result<()> {
let harness = shell_command_harness_with(|builder| builder.with_model("gpt-5.1")).await?;
let call_id = "shell-command-call";
mount_shell_responses(&harness, call_id, "echo 'hello, world'", None).await;
mount_shell_responses(
&harness,
call_id,
"echo 'hello, world'",
/*login*/ None,
)
.await;
harness.submit("run the echo command").await?;
let output = harness.function_call_stdout(call_id).await;
@@ -183,7 +189,13 @@ async fn pipe_output_with_login() -> anyhow::Result<()> {
let harness = shell_command_harness_with(|builder| builder.with_model("gpt-5.1")).await?;
let call_id = "shell-command-call-second-extra-no-login";
mount_shell_responses(&harness, call_id, "echo 'hello, world' | cat", None).await;
mount_shell_responses(
&harness,
call_id,
"echo 'hello, world' | cat",
/*login*/ None,
)
.await;
harness.submit("run the command without login").await?;
let output = harness.function_call_stdout(call_id).await;
@@ -224,7 +236,7 @@ async fn shell_command_times_out_with_timeout_ms() -> anyhow::Result<()> {
&harness,
call_id,
command,
None,
/*login*/ None,
Duration::from_millis(200),
)
.await;
@@ -125,7 +125,11 @@ async fn shell_output_stays_json_without_freeform_apply_patch(
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = configure_shell_model(test_codex(), output_type, false);
let mut builder = configure_shell_model(
test_codex(),
output_type,
/*include_apply_patch_tool*/ false,
);
let test = builder.build(&server).await?;
let call_id = "shell-json";
@@ -177,7 +181,11 @@ async fn shell_output_is_structured_with_freeform_apply_patch(
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = configure_shell_model(test_codex(), output_type, true);
let mut builder = configure_shell_model(
test_codex(),
output_type,
/*include_apply_patch_tool*/ true,
);
let test = builder.build(&server).await?;
let call_id = "shell-structured";
@@ -222,7 +230,11 @@ async fn shell_output_preserves_fixture_json_without_serialization(
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = configure_shell_model(test_codex(), output_type, false);
let mut builder = configure_shell_model(
test_codex(),
output_type,
/*include_apply_patch_tool*/ false,
);
let test = builder.build(&server).await?;
let fixture_path = test.cwd.path().join("fixture.json");
@@ -286,7 +298,11 @@ async fn shell_output_structures_fixture_with_serialization(
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = configure_shell_model(test_codex(), output_type, true);
let mut builder = configure_shell_model(
test_codex(),
output_type,
/*include_apply_patch_tool*/ true,
);
let test = builder.build(&server).await?;
let fixture_path = test.cwd.path().join("fixture.json");
@@ -345,7 +361,11 @@ async fn shell_output_for_freeform_tool_records_duration(
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = configure_shell_model(test_codex(), output_type, true);
let mut builder = configure_shell_model(
test_codex(),
output_type,
/*include_apply_patch_tool*/ true,
);
let test = builder.build(&server).await?;
let call_id = "shell-structured";
@@ -395,10 +415,14 @@ async fn shell_output_reserializes_truncated_content(output_type: ShellModelOutp
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder =
configure_shell_model(test_codex(), output_type, true).with_config(move |config| {
config.tool_output_token_limit = Some(200);
});
let mut builder = configure_shell_model(
test_codex(),
output_type,
/*include_apply_patch_tool*/ true,
)
.with_config(move |config| {
config.tool_output_token_limit = Some(200);
});
let test = builder.build(&server).await?;
let call_id = "shell-truncated";
@@ -77,7 +77,7 @@ fn test_model_info(
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -144,7 +144,7 @@ async fn setup_turn_one_with_spawned_child(
"message": CHILD_PROMPT,
}),
child_response_delay,
true,
/*wait_for_parent_notification*/ true,
|builder| builder,
)
.await
@@ -253,9 +253,14 @@ async fn spawn_child_and_capture_snapshot(
core_test_support::test_codex::TestCodexBuilder,
) -> core_test_support::test_codex::TestCodexBuilder,
) -> Result<ThreadConfigSnapshot> {
let (test, spawned_id) =
setup_turn_one_with_custom_spawned_child(server, spawn_args, None, false, configure_test)
.await?;
let (test, spawned_id) = setup_turn_one_with_custom_spawned_child(
server,
spawn_args,
/*child_response_delay*/ None,
/*wait_for_parent_notification*/ false,
configure_test,
)
.await?;
let thread_id = ThreadId::from_string(&spawned_id)?;
Ok(test
.thread_manager
@@ -270,7 +275,8 @@ async fn subagent_notification_is_included_without_wait() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let (test, _spawned_id) = setup_turn_one_with_spawned_child(&server, None).await?;
let (test, _spawned_id) =
setup_turn_one_with_spawned_child(&server, /*child_response_delay*/ None).await?;
let turn2 = mount_sse_once_match(
&server,
+2 -2
View File
@@ -321,7 +321,7 @@ async fn collect_tools(use_unified_exec: bool) -> Result<Vec<String>> {
async fn unified_exec_spec_toggle_end_to_end() -> Result<()> {
skip_if_no_network!(Ok(()));
let tools_disabled = collect_tools(false).await?;
let tools_disabled = collect_tools(/*use_unified_exec*/ false).await?;
assert!(
!tools_disabled.iter().any(|name| name == "exec_command"),
"tools list should not include exec_command when disabled: {tools_disabled:?}"
@@ -331,7 +331,7 @@ async fn unified_exec_spec_toggle_end_to_end() -> Result<()> {
"tools list should not include write_stdin when disabled: {tools_disabled:?}"
);
let tools_enabled = collect_tools(true).await?;
let tools_enabled = collect_tools(/*use_unified_exec*/ true).await?;
assert!(
tools_enabled.iter().any(|name| name == "exec_command"),
"tools list should include exec_command when enabled: {tools_enabled:?}"
@@ -42,7 +42,13 @@ async fn emits_warning_when_unstable_features_enabled_via_config() {
thread: conversation,
..
} = thread_manager
.resume_thread_with_history(config, InitialHistory::New, auth_manager, false, None)
.resume_thread_with_history(
config,
InitialHistory::New,
auth_manager,
/*persist_extended_history*/ false,
/*parent_trace*/ None,
)
.await
.expect("spawn conversation");
@@ -83,7 +89,13 @@ async fn suppresses_warning_when_configured() {
thread: conversation,
..
} = thread_manager
.resume_thread_with_history(config, InitialHistory::New, auth_manager, false, None)
.resume_thread_with_history(
config,
InitialHistory::New,
auth_manager,
/*persist_extended_history*/ false,
/*parent_trace*/ None,
)
.await
.expect("spawn conversation");
+17 -3
View File
@@ -476,7 +476,14 @@ async fn view_image_tool_errors_clearly_for_unsupported_detail_values() -> anyho
} = &test;
let rel_path = "assets/unsupported-detail.png";
write_workspace_png(&test, rel_path, 256, 128, [0u8, 80, 255, 255]).await?;
write_workspace_png(
&test,
rel_path,
/*width*/ 256,
/*height*/ 128,
[0u8, 80, 255, 255],
)
.await?;
let call_id = "view-image-unsupported-detail";
let arguments = serde_json::json!({ "path": rel_path, "detail": "low" }).to_string();
@@ -1352,7 +1359,7 @@ async fn view_image_tool_returns_unsupported_message_for_text_only_model() -> an
availability_nux: None,
apply_patch_tool_type: None,
web_search_tool_type: Default::default(),
truncation_policy: TruncationPolicyConfig::bytes(10_000),
truncation_policy: TruncationPolicyConfig::bytes(/*limit*/ 10_000),
supports_parallel_tool_calls: false,
supports_image_detail_original: false,
context_window: Some(272_000),
@@ -1377,7 +1384,14 @@ async fn view_image_tool_returns_unsupported_message_for_text_only_model() -> an
let TestCodex { codex, config, .. } = &test;
let rel_path = "assets/example.png";
write_workspace_png(&test, rel_path, 20, 20, [255u8, 0, 0, 255]).await?;
write_workspace_png(
&test,
rel_path,
/*width*/ 20,
/*height*/ 20,
[255u8, 0, 0, 255],
)
.await?;
let call_id = "view-image-unsupported-model";
let arguments = serde_json::json!({ "path": rel_path }).to_string();