openai/codex
1
0
Fork 0
mirror of https://github.com/openai/codex.git synced 2026-11-05 20:10:33 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'main' into jif/client-2

Browse Source 
# Conflicts:
#	codex-rs/core/src/client.rs
#	codex-rs/core/src/client/chat_completions.rs
#	codex-rs/core/src/client_common.rs
This commit is contained in:
jif-oai
2025-11-14 14:04:43 +01:00
parent c8a6e4ddd7 6c384eb9c6
commit cf15b86f5c
79 changed files with 2575 additions and 3148 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/close-stale-contributor-prs.yml
+105
View File
@@ -0,0 +1,105 @@
name: Close stale contributor PRs
on:
workflow_dispatch:
schedule:
- cron: "0 6 * * *"
permissions:
contents: read
issues: write
pull-requests: write
jobs:
close-stale-contributor-prs:
runs-on: ubuntu-latest
steps:
- name: Close inactive PRs from contributors
uses: actions/github-script@v7
with:
github-token: ${{ secrets.GITHUB_TOKEN }}
script: |
const DAYS_INACTIVE = 14;
const cutoff = new Date(Date.now() - DAYS_INACTIVE * 24 * 60 * 60 * 1000);
const { owner, repo } = context.repo;
const dryRun = false;
const stalePrs = [];
core.info(`Dry run mode: ${dryRun}`);
const prs = await github.paginate(github.rest.pulls.list, {
owner,
repo,
state: "open",
per_page: 100,
sort: "updated",
direction: "asc",
});
for (const pr of prs) {
const lastUpdated = new Date(pr.updated_at);
if (lastUpdated > cutoff) {
core.info(`PR ${pr.number} is fresh`);
continue;
}
if (!pr.user || pr.user.type !== "User") {
core.info(`PR ${pr.number} wasn't created by a user`);
continue;
}
let permission;
try {
const permissionResponse = await github.rest.repos.getCollaboratorPermissionLevel({
owner,
repo,
username: pr.user.login,
});
permission = permissionResponse.data.permission;
} catch (error) {
if (error.status === 404) {
core.info(`Author ${pr.user.login} is not a collaborator; skipping #${pr.number}`);
continue;
}
throw error;
}
const hasContributorAccess = ["admin", "maintain", "write"].includes(permission);
if (!hasContributorAccess) {
core.info(`Author ${pr.user.login} has ${permission} access; skipping #${pr.number}`);
continue;
}
stalePrs.push(pr);
}
if (!stalePrs.length) {
core.info("No stale contributor pull requests found.");
return;
}
for (const pr of stalePrs) {
const issue_number = pr.number;
const closeComment = `Closing this pull request because it has had no updates for more than ${DAYS_INACTIVE} days. If you plan to continue working on it, feel free to reopen or open a new PR.`;
if (dryRun) {
core.info(`[dry-run] Would close contributor PR #${issue_number} from ${pr.user.login}`);
continue;
}
await github.rest.issues.createComment({
owner,
repo,
issue_number,
body: closeComment,
});
await github.rest.pulls.update({
owner,
repo,
pull_number: issue_number,
state: "closed",
});
core.info(`Closed contributor PR #${issue_number} from ${pr.user.login}`);
}
.github/workflows/rust-ci.yml
+1 -1
View File
@@ -406,10 +406,10 @@ jobs:
- name: tests
id: test
continue-on-error: true
run: cargo nextest run --all-features --no-fail-fast --target ${{ matrix.target }} --cargo-profile ci-test
env:
RUST_BACKTRACE: 1
NEXTEST_STATUS_LEVEL: leak
- name: Save cargo home cache
if: always() && !cancelled() && steps.cache_cargo_home_restore.outputs.cache-hit != 'true'
AGENTS.md
+1
View File
@@ -85,6 +85,7 @@ If you don’t have the tool:
- `ResponsesRequest` exposes helpers (`body_json`, `input`, `function_call_output`, `custom_tool_call_output`, `call_output`, `header`, `path`, `query_param`) so assertions can target structured payloads instead of manual JSON digging.
- Build SSE payloads with the provided `ev_*` constructors and the `sse(...)`.
- Prefer `wait_for_event` over `wait_for_event_with_timeout`.
- Prefer `mount_sse_once` over `mount_sse_once_match` or `mount_sse_sequence`
- Typical pattern:
codex-rs/Cargo.lock
Generated
+34
View File
@@ -1115,6 +1115,7 @@ dependencies = [
"similar",
"strum_macros 0.27.2",
"tempfile",
"test-case",
"test-log",
"thiserror 2.0.17",
"time",
@@ -6161,6 +6162,39 @@ version = "0.5.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8f50febec83f5ee1df3015341d8bd429f2d1cc62bcba7ea2076759d315084683"
[[package]]
name = "test-case"
version = "3.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "eb2550dd13afcd286853192af8601920d959b14c401fcece38071d53bf0768a8"
dependencies = [
"test-case-macros",
]
[[package]]
name = "test-case-core"
version = "3.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "adcb7fd841cd518e279be3d5a3eb0636409487998a4aff22f3de87b81e88384f"
dependencies = [
"cfg-if",
"proc-macro2",
"quote",
"syn 2.0.104",
]
[[package]]
name = "test-case-macros"
version = "3.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5c89e72a01ed4c579669add59014b9a524d609c0c88c6a585ce37485879f6ffb"
dependencies = [
"proc-macro2",
"quote",
"syn 2.0.104",
"test-case-core",
]
[[package]]
name = "test-log"
version = "0.2.18"
codex-rs/app-server-protocol/src/export.rs
+8 -19
View File
@@ -13,10 +13,7 @@ use crate::export_server_responses;
use anyhow::Context;
use anyhow::Result;
use anyhow::anyhow;
use codex_protocol::parse_command::ParsedCommand;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::FileChange;
use codex_protocol::protocol::SandboxPolicy;
use schemars::JsonSchema;
use schemars::schema_for;
use serde::Serialize;
@@ -120,10 +117,6 @@ pub fn generate_json(out_dir: &Path) -> Result<()> {
|d| write_json_schema_with_return::<crate::ClientNotification>(d, "ClientNotification"),
|d| write_json_schema_with_return::<crate::ServerNotification>(d, "ServerNotification"),
|d| write_json_schema_with_return::<EventMsg>(d, "EventMsg"),
|d| write_json_schema_with_return::<FileChange>(d, "FileChange"),
|d| write_json_schema_with_return::<crate::protocol::v1::InputItem>(d, "InputItem"),
|d| write_json_schema_with_return::<ParsedCommand>(d, "ParsedCommand"),
|d| write_json_schema_with_return::<SandboxPolicy>(d, "SandboxPolicy"),
];
let mut schemas: Vec<GeneratedSchema> = Vec::new();
@@ -152,13 +145,10 @@ fn build_schema_bundle(schemas: Vec<GeneratedSchema>) -> Result<Value> {
"ClientNotification",
"ClientRequest",
"EventMsg",
"FileChange",
"InputItem",
"ParsedCommand",
"SandboxPolicy",
"ServerNotification",
"ServerRequest",
];
const IGNORED_DEFINITIONS: &[&str] = &["Option<()>"];
let namespaced_types = collect_namespaced_types(&schemas);
let mut definitions = Map::new();
@@ -171,6 +161,10 @@ fn build_schema_bundle(schemas: Vec<GeneratedSchema>) -> Result<Value> {
in_v1_dir,
} = schema;
if IGNORED_DEFINITIONS.contains(&logical_name.as_str()) {
continue;
}
if let Some(ref ns) = namespace {
rewrite_refs_to_namespace(&mut value, ns);
}
@@ -181,6 +175,9 @@ fn build_schema_bundle(schemas: Vec<GeneratedSchema>) -> Result<Value> {
&& let Value::Object(defs_obj) = defs
{
for (def_name, mut def_schema) in defs_obj {
if IGNORED_DEFINITIONS.contains(&def_name.as_str()) {
continue;
}
if SPECIAL_DEFINITIONS.contains(&def_name.as_str()) {
continue;
}
@@ -386,14 +383,6 @@ fn variant_definition_name(base: &str, variant: &Value) -> Option<String> {
});
}
if let Some(mode_literal) = literal_from_property(props, "mode") {
let pascal = to_pascal_case(mode_literal);
return Some(match base {
"SandboxPolicy" => format!("{pascal}SandboxPolicy"),
_ => format!("{pascal}{base}"),
});
}
if props.len() == 1
&& let Some(key) = props.keys().next()
{
codex-rs/app-server-protocol/src/protocol/common.rs
+3
View File
@@ -508,6 +508,9 @@ server_notification_definitions! {
McpToolCallProgress => "item/mcpToolCall/progress" (v2::McpToolCallProgressNotification),
AccountUpdated => "account/updated" (v2::AccountUpdatedNotification),
AccountRateLimitsUpdated => "account/rateLimits/updated" (v2::AccountRateLimitsUpdatedNotification),
ReasoningSummaryTextDelta => "item/reasoning/summaryTextDelta" (v2::ReasoningSummaryTextDeltaNotification),
ReasoningSummaryPartAdded => "item/reasoning/summaryPartAdded" (v2::ReasoningSummaryPartAddedNotification),
ReasoningTextDelta => "item/reasoning/textDelta" (v2::ReasoningTextDeltaNotification),
#[serde(rename = "account/login/completed")]
#[ts(rename = "account/login/completed")]
codex-rs/app-server-protocol/src/protocol/v2.rs
+39 -15
View File
@@ -57,8 +57,8 @@ v2_enum_from_core!(
);
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema, TS)]
#[serde(tag = "mode", rename_all = "camelCase")]
#[ts(tag = "mode")]
#[serde(tag = "type", rename_all = "camelCase")]
#[ts(tag = "type")]
#[ts(export_to = "v2/")]
pub enum SandboxPolicy {
DangerFullAccess,
@@ -516,7 +516,10 @@ pub enum ThreadItem {
},
Reasoning {
id: String,
text: String,
#[serde(default)]
summary: Vec<String>,
#[serde(default)]
content: Vec<String>,
},
CommandExecution {
id: String,
@@ -575,17 +578,11 @@ impl From<CoreTurnItem> for ThreadItem {
.collect::<String>();
ThreadItem::AgentMessage { id: agent.id, text }
}
CoreTurnItem::Reasoning(reasoning) => {
let text = if !reasoning.summary_text.is_empty() {
reasoning.summary_text.join("\n")
} else {
reasoning.raw_content.join("\n")
};
ThreadItem::Reasoning {
id: reasoning.id,
text,
}
}
CoreTurnItem::Reasoning(reasoning) => ThreadItem::Reasoning {
id: reasoning.id,
summary: reasoning.summary_text,
content: reasoning.raw_content,
},
CoreTurnItem::WebSearch(search) => ThreadItem::WebSearch {
id: search.id,
query: search.query,
@@ -719,6 +716,32 @@ pub struct AgentMessageDeltaNotification {
pub delta: String,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct ReasoningSummaryTextDeltaNotification {
pub item_id: String,
pub delta: String,
pub summary_index: i64,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct ReasoningSummaryPartAddedNotification {
pub item_id: String,
pub summary_index: i64,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct ReasoningTextDeltaNotification {
pub item_id: String,
pub delta: String,
pub content_index: i64,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
@@ -867,7 +890,8 @@ mod tests {
ThreadItem::from(reasoning_item),
ThreadItem::Reasoning {
id: "reasoning-1".to_string(),
text: "line one\nline two".to_string(),
summary: vec!["line one".to_string(), "line two".to_string()],
content: vec![],
}
);
codex-rs/app-server/src/codex_message_processor.rs
+46
View File
@@ -12,6 +12,7 @@ use codex_app_server_protocol::AccountRateLimitsUpdatedNotification;
use codex_app_server_protocol::AccountUpdatedNotification;
use codex_app_server_protocol::AddConversationListenerParams;
use codex_app_server_protocol::AddConversationSubscriptionResponse;
use codex_app_server_protocol::AgentMessageDeltaNotification;
use codex_app_server_protocol::ApplyPatchApprovalParams;
use codex_app_server_protocol::ApplyPatchApprovalResponse;
use codex_app_server_protocol::ArchiveConversationParams;
@@ -62,6 +63,9 @@ use codex_app_server_protocol::ModelListParams;
use codex_app_server_protocol::ModelListResponse;
use codex_app_server_protocol::NewConversationParams;
use codex_app_server_protocol::NewConversationResponse;
use codex_app_server_protocol::ReasoningSummaryPartAddedNotification;
use codex_app_server_protocol::ReasoningSummaryTextDeltaNotification;
use codex_app_server_protocol::ReasoningTextDeltaNotification;
use codex_app_server_protocol::RemoveConversationListenerParams;
use codex_app_server_protocol::RemoveConversationSubscriptionResponse;
use codex_app_server_protocol::RequestId;
@@ -2681,6 +2685,48 @@ async fn apply_bespoke_event_handling(
on_patch_approval_response(event_id, rx, conversation).await;
});
}
EventMsg::AgentMessageContentDelta(event) => {
let notification = AgentMessageDeltaNotification {
item_id: event.item_id,
delta: event.delta,
};
outgoing
.send_server_notification(ServerNotification::AgentMessageDelta(notification))
.await;
}
EventMsg::ReasoningContentDelta(event) => {
let notification = ReasoningSummaryTextDeltaNotification {
item_id: event.item_id,
delta: event.delta,
summary_index: event.summary_index,
};
outgoing
.send_server_notification(ServerNotification::ReasoningSummaryTextDelta(
notification,
))
.await;
}
EventMsg::ReasoningRawContentDelta(event) => {
let notification = ReasoningTextDeltaNotification {
item_id: event.item_id,
delta: event.delta,
content_index: event.content_index,
};
outgoing
.send_server_notification(ServerNotification::ReasoningTextDelta(notification))
.await;
}
EventMsg::AgentReasoningSectionBreak(event) => {
let notification = ReasoningSummaryPartAddedNotification {
item_id: event.item_id,
summary_index: event.summary_index,
};
outgoing
.send_server_notification(ServerNotification::ReasoningSummaryPartAdded(
notification,
))
.await;
}
EventMsg::ExecApprovalRequest(ExecApprovalRequestEvent {
call_id,
command,
codex-rs/app-server/tests/suite/v2/thread_list.rs
+7 -42
View File
@@ -6,10 +6,8 @@ use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::ThreadListParams;
use codex_app_server_protocol::ThreadListResponse;
use serde_json::json;
use tempfile::TempDir;
use tokio::time::timeout;
use uuid::Uuid;
const DEFAULT_READ_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(10);
@@ -150,46 +148,13 @@ async fn thread_list_respects_provider_filter() -> Result<()> {
"X",
Some("mock_provider"),
)?; // mock_provider
// one with a different provider
let uuid = Uuid::new_v4();
let dir = codex_home
.path()
.join("sessions")
.join("2025")
.join("01")
.join("02");
std::fs::create_dir_all(&dir)?;
let file_path = dir.join(format!("rollout-2025-01-02T11-00-00-{uuid}.jsonl"));
let lines = [
json!({
"timestamp": "2025-01-02T11:00:00Z",
"type": "session_meta",
"payload": {
"id": uuid,
"timestamp": "2025-01-02T11:00:00Z",
"cwd": "/",
"originator": "codex",
"cli_version": "0.0.0",
"instructions": null,
"source": "vscode",
"model_provider": "other_provider"
}
})
.to_string(),
json!({
"timestamp": "2025-01-02T11:00:00Z",
"type":"response_item",
"payload": {"type":"message","role":"user","content":[{"type":"input_text","text":"X"}]}
})
.to_string(),
json!({
"timestamp": "2025-01-02T11:00:00Z",
"type":"event_msg",
"payload": {"type":"user_message","message":"X","kind":"plain"}
})
.to_string(),
];
std::fs::write(file_path, lines.join("\n") + "\n")?;
let _b = create_fake_rollout(
codex_home.path(),
"2025-01-02T11-00-00",
"2025-01-02T11:00:00Z",
"X",
Some("other_provider"),
)?;
let mut mcp = McpProcess::new(codex_home.path()).await?;
timeout(DEFAULT_READ_TIMEOUT, mcp.initialize()).await??;
codex-rs/core/Cargo.toml
+1
View File
@@ -60,6 +60,7 @@ shlex = { workspace = true }
similar = { workspace = true }
strum_macros = { workspace = true }
tempfile = { workspace = true }
test-case = "3.3.1"
test-log = { workspace = true }
thiserror = { workspace = true }
time = { workspace = true, features = [
codex-rs/core/gpt_5_1_prompt.md
+96 -36
View File
@@ -1,4 +1,4 @@
You are a coding agent running in the Codex CLI, a terminal-based coding assistant. Codex CLI is an open source project led by OpenAI. You are expected to be precise, safe, and helpful.
You are GPT-5.1 running in the Codex CLI, a terminal-based coding assistant. Codex CLI is an open source project led by OpenAI. You are expected to be precise, safe, and helpful.
Your capabilities:
@@ -26,17 +26,28 @@ Your default personality and tone is concise, direct, and friendly. You communic
- Direct system/developer/user instructions (as part of a prompt) take precedence over AGENTS.md instructions.
- The contents of the AGENTS.md file at the root of the repo and any directories from the CWD up to the root are included with the developer message and don't need to be re-read. When working in a subdirectory of CWD, or a directory outside the CWD, check for any AGENTS.md files that may be applicable.
## Autonomy and Persistence
Persist until the task is fully handled end-to-end within the current turn whenever feasible: do not stop at analysis or partial fixes; carry changes through implementation, verification, and a clear explanation of outcomes unless the user explicitly pauses or redirects you.
Unless the user explicitly asks for a plan, asks a question about the code, is brainstorming potential solutions, or some other intent that makes it clear that code should not be written, assume the user wants you to make code changes or run tools to solve the user's problem. In these cases, it's bad to output your proposed solution in a message, you should go ahead and actually implement the change. If you encounter challenges or blockers, you should attempt to resolve them yourself.
## Responsiveness
### Preamble messages
### User Updates Spec
You'll work for stretches with tool calls — it's critical to keep the user updated as you work.
Before making tool calls, send a brief preamble to the user explaining what you’re about to do. When sending preamble messages, follow these principles and examples:
Frequency & Length:
- Send short updates (1–2 sentences) whenever there is a meaningful, important insight you need to share with the user to keep them informed.
- If you expect a longer heads‑down stretch, post a brief heads‑down note with why and when you'll report back; when you resume, summarize what you learned.
- Only the initial plan, plan updates, and final recap can be longer, with multiple bullets and paragraphs
- **Logically group related actions**: if you’re about to run several related commands, describe them together in one preamble rather than sending a separate note for each.
- **Keep it concise**: be no more than 1-2 sentences, focused on immediate, tangible next steps. (8–12 words for quick updates).
- **Build on prior context**: if this is not your first tool call, use the preamble message to connect the dots with what’s been done so far and create a sense of momentum and clarity for the user to understand your next actions.
- **Keep your tone light, friendly and curious**: add small touches of personality in preambles feel collaborative and engaging.
- **Exception**: Avoid adding a preamble for every trivial read (e.g., `cat` a single file) unless it’s part of a larger grouped action.
Tone:
- Friendly, confident, senior-engineer energy. Positive, collaborative, humble; fix mistakes quickly.
Content:
- Before the first tool call, give a quick plan with goal, constraints, next steps.
- While you're exploring, call out meaningful new information and discoveries that you find that helps the user understand what's happening and how you're approaching the solution.
- If you change the plan (e.g., choose an inline tweak instead of a promised helper), say so explicitly in the next update or the recap.
**Examples:**
@@ -59,6 +70,8 @@ Do not repeat the full contents of the plan after an `update_plan` call — the
Before running a command, consider whether or not you have completed the previous step, and make sure to mark it as completed before moving on to the next step. It may be the case that you complete all steps in your plan after a single pass of implementation. If this is the case, you can simply mark all the planned steps as completed. Sometimes, you may need to change plans in the middle of a task: call `update_plan` with the updated plan and make sure to provide an `explanation` of the rationale when doing so.
Maintain statuses in the tool: exactly one item in_progress at a time; mark items complete when done; post timely status transitions. Do not jump an item from pending to completed: always set it to in_progress first. Do not batch-complete multiple items after the fact. Finish with all items completed or explicitly canceled/deferred before ending the turn. Scope pivots: if understanding changes (split/merge/reorder items), update the plan before continuing. Do not let the plan go stale while coding.
Use a plan when:
- The task is non-trivial and will require multiple actions over a long time horizon.
@@ -122,14 +135,14 @@ If you need to write a plan, only write high quality plans, not low quality ones
## Task execution
You are a coding agent. Please keep going until the query is completely resolved, before ending your turn and yielding back to the user. Only terminate your turn when you are sure that the problem is solved. Autonomously resolve the query to the best of your ability, using the tools available to you, before coming back to the user. Do NOT guess or make up an answer.
You are a coding agent. You must keep going until the query or task is completely resolved, before ending your turn and yielding back to the user. Persist until the task is fully handled end-to-end within the current turn whenever feasible and persevere even when function calls fail. Only terminate your turn when you are sure that the problem is solved. Autonomously resolve the query to the best of your ability, using the tools available to you, before coming back to the user. Do NOT guess or make up an answer.
You MUST adhere to the following criteria when solving queries:
- Working on the repo(s) in the current environment is allowed, even if they are proprietary.
- Analyzing code for vulnerabilities is allowed.
- Showing user code and tool call details is allowed.
- Use the `apply_patch` tool to edit files (NEVER try `applypatch` or `apply-patch`, only `apply_patch`): {"command":["apply_patch","*** Begin Patch\\n*** Update File: path/to/file.py\\n@@ def example():\\n- pass\\n+ return 123\\n*** End Patch"]}
- Use the `apply_patch` tool to edit files (NEVER try `applypatch` or `apply-patch`, only `apply_patch`). This is a FREEFORM tool, so do not wrap the patch in JSON.
If completing the user's task requires writing or modifying files, your code and final answer should follow these coding guidelines, though user instructions (i.e. AGENTS.md) may override these guidelines:
@@ -146,44 +159,46 @@ If completing the user's task requires writing or modifying files, your code and
- Do not use one-letter variable names unless explicitly requested.
- NEVER output inline citations like "【F:README.md†L5-L14】" in your outputs. The CLI is not able to render these so they will just be broken in the UI. Instead, if you output valid filepaths, users will be able to click on them to open the files in their editor.
## Sandbox and approvals
## Codex CLI harness, sandboxing, and approvals
The Codex CLI harness supports several different sandboxing, and approval configurations that the user can choose from.
The Codex CLI harness supports several different configurations for sandboxing and escalation approvals that the user can choose from.
Filesystem sandboxing prevents you from editing files without user approval. The options are:
Filesystem sandboxing defines which files can be read or written. The options for `sandbox_mode` are:
- **read-only**: The sandbox only permits reading files.
- **workspace-write**: The sandbox permits reading files, and editing files in `cwd` and `writable_roots`. Editing files in other directories requires approval.
- **danger-full-access**: No filesystem sandboxing - all commands are permitted.
- **read-only**: You can only read files.
- **workspace-write**: You can read files. You can write to files in your workspace folder, but not outside it.
- **danger-full-access**: No filesystem sandboxing.
Network sandboxing prevents you from accessing network without approval. Options are
- **restricted**
- **enabled**
Approvals are your mechanism to get user consent to perform more privileged actions. Although they introduce friction to the user because your work is paused until the user responds, you should leverage them to accomplish your important work. Do not let these settings or the sandbox deter you from attempting to accomplish the user's task. Approval options are
Network sandboxing defines whether network can be accessed without approval. Options for `network_access` are:
- **restricted**: Requires approval
- **enabled**: No approval needed
Approvals are your mechanism to get user consent to run shell commands without the sandbox. Possible configuration options for `approval_policy` are
- **untrusted**: The harness will escalate most commands for user approval, apart from a limited allowlist of safe "read" commands.
- **on-failure**: The harness will allow all commands to run in the sandbox (if enabled), and failures will be escalated to the user for approval to run again without the sandbox.
- **on-request**: Commands will be run in the sandbox by default, and you can specify in your tool call if you want to escalate a command to run without sandboxing. (Note that this mode is not always available. If it is, you'll see parameters for it in the `shell` command description.)
- **never**: This is a non-interactive mode where you may NEVER ask the user for approval to run commands. Instead, you must always persist and work around constraints to solve the task for the user. You MUST do your utmost best to finish the task and validate your work before yielding. If this mode is pared with `danger-full-access`, take advantage of it to deliver the best outcome for the user. Further, in this mode, your default testing philosophy is overridden: Even if you don't see local patterns for testing, you may add tests and scripts to validate your work. Just remove them before yielding.
- **on-request**: Commands will be run in the sandbox by default, and you can specify in your tool call if you want to escalate a command to run without sandboxing. (Note that this mode is not always available. If it is, you'll see parameters for escalating in the tool definition.)
- **never**: This is a non-interactive mode where you may NEVER ask the user for approval to run commands. Instead, you must always persist and work around constraints to solve the task for the user. You MUST do your utmost best to finish the task and validate your work before yielding. If this mode is paired with `danger-full-access`, take advantage of it to deliver the best outcome for the user. Further, in this mode, your default testing philosophy is overridden: Even if you don't see local patterns for testing, you may add tests and scripts to validate your work. Just remove them before yielding.
When you are running with approvals `on-request`, and sandboxing enabled, here are scenarios where you'll need to request approval:
- You need to run a command that writes to a directory that requires it (e.g. running tests that write to /tmp)
When you are running with `approval_policy == on-request`, and sandboxing enabled, here are scenarios where you'll need to request approval:
- You need to run a command that writes to a directory that requires it (e.g. running tests that write to /var)
- You need to run a GUI app (e.g., open/xdg-open/osascript) to open browsers or files.
- You are running sandboxed and need to run a command that requires network access (e.g. installing packages)
- If you run a command that is important to solving the user's query, but it fails because of sandboxing, rerun the command with approval.
- If you run a command that is important to solving the user's query, but it fails because of sandboxing, rerun the command with approval. ALWAYS proceed to use the `with_escalated_permissions` and `justification` parameters. Within this harness, prefer requesting approval via the tool over asking in natural language.
- You are about to take a potentially destructive action such as an `rm` or `git reset` that the user did not explicitly ask for
- (For all of these, you should weigh alternative paths that do not require approval.)
- (for all of these, you should weigh alternative paths that do not require approval)
Note that when sandboxing is set to read-only, you'll need to request approval for any command that isn't a read.
When `sandbox_mode` is set to read-only, you'll need to request approval for any command that isn't a read.
You will be told what filesystem sandboxing, network sandboxing, and approval mode are active in a developer or user message. If you are not told about this, assume that you are running with workspace-write, network sandboxing ON, and approval on-failure.
You will be told what filesystem sandboxing, network sandboxing, and approval mode are active in a developer or user message. If you are not told about this, assume that you are running with workspace-write, network sandboxing enabled, and approval on-failure.
Although they introduce friction to the user because your work is paused until the user responds, you should leverage them when necessary to accomplish important work. If the completing the task requires escalated permissions, Do not let these settings or the sandbox deter you from attempting to accomplish the user's task unless it is set to "never", in which case never ask for approvals.
When requesting approval to execute a command that will require escalated privileges:
- Provide the `with_escalated_permissions` parameter with the boolean value true
- Include a short, 1 sentence explanation for why you need to enable `with_escalated_permissions` in the justification parameter
## Validating your work
If the codebase has tests or the ability to build or run, consider using them to verify that your work is complete.
If the codebase has tests or the ability to build or run, consider using them to verify changes once your work is complete.
When testing, your philosophy should be to start as specific as possible to the code you changed so that you can catch issues efficiently, then make your way to broader tests as you build confidence. If there's no test for the code you changed, and if the adjacent patterns in the codebases show that there's a logical place for you to add a test, you may do so. However, do not add tests to codebases with no tests.
@@ -193,7 +208,7 @@ For all of testing, running, building, and formatting, do not attempt to fix unr
Be mindful of whether to run validation commands proactively. In the absence of behavioral guidance:
- When running in non-interactive approval modes like **never** or **on-failure**, proactively run tests, lint and do whatever you need to ensure you've completed the task.
- When running in non-interactive approval modes like **never** or **on-failure**, you can proactively run tests, lint and do whatever you need to ensure you've completed the task. If you are unable to run tests, you must still do your utmost best to complete the task.
- When working in interactive approval modes like **untrusted**, or **on-request**, hold off on running tests or lint commands until the user is ready for you to finalize your output, because these commands take time to run and slow down iteration. Instead suggest what you want to do next, and let the user confirm first.
- When working on test-related tasks, such as adding tests, fixing tests, or reproducing a bug to verify behavior, you may proactively run tests regardless of approval mode. Use your judgement to decide whether this is a test-related task.
@@ -219,7 +234,7 @@ Your final message should read naturally, like an update from a concise teammate
You can skip heavy formatting for single, simple actions or confirmations. In these cases, respond in plain sentences with any relevant next step or quick option. Reserve multi-section structured responses for results that need grouping or explanation.
The user is working on the same computer as you, and has access to your work. As such there's no need to show the full contents of large files you have already written unless the user explicitly asks for them. Similarly, if you've created or modified files using `apply_patch`, there's no need to tell users to "save the file" or "copy the code into a file"—just reference the file path.
The user is working on the same computer as you, and has access to your work. As such there's no need to show the contents of files you have already written unless the user explicitly asks for them. Similarly, if you've created or modified files using `apply_patch`, there's no need to tell users to "save the file" or "copy the code into a file"—just reference the file path.
If there's something that you think you could help with as a logical next step, concisely ask the user if they want you to do so. Good examples of this are running tests, committing changes, or building out the next logical component. If there’s something that you couldn't do (even with approval) but that the user might want to do (such as verifying changes by running the app), include those instructions succinctly.
@@ -247,7 +262,7 @@ You are producing plain text that will later be styled by the CLI. Follow these
**Monospace**
- Wrap all commands, file paths, env vars, and code identifiers in backticks (`` `...` ``).
- Wrap all commands, file paths, env vars, code identifiers, and code samples in backticks (`` `...` ``).
- Apply to inline examples and to bullet keywords if the keyword itself is a literal file/command.
- Never mix monospace and bold markers; choose one based on whether it’s a keyword (`**`) or inline code/path (`` ` ``).
@@ -278,6 +293,13 @@ When referencing files in your response, make sure to include the relevant start
- Keep descriptions self-contained; don’t refer to “above” or “below”.
- Use parallel structure in lists for consistency.
**Verbosity**
- Final answer compactness rules (enforced):
- Tiny/small single-file change (≤ ~10 lines): 2–5 sentences or ≤3 bullets. No headings. 0–1 short snippet (≤3 lines) only if essential.
- Medium change (single area or a few files): ≤6 bullets or 6–10 sentences. At most 1–2 short snippets total (≤8 lines each).
- Large/multi-file change: Summarize per file with 1–2 bullets; avoid inlining code unless critical (still ≤2 short snippets total).
- Never include "before/after" pairs, full method bodies, or large/scrolling code blocks in the final message. Prefer referencing file/symbol names instead.
**Don’t**
- Don’t use literal words “bold” or “monospace” in the content.
@@ -296,9 +318,47 @@ For casual greetings, acknowledgements, or other one-off conversational messages
When using the shell, you must adhere to the following guidelines:
- The arguments to `shell` will be passed to execvp().
- Always set the `workdir` param when using the shell function. Do not use `cd` unless absolutely necessary.
- When searching for text or files, prefer using `rg` or `rg --files` respectively because `rg` is much faster than alternatives like `grep`. (If the `rg` command is not found, then use alternatives.)
- Read files in chunks with a max chunk size of 250 lines. Do not use python scripts to attempt to output larger chunks of a file. Command line output will be truncated after 10 kilobytes or 256 lines of output, regardless of the command used.
## apply_patch
Use the `apply_patch` tool to edit files. Your patch language is a stripped‑down, file‑oriented diff format designed to be easy to parse and safe to apply. You can think of it as a high‑level envelope:
*** Begin Patch
[ one or more file sections ]
*** End Patch
Within that envelope, you get a sequence of file operations.
You MUST include a header to specify the action you are taking.
Each operation starts with one of three headers:
*** Add File: <path> - create a new file. Every following line is a + line (the initial contents).
*** Delete File: <path> - remove an existing file. Nothing follows.
*** Update File: <path> - patch an existing file in place (optionally with a rename).
Example patch:
```
*** Begin Patch
*** Add File: hello.txt
+Hello world
*** Update File: src/app.py
*** Move to: src/main.py
@@ def greet():
-print("Hi")
+print("Hello, world!")
*** Delete File: obsolete.txt
*** End Patch
```
It is important to remember:
- You must include a header with your intended action (Add/Delete/Update)
- You must prefix new lines with `+` even when creating a new file
## `update_plan`
A tool named `update_plan` is available to you. You can use it to keep an up‑to‑date, step‑by‑step plan for the task.
codex-rs/core/src/bash.rs
+11 -13
View File
@@ -1,8 +1,13 @@
use std::path::PathBuf;
use tree_sitter::Node;
use tree_sitter::Parser;
use tree_sitter::Tree;
use tree_sitter_bash::LANGUAGE as BASH;
use crate::shell::ShellType;
use crate::shell::detect_shell_type;
/// Parse the provided bash source using tree-sitter-bash, returning a Tree on
/// success or None if parsing failed.
pub fn try_parse_shell(shell_lc_arg: &str) -> Option<Tree> {
@@ -88,23 +93,16 @@ pub fn try_parse_word_only_commands_sequence(tree: &Tree, src: &str) -> Option<V
Some(commands)
}
pub fn is_well_known_sh_shell(shell: &str) -> bool {
if shell == "bash" || shell == "zsh" {
return true;
}
let shell_name = std::path::Path::new(shell)
.file_name()
.and_then(|s| s.to_str())
.unwrap_or(shell);
matches!(shell_name, "bash" | "zsh")
}
pub fn extract_bash_command(command: &[String]) -> Option<(&str, &str)> {
let [shell, flag, script] = command else {
return None;
};
if !matches!(flag.as_str(), "-lc" | "-c") || !is_well_known_sh_shell(shell) {
if !matches!(flag.as_str(), "-lc" | "-c")
|| !matches!(
detect_shell_type(&PathBuf::from(shell)),
Some(ShellType::Zsh) | Some(ShellType::Bash)
)
{
return None;
}
Some((shell, script))
codex-rs/core/src/client/aggregation.rs
+10 -4
View File
@@ -175,16 +175,22 @@ where
return Poll::Ready(Some(Ok(ResponseEvent::OutputTextDelta(delta))));
}
}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningContentDelta(delta)))) => {
Poll::Ready(Some(Ok(ResponseEvent::ReasoningContentDelta {
delta,
content_index,
}))) => {
// Always accumulate reasoning deltas so we can emit a final Reasoning item at Completed.
this.cumulative_reasoning.push_str(&delta);
if matches!(this.mode, AggregateMode::Streaming) {
// In streaming mode, also forward the delta immediately.
return Poll::Ready(Some(Ok(ResponseEvent::ReasoningContentDelta(delta))));
return Poll::Ready(Some(Ok(ResponseEvent::ReasoningContentDelta {
delta,
content_index,
})));
}
}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningSummaryDelta(_)))) => {}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningSummaryPartAdded))) => {}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningSummaryDelta { .. }))) => {}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningSummaryPartAdded { .. }))) => {}
Poll::Ready(Some(Ok(ResponseEvent::OutputItemAdded(item)))) => {
return Poll::Ready(Some(Ok(ResponseEvent::OutputItemAdded(item))));
}
codex-rs/core/src/client/chat_completions.rs
+5 -1
View File
@@ -513,10 +513,14 @@ async fn append_reasoning_text(
..
}) = reasoning_item
{
let content_index = content.len() as i64;
content.push(ReasoningItemContent::ReasoningText { text: text.clone() });
let _ = tx_event
.send(Ok(ResponseEvent::ReasoningContentDelta(text.clone())))
.send(Ok(ResponseEvent::ReasoningContentDelta {
delta: text.clone(),
content_index,
}))
.await;
}
}
codex-rs/core/src/client/responses.rs
+17 -7
View File
@@ -668,6 +668,8 @@ struct SseEventWrapper {
item: Option<Value>,
response: Option<Value>,
delta: Option<String>,
summary_index: Option<i64>,
content_index: Option<i64>,
}
#[derive(Debug, Deserialize)]
@@ -786,16 +788,22 @@ async fn handle_sse_event(
}
}
"response.reasoning_summary_text.delta" => {
if let Some(delta) = event.delta {
let event = ResponseEvent::ReasoningSummaryDelta(delta);
if let (Some(delta), Some(summary_index)) = (event.delta, event.summary_index) {
let event = ResponseEvent::ReasoningSummaryDelta {
delta,
summary_index,
};
if tx_event.send(Ok(event)).await.is_err() {
return Ok(());
}
}
}
"response.reasoning_text.delta" => {
if let Some(delta) = event.delta {
let event = ResponseEvent::ReasoningContentDelta(delta);
if let (Some(delta), Some(content_index)) = (event.delta, event.content_index) {
let event = ResponseEvent::ReasoningContentDelta {
delta,
content_index,
};
if tx_event.send(Ok(event)).await.is_err() {
return Ok(());
}
@@ -853,9 +861,11 @@ async fn handle_sse_event(
}
"response.reasoning_summary_part.added" => {
// Boundary between reasoning summary sections (e.g., titles).
let event = ResponseEvent::ReasoningSummaryPartAdded;
if tx_event.send(Ok(event)).await.is_err() {
return Ok(());
if let Some(summary_index) = event.summary_index {
let event = ResponseEvent::ReasoningSummaryPartAdded { summary_index };
if tx_event.send(Ok(event)).await.is_err() {
return Ok(());
}
}
}
"response.content_part.done"
codex-rs/core/src/client/types.rs
+11 -3
View File
@@ -25,9 +25,17 @@ pub enum ResponseEvent {
token_usage: Option<TokenUsage>,
},
OutputTextDelta(String),
ReasoningSummaryDelta(String),
ReasoningContentDelta(String),
ReasoningSummaryPartAdded,
ReasoningSummaryDelta {
delta: String,
summary_index: i64,
},
ReasoningContentDelta {
delta: String,
content_index: i64,
},
ReasoningSummaryPartAdded {
summary_index: i64,
},
RateLimits(RateLimitSnapshot),
}
codex-rs/core/src/codex.rs
+54 -8
View File
@@ -96,6 +96,7 @@ use crate::protocol::StreamErrorEvent;
use crate::protocol::Submission;
use crate::protocol::TokenCountEvent;
use crate::protocol::TokenUsage;
use crate::protocol::TokenUsageInfo;
use crate::protocol::TurnDiffEvent;
use crate::protocol::WarningEvent;
use crate::rollout::RolloutRecorder;
@@ -1082,6 +1083,36 @@ impl Session {
self.send_token_count_event(turn_context).await;
}
pub(crate) async fn override_last_token_usage_estimate(
&self,
turn_context: &TurnContext,
estimated_total_tokens: i64,
) {
{
let mut state = self.state.lock().await;
let mut info = state.token_info().unwrap_or(TokenUsageInfo {
total_token_usage: TokenUsage::default(),
last_token_usage: TokenUsage::default(),
model_context_window: None,
});
info.last_token_usage = TokenUsage {
input_tokens: 0,
cached_input_tokens: 0,
output_tokens: 0,
reasoning_output_tokens: 0,
total_tokens: estimated_total_tokens.max(0),
};
if info.model_context_window.is_none() {
info.model_context_window = turn_context.client.get_model_context_window();
}
state.set_token_info(Some(info));
}
self.send_token_count_event(turn_context).await;
}
pub(crate) async fn update_rate_limits(
&self,
turn_context: &TurnContext,
@@ -2201,13 +2232,17 @@ async fn try_run_turn(
error_or_panic("ReasoningSummaryDelta without active item".to_string());
}
}
ResponseEvent::ReasoningSummaryDelta(delta) => {
ResponseEvent::ReasoningSummaryDelta {
delta,
summary_index,
} => {
if let Some(active) = active_item.as_ref() {
let event = ReasoningContentDeltaEvent {
thread_id: sess.conversation_id.to_string(),
turn_id: turn_context.sub_id.clone(),
item_id: active.id(),
delta: delta.clone(),
delta,
summary_index,
};
sess.send_event(&turn_context, EventMsg::ReasoningContentDelta(event))
.await;
@@ -2215,18 +2250,29 @@ async fn try_run_turn(
error_or_panic("ReasoningSummaryDelta without active item".to_string());
}
}
ResponseEvent::ReasoningSummaryPartAdded => {
let event =
EventMsg::AgentReasoningSectionBreak(AgentReasoningSectionBreakEvent {});
sess.send_event(&turn_context, event).await;
ResponseEvent::ReasoningSummaryPartAdded { summary_index } => {
if let Some(active) = active_item.as_ref() {
let event =
EventMsg::AgentReasoningSectionBreak(AgentReasoningSectionBreakEvent {
item_id: active.id(),
summary_index,
});
sess.send_event(&turn_context, event).await;
} else {
error_or_panic("ReasoningSummaryPartAdded without active item".to_string());
}
}
ResponseEvent::ReasoningContentDelta(delta) => {
ResponseEvent::ReasoningContentDelta {
delta,
content_index,
} => {
if let Some(active) = active_item.as_ref() {
let event = ReasoningRawContentDeltaEvent {
thread_id: sess.conversation_id.to_string(),
turn_id: turn_context.sub_id.clone(),
item_id: active.id(),
delta: delta.clone(),
delta,
content_index,
};
sess.send_event(&turn_context, EventMsg::ReasoningRawContentDelta(event))
.await;
codex-rs/core/src/command_safety/is_safe_command.rs
+4 -6
View File
@@ -1,4 +1,5 @@
use crate::bash::parse_shell_lc_plain_commands;
use crate::command_safety::windows_safe_commands::is_safe_command_windows;
pub fn is_known_safe_command(command: &[String]) -> bool {
let command: Vec<String> = command
@@ -11,12 +12,9 @@ pub fn is_known_safe_command(command: &[String]) -> bool {
}
})
.collect();
#[cfg(target_os = "windows")]
{
use super::windows_safe_commands::is_safe_command_windows;
if is_safe_command_windows(&command) {
return true;
}
if is_safe_command_windows(&command) {
return true;
}
if is_safe_to_call_with_exec(&command) {
codex-rs/core/src/command_safety/mod.rs
-1
View File
@@ -1,4 +1,3 @@
pub mod is_dangerous_command;
pub mod is_safe_command;
#[cfg(target_os = "windows")]
pub mod windows_safe_commands;
codex-rs/core/src/compact.rs
+9
View File
@@ -153,6 +153,15 @@ async fn run_compact_task_inner(
new_history.extend(ghost_snapshots);
sess.replace_history(new_history).await;
if let Some(estimated_tokens) = sess
.clone_history()
.await
.estimate_token_count(&turn_context)
{
sess.override_last_token_usage_estimate(&turn_context, estimated_tokens)
.await;
}
let rollout_item = RolloutItem::Compacted(CompactedItem {
message: summary_text.clone(),
});
codex-rs/core/src/context_manager/history.rs
+47 -5
View File
@@ -1,12 +1,20 @@
use crate::codex::TurnContext;
use crate::context_manager::normalize;
use crate::context_manager::truncate;
use crate::context_manager::truncate::format_output_for_model_body;
use crate::context_manager::truncate::globally_truncate_function_output_items;
use codex_protocol::models::FunctionCallOutputPayload;
use codex_protocol::models::ResponseItem;
use codex_protocol::protocol::TokenUsage;
use codex_protocol::protocol::TokenUsageInfo;
use codex_utils_tokenizer::Tokenizer;
use std::ops::Deref;
use crate::context_manager::normalize;
use crate::context_manager::truncate::format_output_for_model_body;
use crate::context_manager::truncate::globally_truncate_function_output_items;
const CONTEXT_WINDOW_HARD_LIMIT_FACTOR: f64 = 1.1;
const CONTEXT_WINDOW_HARD_LIMIT_BYTES: usize =
(truncate::MODEL_FORMAT_MAX_BYTES as f64 * CONTEXT_WINDOW_HARD_LIMIT_FACTOR) as usize;
const CONTEXT_WINDOW_HARD_LIMIT_LINES: usize =
(truncate::MODEL_FORMAT_MAX_LINES as f64 * CONTEXT_WINDOW_HARD_LIMIT_FACTOR) as usize;
/// Transcript of conversation history
#[derive(Debug, Clone, Default)]
@@ -28,6 +36,10 @@ impl ContextManager {
self.token_info.clone()
}
pub(crate) fn set_token_info(&mut self, info: Option<TokenUsageInfo>) {
self.token_info = info;
}
pub(crate) fn set_token_usage_full(&mut self, context_window: i64) {
match &mut self.token_info {
Some(info) => info.fill_to_context_window(context_window),
@@ -68,6 +80,28 @@ impl ContextManager {
history
}
// Estimate the number of tokens in the history. Return None if no tokenizer
// is available. This does not consider the reasoning traces.
// /!\ The value is a lower bound estimate and does not represent the exact
// context length.
pub(crate) fn estimate_token_count(&self, turn_context: &TurnContext) -> Option<i64> {
let model = turn_context.client.get_model();
let tokenizer = Tokenizer::for_model(model.as_str()).ok()?;
let model_family = turn_context.client.get_model_family();
Some(
self.items
.iter()
.map(|item| {
serde_json::to_string(&item)
.map(|item| tokenizer.count(&item))
.unwrap_or_default()
})
.sum::<i64>()
+ tokenizer.count(model_family.base_instructions.as_str()),
)
}
pub(crate) fn remove_first_item(&mut self) {
if !self.items.is_empty() {
// Remove the oldest item (front of the list). Items are ordered from
@@ -119,7 +153,11 @@ impl ContextManager {
fn process_item(item: &ResponseItem) -> ResponseItem {
match item {
ResponseItem::FunctionCallOutput { call_id, output } => {
let truncated = format_output_for_model_body(output.content.as_str());
let truncated = format_output_for_model_body(
output.content.as_str(),
CONTEXT_WINDOW_HARD_LIMIT_BYTES,
CONTEXT_WINDOW_HARD_LIMIT_LINES,
);
let truncated_items = output
.content_items
.as_ref()
@@ -134,7 +172,11 @@ impl ContextManager {
}
}
ResponseItem::CustomToolCallOutput { call_id, output } => {
let truncated = format_output_for_model_body(output);
let truncated = format_output_for_model_body(
output,
CONTEXT_WINDOW_HARD_LIMIT_BYTES,
CONTEXT_WINDOW_HARD_LIMIT_LINES,
);
ResponseItem::CustomToolCallOutput {
call_id: call_id.clone(),
output: truncated,
codex-rs/core/src/context_manager/history_tests.rs
+33 -7
View File
@@ -1,4 +1,5 @@
use super::*;
use crate::context_manager::MODEL_FORMAT_MAX_LINES;
use crate::context_manager::truncate;
use codex_git::GhostCommit;
use codex_protocol::models::ContentItem;
@@ -308,8 +309,10 @@ fn assert_truncated_message_matches(message: &str, line: &str, total_lines: usiz
}
fn truncated_message_pattern(line: &str, total_lines: usize) -> String {
let head_take = truncate::MODEL_FORMAT_HEAD_LINES.min(total_lines);
let tail_take = truncate::MODEL_FORMAT_TAIL_LINES.min(total_lines.saturating_sub(head_take));
let head_lines = MODEL_FORMAT_MAX_LINES / 2;
let tail_lines = MODEL_FORMAT_MAX_LINES - head_lines;
let head_take = head_lines.min(total_lines);
let tail_take = tail_lines.min(total_lines.saturating_sub(head_take));
let omitted = total_lines.saturating_sub(head_take + tail_take);
let escaped_line = regex_lite::escape(line);
if omitted == 0 {
@@ -328,7 +331,11 @@ fn format_exec_output_truncates_large_error() {
let line = "very long execution error line that should trigger truncation\n";
let large_error = line.repeat(2_500); // way beyond both byte and line limits
let truncated = truncate::format_output_for_model_body(&large_error);
let truncated = truncate::format_output_for_model_body(
&large_error,
truncate::MODEL_FORMAT_MAX_BYTES,
truncate::MODEL_FORMAT_MAX_LINES,
);
let total_lines = large_error.lines().count();
assert_truncated_message_matches(&truncated, line, total_lines);
@@ -338,7 +345,11 @@ fn format_exec_output_truncates_large_error() {
#[test]
fn format_exec_output_marks_byte_truncation_without_omitted_lines() {
let long_line = "a".repeat(truncate::MODEL_FORMAT_MAX_BYTES + 50);
let truncated = truncate::format_output_for_model_body(&long_line);
let truncated = truncate::format_output_for_model_body(
&long_line,
truncate::MODEL_FORMAT_MAX_BYTES,
truncate::MODEL_FORMAT_MAX_LINES,
);
assert_ne!(truncated, long_line);
let marker_line = format!(
@@ -359,7 +370,14 @@ fn format_exec_output_marks_byte_truncation_without_omitted_lines() {
fn format_exec_output_returns_original_when_within_limits() {
let content = "example output\n".repeat(10);
assert_eq!(truncate::format_output_for_model_body(&content), content);
assert_eq!(
truncate::format_output_for_model_body(
&content,
truncate::MODEL_FORMAT_MAX_BYTES,
truncate::MODEL_FORMAT_MAX_LINES
),
content
);
}
#[test]
@@ -369,7 +387,11 @@ fn format_exec_output_reports_omitted_lines_and_keeps_head_and_tail() {
.map(|idx| format!("line-{idx}\n"))
.collect();
let truncated = truncate::format_output_for_model_body(&content);
let truncated = truncate::format_output_for_model_body(
&content,
truncate::MODEL_FORMAT_MAX_BYTES,
truncate::MODEL_FORMAT_MAX_LINES,
);
let omitted = total_lines - truncate::MODEL_FORMAT_MAX_LINES;
let expected_marker = format!("[... omitted {omitted} of {total_lines} lines ...]");
@@ -397,7 +419,11 @@ fn format_exec_output_prefers_line_marker_when_both_limits_exceeded() {
.map(|idx| format!("line-{idx}-{long_line}\n"))
.collect();
let truncated = truncate::format_output_for_model_body(&content);
let truncated = truncate::format_output_for_model_body(
&content,
truncate::MODEL_FORMAT_MAX_BYTES,
truncate::MODEL_FORMAT_MAX_LINES,
);
assert!(
truncated.contains("[... omitted 42 of 298 lines ...]"),
codex-rs/core/src/context_manager/mod.rs
+2
View File
@@ -3,4 +3,6 @@ mod normalize;
mod truncate;
pub(crate) use history::ContextManager;
pub(crate) use truncate::MODEL_FORMAT_MAX_BYTES;
pub(crate) use truncate::MODEL_FORMAT_MAX_LINES;
pub(crate) use truncate::format_output_for_model_body;
codex-rs/core/src/context_manager/truncate.rs
+37 -17
View File
@@ -2,12 +2,11 @@ use codex_protocol::models::FunctionCallOutputContentItem;
use codex_utils_string::take_bytes_at_char_boundary;
use codex_utils_string::take_last_bytes_at_char_boundary;
use crate::util::error_or_panic;
// Model-formatting limits: clients get full streams; only content sent to the model is truncated.
pub(crate) const MODEL_FORMAT_MAX_BYTES: usize = 10 * 1024; // 10 KiB
pub(crate) const MODEL_FORMAT_MAX_LINES: usize = 256; // lines
pub(crate) const MODEL_FORMAT_HEAD_LINES: usize = MODEL_FORMAT_MAX_LINES / 2;
pub(crate) const MODEL_FORMAT_TAIL_LINES: usize = MODEL_FORMAT_MAX_LINES - MODEL_FORMAT_HEAD_LINES; // 128
pub(crate) const MODEL_FORMAT_HEAD_BYTES: usize = MODEL_FORMAT_MAX_BYTES / 2;
pub const MODEL_FORMAT_MAX_BYTES: usize = 10 * 1024; // 10 KiB
pub const MODEL_FORMAT_MAX_LINES: usize = 256; // lines
pub(crate) fn globally_truncate_function_output_items(
items: &[FunctionCallOutputContentItem],
@@ -56,21 +55,34 @@ pub(crate) fn globally_truncate_function_output_items(
out
}
pub(crate) fn format_output_for_model_body(content: &str) -> String {
pub(crate) fn format_output_for_model_body(
content: &str,
limit_bytes: usize,
limit_lines: usize,
) -> String {
// Head+tail truncation for the model: show the beginning and end with an elision.
// Clients still receive full streams; only this formatted summary is capped.
let total_lines = content.lines().count();
if content.len() <= MODEL_FORMAT_MAX_BYTES && total_lines <= MODEL_FORMAT_MAX_LINES {
if content.len() <= limit_bytes && total_lines <= limit_lines {
return content.to_string();
}
let output = truncate_formatted_exec_output(content, total_lines);
let output = truncate_formatted_exec_output(content, total_lines, limit_bytes, limit_lines);
format!("Total output lines: {total_lines}\n\n{output}")
}
fn truncate_formatted_exec_output(content: &str, total_lines: usize) -> String {
fn truncate_formatted_exec_output(
content: &str,
total_lines: usize,
limit_bytes: usize,
limit_lines: usize,
) -> String {
debug_panic_on_double_truncation(content);
let head_lines: usize = limit_lines / 2;
let tail_lines: usize = limit_lines - head_lines; // 128
let head_bytes: usize = limit_bytes / 2;
let segments: Vec<&str> = content.split_inclusive('\n').collect();
let head_take = MODEL_FORMAT_HEAD_LINES.min(segments.len());
let tail_take = MODEL_FORMAT_TAIL_LINES.min(segments.len().saturating_sub(head_take));
let head_take = head_lines.min(segments.len());
let tail_take = tail_lines.min(segments.len().saturating_sub(head_take));
let omitted = segments.len().saturating_sub(head_take + tail_take);
let head_slice_end: usize = segments
@@ -91,7 +103,7 @@ fn truncate_formatted_exec_output(content: &str, total_lines: usize) -> String {
};
let head_slice = &content[..head_slice_end];
let tail_slice = &content[tail_slice_start..];
let truncated_by_bytes = content.len() > MODEL_FORMAT_MAX_BYTES;
let truncated_by_bytes = content.len() > limit_bytes;
// this is a bit wrong. We are counting metadata lines and not just shell output lines.
let marker = if omitted > 0 {
Some(format!(
@@ -99,24 +111,24 @@ fn truncate_formatted_exec_output(content: &str, total_lines: usize) -> String {
))
} else if truncated_by_bytes {
Some(format!(
"\n[... output truncated to fit {MODEL_FORMAT_MAX_BYTES} bytes ...]\n\n"
"\n[... output truncated to fit {limit_bytes} bytes ...]\n\n"
))
} else {
None
};
let marker_len = marker.as_ref().map_or(0, String::len);
let base_head_budget = MODEL_FORMAT_HEAD_BYTES.min(MODEL_FORMAT_MAX_BYTES);
let head_budget = base_head_budget.min(MODEL_FORMAT_MAX_BYTES.saturating_sub(marker_len));
let base_head_budget = head_bytes.min(limit_bytes);
let head_budget = base_head_budget.min(limit_bytes.saturating_sub(marker_len));
let head_part = take_bytes_at_char_boundary(head_slice, head_budget);
let mut result = String::with_capacity(MODEL_FORMAT_MAX_BYTES.min(content.len()));
let mut result = String::with_capacity(limit_bytes.min(content.len()));
result.push_str(head_part);
if let Some(marker_text) = marker.as_ref() {
result.push_str(marker_text);
}
let remaining = MODEL_FORMAT_MAX_BYTES.saturating_sub(result.len());
let remaining = limit_bytes.saturating_sub(result.len());
if remaining == 0 {
return result;
}
@@ -126,3 +138,11 @@ fn truncate_formatted_exec_output(content: &str, total_lines: usize) -> String {
result
}
fn debug_panic_on_double_truncation(content: &str) {
if content.contains("Total output lines:") && content.contains("omitted") {
error_or_panic(format!(
"FunctionCallOutput content was already truncated before ContextManager::record_items; this would cause double truncation {content}"
));
}
}
codex-rs/core/src/environment_context.rs
-2
View File
@@ -329,7 +329,6 @@ mod tests {
Some(workspace_write_policy(vec!["/repo"], false)),
Some(Shell::Bash(BashShell {
shell_path: "/bin/bash".into(),
bashrc_path: "/home/user/.bashrc".into(),
})),
);
let context2 = EnvironmentContext::new(
@@ -338,7 +337,6 @@ mod tests {
Some(workspace_write_policy(vec!["/repo"], false)),
Some(Shell::Zsh(ZshShell {
shell_path: "/bin/zsh".into(),
zshrc_path: "/home/user/.zshrc".into(),
})),
);
codex-rs/core/src/shell.rs
+231 -338
View File
@@ -2,22 +2,26 @@ use serde::Deserialize;
use serde::Serialize;
use std::path::PathBuf;
#[derive(Debug, PartialEq, Eq, Clone, Serialize, Deserialize)]
pub enum ShellType {
Zsh,
Bash,
PowerShell,
}
#[derive(Debug, PartialEq, Eq, Clone, Serialize, Deserialize)]
pub struct ZshShell {
pub(crate) shell_path: String,
pub(crate) zshrc_path: String,
pub(crate) shell_path: PathBuf,
}
#[derive(Debug, PartialEq, Eq, Clone, Serialize, Deserialize)]
pub struct BashShell {
pub(crate) shell_path: String,
pub(crate) bashrc_path: String,
pub(crate) shell_path: PathBuf,
}
#[derive(Debug, PartialEq, Eq, Clone, Serialize, Deserialize)]
pub struct PowerShellConfig {
pub(crate) exe: String, // Executable name or path, e.g. "pwsh" or "powershell.exe".
pub(crate) bash_exe_fallback: Option<PathBuf>, // In case the model generates a bash command.
pub(crate) shell_path: PathBuf, // Executable name or path, e.g. "pwsh" or "powershell.exe".
}
#[derive(Debug, PartialEq, Eq, Clone, Serialize, Deserialize)]
@@ -36,7 +40,10 @@ impl Shell {
.file_name()
.map(|s| s.to_string_lossy().to_string())
}
Shell::PowerShell(ps) => Some(ps.exe.clone()),
Shell::PowerShell(ps) => ps
.shell_path
.file_stem()
.map(|s| s.to_string_lossy().to_string()),
Shell::Unknown => None,
}
}
@@ -47,10 +54,17 @@ impl Shell {
match self {
Shell::Zsh(ZshShell { shell_path, .. }) | Shell::Bash(BashShell { shell_path, .. }) => {
let arg = if use_login_shell { "-lc" } else { "-c" };
vec![shell_path.clone(), arg.to_string(), command.to_string()]
vec![
shell_path.to_string_lossy().to_string(),
arg.to_string(),
command.to_string(),
]
}
Shell::PowerShell(ps) => {
let mut args = vec![ps.exe.clone(), "-NoLogo".to_string()];
let mut args = vec![
ps.shell_path.to_string_lossy().to_string(),
"-NoLogo".to_string(),
];
if !use_login_shell {
args.push("-NoProfile".to_string());
}
@@ -65,7 +79,7 @@ impl Shell {
}
#[cfg(unix)]
fn detect_default_user_shell() -> Shell {
fn get_user_shell_path() -> Option<PathBuf> {
use libc::getpwuid;
use libc::getuid;
use std::ffi::CStr;
@@ -78,75 +92,174 @@ fn detect_default_user_shell() -> Shell {
let shell_path = CStr::from_ptr((*pw).pw_shell)
.to_string_lossy()
.into_owned();
let home_path = CStr::from_ptr((*pw).pw_dir).to_string_lossy().into_owned();
Some(PathBuf::from(shell_path))
} else {
None
}
}
}
if shell_path.ends_with("/zsh") {
return Shell::Zsh(ZshShell {
shell_path,
zshrc_path: format!("{home_path}/.zshrc"),
});
}
#[cfg(not(unix))]
fn get_user_shell_path() -> Option<PathBuf> {
None
}
if shell_path.ends_with("/bash") {
return Shell::Bash(BashShell {
shell_path,
bashrc_path: format!("{home_path}/.bashrc"),
});
fn file_exists(path: &PathBuf) -> Option<PathBuf> {
if std::fs::metadata(path).is_ok_and(|metadata| metadata.is_file()) {
Some(PathBuf::from(path))
} else {
None
}
}
fn get_shell_path(
shell_type: ShellType,
provided_path: Option<&PathBuf>,
binary_name: &str,
fallback_paths: Vec<&str>,
) -> Option<PathBuf> {
// If exact provided path exists, use it
if provided_path.and_then(file_exists).is_some() {
return provided_path.cloned();
}
// Check if the shell we are trying to load is user's default shell
// if just use it
let default_shell_path = get_user_shell_path();
if let Some(default_shell_path) = default_shell_path
&& detect_shell_type(&default_shell_path) == Some(shell_type)
{
return Some(default_shell_path);
}
if let Ok(path) = which::which(binary_name) {
return Some(path);
}
for path in fallback_paths {
//check exists
if let Some(path) = file_exists(&PathBuf::from(path)) {
return Some(path);
}
}
None
}
fn get_zsh_shell(path: Option<&PathBuf>) -> Option<ZshShell> {
let shell_path = get_shell_path(ShellType::Zsh, path, "zsh", vec!["/bin/zsh"]);
shell_path.map(|shell_path| ZshShell { shell_path })
}
fn get_bash_shell(path: Option<&PathBuf>) -> Option<BashShell> {
let shell_path = get_shell_path(ShellType::Bash, path, "bash", vec!["/bin/bash"]);
shell_path.map(|shell_path| BashShell { shell_path })
}
fn get_powershell_shell(path: Option<&PathBuf>) -> Option<PowerShellConfig> {
let shell_path = get_shell_path(
ShellType::PowerShell,
path,
"pwsh",
vec!["/usr/local/bin/pwsh"],
)
.or_else(|| get_shell_path(ShellType::PowerShell, path, "powershell", vec![]));
shell_path.map(|shell_path| PowerShellConfig { shell_path })
}
pub fn get_shell_by_model_provided_path(shell_path: &PathBuf) -> Shell {
detect_shell_type(shell_path)
.and_then(|shell_type| get_shell(shell_type, Some(shell_path)))
.unwrap_or(Shell::Unknown)
}
pub fn get_shell(shell_type: ShellType, path: Option<&PathBuf>) -> Option<Shell> {
match shell_type {
ShellType::Zsh => get_zsh_shell(path).map(Shell::Zsh),
ShellType::Bash => get_bash_shell(path).map(Shell::Bash),
ShellType::PowerShell => get_powershell_shell(path).map(Shell::PowerShell),
}
}
pub fn detect_shell_type(shell_path: &PathBuf) -> Option<ShellType> {
match shell_path.as_os_str().to_str() {
Some("zsh") => Some(ShellType::Zsh),
Some("bash") => Some(ShellType::Bash),
Some("pwsh") => Some(ShellType::PowerShell),
Some("powershell") => Some(ShellType::PowerShell),
_ => {
let shell_name = shell_path.file_stem();
if let Some(shell_name) = shell_name
&& shell_name != shell_path
{
detect_shell_type(&PathBuf::from(shell_name))
} else {
None
}
}
}
Shell::Unknown
}
#[cfg(unix)]
pub async fn default_user_shell() -> Shell {
detect_default_user_shell()
}
#[cfg(target_os = "windows")]
pub async fn default_user_shell() -> Shell {
use tokio::process::Command;
// Prefer PowerShell 7+ (`pwsh`) if available, otherwise fall back to Windows PowerShell.
let has_pwsh = Command::new("pwsh")
.arg("-NoLogo")
.arg("-NoProfile")
.arg("-Command")
.arg("$PSVersionTable.PSVersion.Major")
.output()
.await
.map(|o| o.status.success())
.unwrap_or(false);
let bash_exe = if Command::new("bash.exe")
.arg("--version")
.stdin(std::process::Stdio::null())
.output()
.await
.ok()
.map(|o| o.status.success())
.unwrap_or(false)
{
which::which("bash.exe").ok()
if cfg!(windows) {
get_shell(ShellType::PowerShell, None).unwrap_or(Shell::Unknown)
} else {
None
};
if has_pwsh {
Shell::PowerShell(PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: bash_exe,
})
} else {
Shell::PowerShell(PowerShellConfig {
exe: "powershell.exe".to_string(),
bash_exe_fallback: bash_exe,
})
get_user_shell_path()
.and_then(|shell| detect_shell_type(&shell))
.and_then(|shell_type| get_shell(shell_type, None))
.unwrap_or(Shell::Unknown)
}
}
#[cfg(all(not(target_os = "windows"), not(unix)))]
pub async fn default_user_shell() -> Shell {
Shell::Unknown
#[cfg(test)]
mod detect_shell_type_tests {
use super::*;
#[test]
fn test_detect_shell_type() {
assert_eq!(
detect_shell_type(&PathBuf::from("zsh")),
Some(ShellType::Zsh)
);
assert_eq!(
detect_shell_type(&PathBuf::from("bash")),
Some(ShellType::Bash)
);
assert_eq!(
detect_shell_type(&PathBuf::from("pwsh")),
Some(ShellType::PowerShell)
);
assert_eq!(
detect_shell_type(&PathBuf::from("powershell")),
Some(ShellType::PowerShell)
);
assert_eq!(detect_shell_type(&PathBuf::from("fish")), None);
assert_eq!(detect_shell_type(&PathBuf::from("other")), None);
assert_eq!(
detect_shell_type(&PathBuf::from("/bin/zsh")),
Some(ShellType::Zsh)
);
assert_eq!(
detect_shell_type(&PathBuf::from("/bin/bash")),
Some(ShellType::Bash)
);
assert_eq!(
detect_shell_type(&PathBuf::from("powershell.exe")),
Some(ShellType::PowerShell)
);
assert_eq!(
detect_shell_type(&PathBuf::from("pwsh.exe")),
Some(ShellType::PowerShell)
);
assert_eq!(
detect_shell_type(&PathBuf::from("/usr/local/bin/pwsh")),
Some(ShellType::PowerShell)
);
}
}
#[cfg(test)]
@@ -156,6 +269,34 @@ mod tests {
use std::path::PathBuf;
use std::process::Command;
#[test]
#[cfg(target_os = "macos")]
fn detects_zsh() {
let zsh_shell = get_shell(ShellType::Zsh, None).unwrap();
let ZshShell { shell_path } = match zsh_shell {
Shell::Zsh(zsh_shell) => zsh_shell,
_ => panic!("expected zsh shell"),
};
assert_eq!(shell_path, PathBuf::from("/bin/zsh"));
}
#[test]
fn detects_bash() {
let bash_shell = get_shell(ShellType::Bash, None).unwrap();
let BashShell { shell_path } = match bash_shell {
Shell::Bash(bash_shell) => bash_shell,
_ => panic!("expected bash shell"),
};
assert!(
shell_path == PathBuf::from("/bin/bash")
|| shell_path == PathBuf::from("/usr/bin/bash"),
"shell path: {shell_path:?}",
);
}
#[tokio::test]
async fn test_current_shell_detects_zsh() {
let shell = Command::new("sh")
@@ -164,292 +305,44 @@ mod tests {
.output()
.unwrap();
let home = std::env::var("HOME").unwrap();
let shell_path = String::from_utf8_lossy(&shell.stdout).trim().to_string();
if shell_path.ends_with("/zsh") {
assert_eq!(
default_user_shell().await,
Shell::Zsh(ZshShell {
shell_path: shell_path.to_string(),
zshrc_path: format!("{home}/.zshrc",),
shell_path: PathBuf::from(shell_path),
})
);
}
}
#[tokio::test]
async fn test_run_with_profile_bash_escaping_and_execution() {
let shell_path = "/bin/bash";
let cases = vec![
(
vec!["myecho"],
vec![shell_path, "-lc", "source BASHRC_PATH && (myecho)"],
Some("It works!\n"),
),
(
vec!["bash", "-lc", "echo 'single' \"double\""],
vec![
shell_path,
"-lc",
"source BASHRC_PATH && (echo 'single' \"double\")",
],
Some("single double\n"),
),
];
for (input, expected_cmd, expected_output) in cases {
use std::collections::HashMap;
use crate::exec::ExecParams;
use crate::exec::SandboxType;
use crate::exec::process_exec_tool_call;
use crate::protocol::SandboxPolicy;
let temp_home = tempfile::tempdir().unwrap();
let bashrc_path = temp_home.path().join(".bashrc");
std::fs::write(
&bashrc_path,
r#"
set -x
function myecho {
echo 'It works!'
}
"#,
)
.unwrap();
let command = expected_cmd
.iter()
.map(|s| s.replace("BASHRC_PATH", bashrc_path.to_str().unwrap()))
.collect::<Vec<_>>();
let output = process_exec_tool_call(
ExecParams {
command: command.clone(),
cwd: PathBuf::from(temp_home.path()),
timeout_ms: None,
env: HashMap::from([(
"HOME".to_string(),
temp_home.path().to_str().unwrap().to_string(),
)]),
with_escalated_permissions: None,
justification: None,
arg0: None,
},
SandboxType::None,
&SandboxPolicy::DangerFullAccess,
temp_home.path(),
&None,
None,
)
.await
.unwrap();
assert_eq!(output.exit_code, 0, "input: {input:?} output: {output:?}");
if let Some(expected) = expected_output {
assert_eq!(
output.stdout.text, expected,
"input: {input:?} output: {output:?}"
);
}
async fn detects_powershell_as_default() {
if !cfg!(windows) {
return;
}
let powershell_shell = default_user_shell().await;
let PowerShellConfig { shell_path } = match powershell_shell {
Shell::PowerShell(powershell_shell) => powershell_shell,
_ => panic!("expected powershell shell"),
};
assert!(shell_path.ends_with("pwsh.exe") || shell_path.ends_with("powershell.exe"));
}
}
#[cfg(test)]
#[cfg(target_os = "macos")]
mod macos_tests {
use std::path::PathBuf;
#[tokio::test]
async fn test_run_with_profile_escaping_and_execution() {
let shell_path = "/bin/zsh";
let cases = vec![
(
vec!["myecho"],
vec![shell_path, "-lc", "source ZSHRC_PATH && (myecho)"],
Some("It works!\n"),
),
(
vec!["myecho"],
vec![shell_path, "-lc", "source ZSHRC_PATH && (myecho)"],
Some("It works!\n"),
),
(
vec!["bash", "-c", "echo 'single' \"double\""],
vec![
shell_path,
"-lc",
"source ZSHRC_PATH && (bash -c \"echo 'single' \\\"double\\\"\")",
],
Some("single double\n"),
),
(
vec!["bash", "-lc", "echo 'single' \"double\""],
vec![
shell_path,
"-lc",
"source ZSHRC_PATH && (echo 'single' \"double\")",
],
Some("single double\n"),
),
];
for (input, expected_cmd, expected_output) in cases {
use std::collections::HashMap;
use crate::exec::ExecParams;
use crate::exec::SandboxType;
use crate::exec::process_exec_tool_call;
use crate::protocol::SandboxPolicy;
let temp_home = tempfile::tempdir().unwrap();
let zshrc_path = temp_home.path().join(".zshrc");
std::fs::write(
&zshrc_path,
r#"
set -x
function myecho {
echo 'It works!'
}
"#,
)
.unwrap();
let command = expected_cmd
.iter()
.map(|s| s.replace("ZSHRC_PATH", zshrc_path.to_str().unwrap()))
.collect::<Vec<_>>();
let output = process_exec_tool_call(
ExecParams {
command: command.clone(),
cwd: PathBuf::from(temp_home.path()),
timeout_ms: None,
env: HashMap::from([(
"HOME".to_string(),
temp_home.path().to_str().unwrap().to_string(),
)]),
with_escalated_permissions: None,
justification: None,
arg0: None,
},
SandboxType::None,
&SandboxPolicy::DangerFullAccess,
temp_home.path(),
&None,
None,
)
.await
.unwrap();
assert_eq!(output.exit_code, 0, "input: {input:?} output: {output:?}");
if let Some(expected) = expected_output {
assert_eq!(
output.stdout.text, expected,
"input: {input:?} output: {output:?}"
);
}
}
}
}
#[cfg(test)]
#[cfg(target_os = "windows")]
mod tests_windows {
use super::*;
#[test]
fn test_format_default_shell_invocation_powershell() {
use std::path::PathBuf;
let cases = vec![
(
PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: None,
},
vec!["bash", "-lc", "echo hello"],
vec!["pwsh.exe", "-NoProfile", "-Command", "echo hello"],
),
(
PowerShellConfig {
exe: "powershell.exe".to_string(),
bash_exe_fallback: None,
},
vec!["bash", "-lc", "echo hello"],
vec!["powershell.exe", "-NoProfile", "-Command", "echo hello"],
),
(
PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: Some(PathBuf::from("bash.exe")),
},
vec!["bash", "-lc", "echo hello"],
vec!["bash.exe", "-lc", "echo hello"],
),
(
PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: Some(PathBuf::from("bash.exe")),
},
vec![
"bash",
"-lc",
"apply_patch <<'EOF'\n*** Begin Patch\n*** Update File: destination_file.txt\n-original content\n+modified content\n*** End Patch\nEOF",
],
vec![
"bash.exe",
"-lc",
"apply_patch <<'EOF'\n*** Begin Patch\n*** Update File: destination_file.txt\n-original content\n+modified content\n*** End Patch\nEOF",
],
),
(
PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: Some(PathBuf::from("bash.exe")),
},
vec!["echo", "hello"],
vec!["pwsh.exe", "-NoProfile", "-Command", "echo hello"],
),
(
PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: Some(PathBuf::from("bash.exe")),
},
vec!["pwsh.exe", "-NoProfile", "-Command", "echo hello"],
vec!["pwsh.exe", "-NoProfile", "-Command", "echo hello"],
),
(
PowerShellConfig {
exe: "powershell.exe".to_string(),
bash_exe_fallback: Some(PathBuf::from("bash.exe")),
},
vec![
"codex-mcp-server.exe",
"--codex-run-as-apply-patch",
"*** Begin Patch\n*** Update File: C:\\Users\\person\\destination_file.txt\n-original content\n+modified content\n*** End Patch",
],
vec![
"codex-mcp-server.exe",
"--codex-run-as-apply-patch",
"*** Begin Patch\n*** Update File: C:\\Users\\person\\destination_file.txt\n-original content\n+modified content\n*** End Patch",
],
),
];
for (config, input, expected_cmd) in cases {
let command = expected_cmd
.iter()
.map(|s| (*s).to_string())
.collect::<Vec<_>>();
// These tests assert the final command for each scenario now that the helper
// has been removed. The inputs remain to document the original coverage.
let expected = expected_cmd
.iter()
.map(|s| (*s).to_string())
.collect::<Vec<_>>();
assert_eq!(command, expected, "input: {input:?} config: {config:?}");
fn finds_poweshell() {
if !cfg!(windows) {
return;
}
let powershell_shell = get_shell(ShellType::PowerShell, None).unwrap();
let PowerShellConfig { shell_path } = match powershell_shell {
Shell::PowerShell(powershell_shell) => powershell_shell,
_ => panic!("expected powershell shell"),
};
assert!(shell_path.ends_with("pwsh.exe") || shell_path.ends_with("powershell.exe"));
}
}
codex-rs/core/src/state/session.rs
+4
View File
@@ -42,6 +42,10 @@ impl SessionState {
self.history.replace(items);
}
pub(crate) fn set_token_info(&mut self, info: Option<TokenUsageInfo>) {
self.history.set_token_info(info);
}
// Token/rate limit helpers
pub(crate) fn update_token_info_from_usage(
&mut self,
codex-rs/core/src/tools/events.rs
+4 -4
View File
@@ -87,7 +87,7 @@ pub(crate) enum ToolEmitter {
auto_approved: bool,
},
UnifiedExec {
command: String,
command: Vec<String>,
cwd: PathBuf,
// True for `exec_command` and false for `write_stdin`.
#[allow(dead_code)]
@@ -111,9 +111,9 @@ impl ToolEmitter {
}
}
pub fn unified_exec(command: String, cwd: PathBuf, is_startup_command: bool) -> Self {
pub fn unified_exec(command: &[String], cwd: PathBuf, is_startup_command: bool) -> Self {
Self::UnifiedExec {
command,
command: command.to_vec(),
cwd,
is_startup_command,
}
@@ -218,7 +218,7 @@ impl ToolEmitter {
emit_patch_end(ctx, String::new(), (*message).to_string(), false).await;
}
(Self::UnifiedExec { command, cwd, .. }, ToolEventStage::Begin) => {
emit_exec_command_begin(ctx, &[command.to_string()], cwd.as_path(), false).await;
emit_exec_command_begin(ctx, command, cwd.as_path(), false).await;
}
(Self::UnifiedExec { .. }, ToolEventStage::Success(output)) => {
emit_exec_end(
codex-rs/core/src/tools/handlers/shell.rs
+9 -14
View File
@@ -324,8 +324,11 @@ impl ShellHandler {
#[cfg(test)]
mod tests {
use std::path::PathBuf;
use crate::is_safe_command::is_known_safe_command;
use crate::shell::BashShell;
use crate::shell::PowerShellConfig;
use crate::shell::Shell;
use crate::shell::ZshShell;
@@ -335,27 +338,19 @@ mod tests {
#[test]
fn commands_generated_by_shell_command_handler_can_be_matched_by_is_known_safe_command() {
let bash_shell = Shell::Bash(BashShell {
shell_path: "/bin/bash".to_string(),
bashrc_path: "/home/user/.bashrc".to_string(),
shell_path: PathBuf::from("/bin/bash"),
});
assert_safe(&bash_shell, "ls -la");
let zsh_shell = Shell::Zsh(ZshShell {
shell_path: "/bin/zsh".to_string(),
zshrc_path: "/home/user/.zshrc".to_string(),
shell_path: PathBuf::from("/bin/zsh"),
});
assert_safe(&zsh_shell, "ls -la");
#[cfg(target_os = "windows")]
{
use crate::shell::PowerShellConfig;
let powershell = Shell::PowerShell(PowerShellConfig {
exe: "pwsh.exe".to_string(),
bash_exe_fallback: None,
});
assert_safe(&powershell, "ls -Name");
}
let powershell = Shell::PowerShell(PowerShellConfig {
shell_path: PathBuf::from("pwsh.exe"),
});
assert_safe(&powershell, "ls -Name");
}
fn assert_safe(shell: &Shell, command: &str) {
codex-rs/core/src/tools/handlers/unified_exec.rs
+26 -12
View File
@@ -5,6 +5,7 @@ use crate::is_safe_command::is_known_safe_command;
use crate::protocol::EventMsg;
use crate::protocol::ExecCommandOutputDeltaEvent;
use crate::protocol::ExecOutputStream;
use crate::shell::get_shell_by_model_provided_path;
use crate::tools::context::ToolInvocation;
use crate::tools::context::ToolOutput;
use crate::tools::context::ToolPayload;
@@ -32,8 +33,8 @@ struct ExecCommandArgs {
shell: String,
#[serde(default = "default_login")]
login: bool,
#[serde(default)]
yield_time_ms: Option<u64>,
#[serde(default = "default_exec_yield_time_ms")]
yield_time_ms: u64,
#[serde(default)]
max_output_tokens: Option<usize>,
#[serde(default)]
@@ -47,12 +48,20 @@ struct WriteStdinArgs {
session_id: i32,
#[serde(default)]
chars: String,
#[serde(default)]
yield_time_ms: Option<u64>,
#[serde(default = "default_write_stdin_yield_time_ms")]
yield_time_ms: u64,
#[serde(default)]
max_output_tokens: Option<usize>,
}
fn default_exec_yield_time_ms() -> u64 {
10000
}
fn default_write_stdin_yield_time_ms() -> u64 {
250
}
fn default_shell() -> String {
"/bin/bash".to_string()
}
@@ -84,7 +93,8 @@ impl ToolHandler for UnifiedExecHandler {
let Ok(params) = serde_json::from_str::<ExecCommandArgs>(arguments) else {
return true;
};
!is_known_safe_command(&["bash".to_string(), "-lc".to_string(), params.cmd])
let command = get_command(&params);
!is_known_safe_command(&command)
}
async fn handle(&self, invocation: ToolInvocation) -> Result<ToolOutput, FunctionCallError> {
@@ -117,15 +127,15 @@ impl ToolHandler for UnifiedExecHandler {
"failed to parse exec_command arguments: {err:?}"
))
})?;
let command = get_command(&args);
let ExecCommandArgs {
cmd,
workdir,
shell,
login,
yield_time_ms,
max_output_tokens,
with_escalated_permissions,
justification,
..
} = args;
if with_escalated_permissions.unwrap_or(false)
@@ -152,15 +162,14 @@ impl ToolHandler for UnifiedExecHandler {
&context.call_id,
None,
);
let emitter = ToolEmitter::unified_exec(cmd.clone(), cwd.clone(), true);
let emitter = ToolEmitter::unified_exec(&command, cwd.clone(), true);
emitter.emit(event_ctx, ToolEventStage::Begin).await;
manager
.exec_command(
ExecCommandRequest {
command: &cmd,
shell: &shell,
login,
command,
yield_time_ms,
max_output_tokens,
workdir,
@@ -221,6 +230,11 @@ impl ToolHandler for UnifiedExecHandler {
}
}
fn get_command(args: &ExecCommandArgs) -> Vec<String> {
let shell = get_shell_by_model_provided_path(&PathBuf::from(args.shell.clone()));
shell.derive_exec_args(&args.cmd, args.login)
}
fn format_response(response: &UnifiedExecResponse) -> String {
let mut sections = Vec::new();
codex-rs/core/src/tools/mod.rs
+3 -1
View File
@@ -9,6 +9,8 @@ pub mod runtimes;
pub mod sandboxing;
pub mod spec;
use crate::context_manager::MODEL_FORMAT_MAX_BYTES;
use crate::context_manager::MODEL_FORMAT_MAX_LINES;
use crate::context_manager::format_output_for_model_body;
use crate::exec::ExecToolCallOutput;
pub use router::ToolRouter;
@@ -75,5 +77,5 @@ pub fn format_exec_output_str(exec_output: &ExecToolCallOutput) -> String {
};
// Truncate for model consumption before serialization.
format_output_for_model_body(&body)
format_output_for_model_body(&body, MODEL_FORMAT_MAX_BYTES, MODEL_FORMAT_MAX_LINES)
}
codex-rs/core/src/unified_exec/mod.rs
+27 -40
View File
@@ -41,7 +41,6 @@ mod session_manager;
pub(crate) use errors::UnifiedExecError;
pub(crate) use session::UnifiedExecSession;
pub(crate) const DEFAULT_YIELD_TIME_MS: u64 = 10_000;
pub(crate) const MIN_YIELD_TIME_MS: u64 = 250;
pub(crate) const MAX_YIELD_TIME_MS: u64 = 30_000;
pub(crate) const DEFAULT_MAX_OUTPUT_TOKENS: usize = 10_000;
@@ -64,11 +63,9 @@ impl UnifiedExecContext {
}
#[derive(Debug)]
pub(crate) struct ExecCommandRequest<'a> {
pub command: &'a str,
pub shell: &'a str,
pub login: bool,
pub yield_time_ms: Option<u64>,
pub(crate) struct ExecCommandRequest {
pub command: Vec<String>,
pub yield_time_ms: u64,
pub max_output_tokens: Option<usize>,
pub workdir: Option<PathBuf>,
pub with_escalated_permissions: Option<bool>,
@@ -79,7 +76,7 @@ pub(crate) struct ExecCommandRequest<'a> {
pub(crate) struct WriteStdinRequest<'a> {
pub session_id: i32,
pub input: &'a str,
pub yield_time_ms: Option<u64>,
pub yield_time_ms: u64,
pub max_output_tokens: Option<usize>,
}
@@ -105,16 +102,13 @@ struct SessionEntry {
session_ref: Arc<Session>,
turn_ref: Arc<TurnContext>,
call_id: String,
command: String,
command: Vec<String>,
cwd: PathBuf,
started_at: tokio::time::Instant,
}
pub(crate) fn clamp_yield_time(yield_time_ms: Option<u64>) -> u64 {
match yield_time_ms {
Some(value) => value.clamp(MIN_YIELD_TIME_MS, MAX_YIELD_TIME_MS),
None => DEFAULT_YIELD_TIME_MS,
}
pub(crate) fn clamp_yield_time(yield_time_ms: u64) -> u64 {
yield_time_ms.clamp(MIN_YIELD_TIME_MS, MAX_YIELD_TIME_MS)
}
pub(crate) fn resolve_max_tokens(max_tokens: Option<usize>) -> usize {
@@ -187,7 +181,7 @@ mod tests {
session: &Arc<Session>,
turn: &Arc<TurnContext>,
cmd: &str,
yield_time_ms: Option<u64>,
yield_time_ms: u64,
) -> Result<UnifiedExecResponse, UnifiedExecError> {
let context =
UnifiedExecContext::new(Arc::clone(session), Arc::clone(turn), "call".to_string());
@@ -197,9 +191,7 @@ mod tests {
.unified_exec_manager
.exec_command(
ExecCommandRequest {
command: cmd,
shell: "/bin/bash",
login: true,
command: vec!["bash".to_string(), "-lc".to_string(), cmd.to_string()],
yield_time_ms,
max_output_tokens: None,
workdir: None,
@@ -215,7 +207,7 @@ mod tests {
session: &Arc<Session>,
session_id: i32,
input: &str,
yield_time_ms: Option<u64>,
yield_time_ms: u64,
) -> Result<UnifiedExecResponse, UnifiedExecError> {
session
.services
@@ -253,14 +245,14 @@ mod tests {
let (session, turn) = test_session_and_turn();
let open_shell = exec_command(&session, &turn, "bash -i", Some(2_500)).await?;
let open_shell = exec_command(&session, &turn, "bash -i", 2_500).await?;
let session_id = open_shell.session_id.expect("expected session_id");
write_stdin(
&session,
session_id,
"export CODEX_INTERACTIVE_SHELL_VAR=codex\n",
Some(2_500),
2_500,
)
.await?;
@@ -268,7 +260,7 @@ mod tests {
&session,
session_id,
"echo $CODEX_INTERACTIVE_SHELL_VAR\n",
Some(2_500),
2_500,
)
.await?;
assert!(
@@ -285,24 +277,19 @@ mod tests {
let (session, turn) = test_session_and_turn();
let shell_a = exec_command(&session, &turn, "bash -i", Some(2_500)).await?;
let shell_a = exec_command(&session, &turn, "bash -i", 2_500).await?;
let session_a = shell_a.session_id.expect("expected session id");
write_stdin(
&session,
session_a,
"export CODEX_INTERACTIVE_SHELL_VAR=codex\n",
Some(2_500),
2_500,
)
.await?;
let out_2 = exec_command(
&session,
&turn,
"echo $CODEX_INTERACTIVE_SHELL_VAR",
Some(2_500),
)
.await?;
let out_2 =
exec_command(&session, &turn, "echo $CODEX_INTERACTIVE_SHELL_VAR", 2_500).await?;
assert!(
out_2.session_id.is_none(),
"short command should not retain a session"
@@ -316,7 +303,7 @@ mod tests {
&session,
session_a,
"echo $CODEX_INTERACTIVE_SHELL_VAR\n",
Some(2_500),
2_500,
)
.await?;
assert!(
@@ -333,14 +320,14 @@ mod tests {
let (session, turn) = test_session_and_turn();
let open_shell = exec_command(&session, &turn, "bash -i", Some(2_500)).await?;
let open_shell = exec_command(&session, &turn, "bash -i", 2_500).await?;
let session_id = open_shell.session_id.expect("expected session id");
write_stdin(
&session,
session_id,
"export CODEX_INTERACTIVE_SHELL_VAR=codex\n",
Some(2_500),
2_500,
)
.await?;
@@ -348,7 +335,7 @@ mod tests {
&session,
session_id,
"sleep 5 && echo $CODEX_INTERACTIVE_SHELL_VAR\n",
Some(10),
10,
)
.await?;
assert!(
@@ -358,7 +345,7 @@ mod tests {
tokio::time::sleep(Duration::from_secs(7)).await;
let out_3 = write_stdin(&session, session_id, "", Some(100)).await?;
let out_3 = write_stdin(&session, session_id, "", 100).await?;
assert!(
out_3.output.contains("codex"),
@@ -373,7 +360,7 @@ mod tests {
async fn requests_with_large_timeout_are_capped() -> anyhow::Result<()> {
let (session, turn) = test_session_and_turn();
let result = exec_command(&session, &turn, "echo codex", Some(120_000)).await?;
let result = exec_command(&session, &turn, "echo codex", 120_000).await?;
assert!(result.session_id.is_none());
assert!(result.output.contains("codex"));
@@ -385,7 +372,7 @@ mod tests {
#[ignore] // Ignored while we have a better way to test this.
async fn completed_commands_do_not_persist_sessions() -> anyhow::Result<()> {
let (session, turn) = test_session_and_turn();
let result = exec_command(&session, &turn, "echo codex", Some(2_500)).await?;
let result = exec_command(&session, &turn, "echo codex", 2_500).await?;
assert!(
result.session_id.is_none(),
@@ -412,14 +399,14 @@ mod tests {
let (session, turn) = test_session_and_turn();
let open_shell = exec_command(&session, &turn, "bash -i", Some(2_500)).await?;
let open_shell = exec_command(&session, &turn, "bash -i", 2_500).await?;
let session_id = open_shell.session_id.expect("expected session id");
write_stdin(&session, session_id, "exit\n", Some(2_500)).await?;
write_stdin(&session, session_id, "exit\n", 2_500).await?;
tokio::time::sleep(Duration::from_millis(200)).await;
let err = write_stdin(&session, session_id, "", Some(100))
let err = write_stdin(&session, session_id, "", 100)
.await
.expect_err("expected unknown session error");
codex-rs/core/src/unified_exec/session_manager.rs
+11 -19
View File
@@ -19,7 +19,6 @@ use crate::tools::runtimes::unified_exec::UnifiedExecRuntime;
use crate::tools::sandboxing::ToolCtx;
use super::ExecCommandRequest;
use super::MIN_YIELD_TIME_MS;
use super::SessionEntry;
use super::UnifiedExecContext;
use super::UnifiedExecError;
@@ -36,23 +35,17 @@ use super::truncate_output_to_tokens;
impl UnifiedExecSessionManager {
pub(crate) async fn exec_command(
&self,
request: ExecCommandRequest<'_>,
request: ExecCommandRequest,
context: &UnifiedExecContext,
) -> Result<UnifiedExecResponse, UnifiedExecError> {
let cwd = request
.workdir
.clone()
.unwrap_or_else(|| context.turn.cwd.clone());
let shell_flag = if request.login { "-lc" } else { "-c" };
let command = vec![
request.shell.to_string(),
shell_flag.to_string(),
request.command.to_string(),
];
let session = self
.open_session_with_sandbox(
command,
&request.command,
cwd.clone(),
request.with_escalated_permissions,
request.justification,
@@ -61,8 +54,7 @@ impl UnifiedExecSessionManager {
.await?;
let max_tokens = resolve_max_tokens(request.max_output_tokens);
let yield_time_ms =
clamp_yield_time(Some(request.yield_time_ms.unwrap_or(MIN_YIELD_TIME_MS)));
let yield_time_ms = clamp_yield_time(request.yield_time_ms);
let start = Instant::now();
let (output_buffer, output_notify) = session.output_handles();
@@ -79,7 +71,7 @@ impl UnifiedExecSessionManager {
None
} else {
Some(
self.store_session(session, context, request.command, cwd.clone(), start)
self.store_session(session, context, &request.command, cwd.clone(), start)
.await,
)
};
@@ -99,7 +91,7 @@ impl UnifiedExecSessionManager {
let exit = response.exit_code.unwrap_or(-1);
Self::emit_exec_end_from_context(
context,
request.command.to_string(),
&request.command,
cwd,
response.output.clone(),
exit,
@@ -224,7 +216,7 @@ impl UnifiedExecSessionManager {
&self,
session: UnifiedExecSession,
context: &UnifiedExecContext,
command: &str,
command: &[String],
cwd: PathBuf,
started_at: Instant,
) -> i32 {
@@ -236,7 +228,7 @@ impl UnifiedExecSessionManager {
session_ref: Arc::clone(&context.session),
turn_ref: Arc::clone(&context.turn),
call_id: context.call_id.clone(),
command: command.to_string(),
command: command.to_vec(),
cwd,
started_at,
};
@@ -264,7 +256,7 @@ impl UnifiedExecSessionManager {
&entry.call_id,
None,
);
let emitter = ToolEmitter::unified_exec(entry.command, entry.cwd, true);
let emitter = ToolEmitter::unified_exec(&entry.command, entry.cwd, true);
emitter
.emit(event_ctx, ToolEventStage::Success(output))
.await;
@@ -272,7 +264,7 @@ impl UnifiedExecSessionManager {
async fn emit_exec_end_from_context(
context: &UnifiedExecContext,
command: String,
command: &[String],
cwd: PathBuf,
aggregated_output: String,
exit_code: i32,
@@ -321,7 +313,7 @@ impl UnifiedExecSessionManager {
pub(super) async fn open_session_with_sandbox(
&self,
command: Vec<String>,
command: &[String],
cwd: PathBuf,
with_escalated_permissions: Option<bool>,
justification: Option<String>,
@@ -330,7 +322,7 @@ impl UnifiedExecSessionManager {
let mut orchestrator = ToolOrchestrator::new();
let mut runtime = UnifiedExecRuntime::new(self);
let req = UnifiedExecToolRequest::new(
command,
command.to_vec(),
cwd,
create_env(&context.turn.shell_environment_policy),
with_escalated_permissions,
codex-rs/core/tests/chat_completions_payload.rs
+9 -53
View File
@@ -10,11 +10,11 @@ use codex_core::ModelProviderInfo;
use codex_core::Prompt;
use codex_core::ResponseItem;
use codex_core::WireApi;
use codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_protocol::ConversationId;
use codex_protocol::models::ReasoningItemContent;
use core_test_support::load_default_config_for_test;
use core_test_support::skip_if_no_network;
use futures::StreamExt;
use serde_json::Value;
use tempfile::TempDir;
@@ -24,10 +24,6 @@ use wiremock::ResponseTemplate;
use wiremock::matchers::method;
use wiremock::matchers::path;
fn network_disabled() -> bool {
std::env::var(CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR).is_ok()
}
async fn run_request(input: Vec<ResponseItem>) -> Value {
let server = MockServer::start().await;
@@ -191,12 +187,7 @@ fn first_assistant(messages: &[Value]) -> &Value {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn omits_reasoning_when_none_present() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![user_message("u1"), assistant_message("a1")]).await;
let messages = messages_from(&body);
@@ -208,12 +199,7 @@ async fn omits_reasoning_when_none_present() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn attaches_reasoning_to_previous_assistant() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![
user_message("u1"),
@@ -230,12 +216,7 @@ async fn attaches_reasoning_to_previous_assistant() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn attaches_reasoning_to_function_call_anchor() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![
user_message("u1"),
@@ -257,12 +238,7 @@ async fn attaches_reasoning_to_function_call_anchor() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn attaches_reasoning_to_local_shell_call() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![
user_message("u1"),
@@ -282,12 +258,7 @@ async fn attaches_reasoning_to_local_shell_call() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn drops_reasoning_when_last_role_is_user() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![
assistant_message("aPrev"),
@@ -301,12 +272,7 @@ async fn drops_reasoning_when_last_role_is_user() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn ignores_reasoning_before_last_user() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![
user_message("u1"),
@@ -321,12 +287,7 @@ async fn ignores_reasoning_before_last_user() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn skips_empty_reasoning_segments() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![
user_message("u1"),
@@ -342,12 +303,7 @@ async fn skips_empty_reasoning_segments() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn suppresses_duplicate_assistant_messages() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let body = run_request(vec![assistant_message("dup"), assistant_message("dup")]).await;
let messages = messages_from(&body);
codex-rs/core/tests/chat_completions_sse.rs
+44 -58
View File
@@ -10,11 +10,11 @@ use codex_core::Prompt;
use codex_core::ResponseEvent;
use codex_core::ResponseItem;
use codex_core::WireApi;
use codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_protocol::ConversationId;
use codex_protocol::models::ReasoningItemContent;
use core_test_support::load_default_config_for_test;
use core_test_support::skip_if_no_network;
use futures::StreamExt;
use tempfile::TempDir;
use wiremock::Mock;
@@ -23,10 +23,6 @@ use wiremock::ResponseTemplate;
use wiremock::matchers::method;
use wiremock::matchers::path;
fn network_disabled() -> bool {
std::env::var(CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR).is_ok()
}
async fn run_stream(sse_body: &str) -> Vec<ResponseEvent> {
run_stream_with_bytes(sse_body.as_bytes()).await
}
@@ -157,12 +153,7 @@ fn assert_reasoning(item: &ResponseItem, expected: &str) {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn streams_text_without_reasoning() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse = concat!(
"data: {\"choices\":[{\"delta\":{\"content\":\"hi\"}}]}\n\n",
@@ -193,12 +184,7 @@ async fn streams_text_without_reasoning() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn streams_reasoning_from_string_delta() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse = concat!(
"data: {\"choices\":[{\"delta\":{\"reasoning\":\"think1\"}}]}\n\n",
@@ -215,7 +201,13 @@ async fn streams_reasoning_from_string_delta() {
}
match &events[1] {
ResponseEvent::ReasoningContentDelta(text) => assert_eq!(text, "think1"),
ResponseEvent::ReasoningContentDelta {
delta,
content_index,
} => {
assert_eq!(delta, "think1");
assert_eq!(content_index, &0);
}
other => panic!("expected reasoning delta, got {other:?}"),
}
@@ -244,12 +236,7 @@ async fn streams_reasoning_from_string_delta() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn streams_reasoning_from_object_delta() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse = concat!(
"data: {\"choices\":[{\"delta\":{\"reasoning\":{\"text\":\"partA\"}}}]}\n\n",
@@ -267,12 +254,24 @@ async fn streams_reasoning_from_object_delta() {
}
match &events[1] {
ResponseEvent::ReasoningContentDelta(text) => assert_eq!(text, "partA"),
ResponseEvent::ReasoningContentDelta {
delta,
content_index,
} => {
assert_eq!(delta, "partA");
assert_eq!(content_index, &0);
}
other => panic!("expected reasoning delta, got {other:?}"),
}
match &events[2] {
ResponseEvent::ReasoningContentDelta(text) => assert_eq!(text, "partB"),
ResponseEvent::ReasoningContentDelta {
delta,
content_index,
} => {
assert_eq!(delta, "partB");
assert_eq!(content_index, &1);
}
other => panic!("expected reasoning delta, got {other:?}"),
}
@@ -301,12 +300,7 @@ async fn streams_reasoning_from_object_delta() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn streams_reasoning_from_final_message() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse = "data: {\"choices\":[{\"message\":{\"reasoning\":\"final-cot\"},\"finish_reason\":\"stop\"}]}\n\n";
@@ -319,7 +313,13 @@ async fn streams_reasoning_from_final_message() {
}
match &events[1] {
ResponseEvent::ReasoningContentDelta(text) => assert_eq!(text, "final-cot"),
ResponseEvent::ReasoningContentDelta {
delta,
content_index,
} => {
assert_eq!(delta, "final-cot");
assert_eq!(content_index, &0);
}
other => panic!("expected reasoning delta, got {other:?}"),
}
@@ -333,12 +333,7 @@ async fn streams_reasoning_from_final_message() {
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn streams_reasoning_before_tool_call() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse = concat!(
"data: {\"choices\":[{\"delta\":{\"reasoning\":\"pre-tool\"}}]}\n\n",
@@ -354,7 +349,13 @@ async fn streams_reasoning_before_tool_call() {
}
match &events[1] {
ResponseEvent::ReasoningContentDelta(text) => assert_eq!(text, "pre-tool"),
ResponseEvent::ReasoningContentDelta {
delta,
content_index,
} => {
assert_eq!(delta, "pre-tool");
assert_eq!(content_index, &0);
}
other => panic!("expected reasoning delta, got {other:?}"),
}
@@ -383,12 +384,7 @@ async fn streams_reasoning_before_tool_call() {
#[tokio::test]
#[traced_test]
async fn chat_sse_emits_failed_on_parse_error() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse_body = concat!("data: not-json\n\n", "data: [DONE]\n\n");
@@ -420,12 +416,7 @@ async fn chat_sse_emits_failed_on_parse_error() {
#[tokio::test]
#[traced_test]
async fn chat_sse_done_chunk_emits_event() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let sse_body = "data: [DONE]\n\n";
@@ -443,12 +434,7 @@ async fn chat_sse_done_chunk_emits_event() {
#[tokio::test]
#[traced_test]
async fn chat_sse_emits_error_on_invalid_utf8() {
if network_disabled() {
println!(
"Skipping test because it cannot execute when network is disabled in a Codex sandbox."
);
return;
}
skip_if_no_network!();
let _ = run_stream_with_bytes(b"data: \x80\x80\n\n").await;
codex-rs/core/tests/common/responses.rs
+106
View File
@@ -1,6 +1,7 @@
use std::sync::Arc;
use std::sync::Mutex;
use anyhow::Result;
use serde_json::Value;
use wiremock::BodyPrintLimit;
use wiremock::Match;
@@ -12,6 +13,8 @@ use wiremock::ResponseTemplate;
use wiremock::matchers::method;
use wiremock::matchers::path_regex;
use crate::test_codex::ApplyPatchModelOutput;
#[derive(Debug, Clone)]
pub struct ResponseMock {
requests: Arc<Mutex<Vec<ResponsesRequest>>>,
@@ -36,6 +39,10 @@ impl ResponseMock {
self.requests.lock().unwrap().clone()
}
pub fn last_request(&self) -> Option<ResponsesRequest> {
self.requests.lock().unwrap().last().cloned()
}
/// Returns true if any captured request contains a `function_call` with the
/// provided `call_id`.
pub fn saw_function_call(&self, call_id: &str) -> bool {
@@ -128,6 +135,42 @@ impl ResponsesRequest {
.map(str::to_string)
}
pub fn function_call_output_content_and_success(
&self,
call_id: &str,
) -> Option<(Option<String>, Option<bool>)> {
self.call_output_content_and_success(call_id, "function_call_output")
}
pub fn custom_tool_call_output_content_and_success(
&self,
call_id: &str,
) -> Option<(Option<String>, Option<bool>)> {
self.call_output_content_and_success(call_id, "custom_tool_call_output")
}
fn call_output_content_and_success(
&self,
call_id: &str,
call_type: &str,
) -> Option<(Option<String>, Option<bool>)> {
let output = self
.call_output(call_id, call_type)
.get("output")
.cloned()
.unwrap_or(Value::Null);
match output {
Value::String(text) => Some((Some(text), None)),
Value::Object(obj) => Some((
obj.get("content")
.and_then(Value::as_str)
.map(str::to_string),
obj.get("success").and_then(Value::as_bool),
)),
_ => Some((None, None)),
}
}
pub fn header(&self, name: &str) -> Option<String> {
self.0
.headers
@@ -294,6 +337,7 @@ pub fn ev_reasoning_summary_text_delta(delta: &str) -> Value {
serde_json::json!({
"type": "response.reasoning_summary_text.delta",
"delta": delta,
"summary_index": 0,
})
}
@@ -301,6 +345,7 @@ pub fn ev_reasoning_text_delta(delta: &str) -> Value {
serde_json::json!({
"type": "response.reasoning_text.delta",
"delta": delta,
"content_index": 0,
})
}
@@ -367,6 +412,21 @@ pub fn ev_local_shell_call(call_id: &str, status: &str, command: Vec<&str>) -> V
})
}
pub fn ev_apply_patch_call(
call_id: &str,
patch: &str,
output_type: ApplyPatchModelOutput,
) -> Value {
match output_type {
ApplyPatchModelOutput::Freeform => ev_apply_patch_custom_tool_call(call_id, patch),
ApplyPatchModelOutput::Function => ev_apply_patch_function_call(call_id, patch),
ApplyPatchModelOutput::Shell => ev_apply_patch_shell_call(call_id, patch),
ApplyPatchModelOutput::ShellViaHeredoc => {
ev_apply_patch_shell_call_via_heredoc(call_id, patch)
}
}
}
/// Convenience: SSE event for an `apply_patch` custom tool call with raw patch
/// text. This mirrors the payload produced by the Responses API when the model
/// invokes `apply_patch` directly (before we convert it to a function call).
@@ -400,6 +460,21 @@ pub fn ev_apply_patch_function_call(call_id: &str, patch: &str) -> Value {
})
}
pub fn ev_apply_patch_shell_call(call_id: &str, patch: &str) -> Value {
let args = serde_json::json!({ "command": ["apply_patch", patch] });
let arguments = serde_json::to_string(&args).expect("serialize apply_patch arguments");
ev_function_call(call_id, "shell", &arguments)
}
pub fn ev_apply_patch_shell_call_via_heredoc(call_id: &str, patch: &str) -> Value {
let script = format!("apply_patch <<'EOF'\n{patch}\nEOF\n");
let args = serde_json::json!({ "command": ["bash", "-lc", script] });
let arguments = serde_json::to_string(&args).expect("serialize apply_patch arguments");
ev_function_call(call_id, "shell", &arguments)
}
pub fn sse_failed(id: &str, code: &str, message: &str) -> String {
sse(vec![serde_json::json!({
"type": "response.failed",
@@ -453,6 +528,37 @@ pub async fn start_mock_server() -> MockServer {
.await
}
#[derive(Clone)]
pub struct FunctionCallResponseMocks {
pub function_call: ResponseMock,
pub completion: ResponseMock,
}
pub async fn mount_function_call_agent_response(
server: &MockServer,
call_id: &str,
arguments: &str,
tool_name: &str,
) -> FunctionCallResponseMocks {
let first_response = sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, tool_name, arguments),
ev_completed("resp-1"),
]);
let function_call = mount_sse_once(server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
let completion = mount_sse_once(server, second_response).await;
FunctionCallResponseMocks {
function_call,
completion,
}
}
/// Mounts a sequence of SSE response bodies and serves them in order for each
/// POST to `/v1/responses`. Panics if more requests are received than bodies
/// provided. Also asserts the exact number of expected calls.
codex-rs/core/tests/common/test_codex.rs
+48 -5
View File
@@ -29,6 +29,15 @@ use crate::wait_for_event;
type ConfigMutator = dyn FnOnce(&mut Config) + Send;
/// A collection of different ways the model can output an apply_patch call
#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
pub enum ApplyPatchModelOutput {
Freeform,
Function,
Shell,
ShellViaHeredoc,
}
pub struct TestCodexBuilder {
config_mutators: Vec<Box<ConfigMutator>>,
}
@@ -71,6 +80,7 @@ impl TestCodexBuilder {
resume_from: Option<PathBuf>,
) -> anyhow::Result<TestCodex> {
let (config, cwd) = self.prepare_config(server, &home).await?;
let conversation_manager = ConversationManager::with_auth(CodexAuth::from_api_key("dummy"));
let new_conversation = match resume_from {
@@ -79,15 +89,20 @@ impl TestCodexBuilder {
CodexAuth::from_api_key("dummy"),
);
conversation_manager
.resume_conversation_from_rollout(config, path, auth_manager)
.resume_conversation_from_rollout(config.clone(), path, auth_manager)
.await?
}
None => {
conversation_manager
.new_conversation(config.clone())
.await?
}
None => conversation_manager.new_conversation(config).await?,
};
Ok(TestCodex {
home,
cwd,
config,
codex: new_conversation.conversation,
session_configured: new_conversation.session_configured,
})
@@ -131,6 +146,7 @@ pub struct TestCodex {
pub cwd: Arc<TempDir>,
pub codex: Arc<CodexConversation>,
pub session_configured: SessionConfiguredEvent,
pub config: Config,
}
impl TestCodex {
@@ -143,14 +159,28 @@ impl TestCodex {
}
pub async fn submit_turn(&self, prompt: &str) -> Result<()> {
self.submit_turn_with_policy(prompt, SandboxPolicy::DangerFullAccess)
.await
self.submit_turn_with_policies(
prompt,
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
)
.await
}
pub async fn submit_turn_with_policy(
&self,
prompt: &str,
sandbox_policy: SandboxPolicy,
) -> Result<()> {
self.submit_turn_with_policies(prompt, AskForApproval::Never, sandbox_policy)
.await
}
pub async fn submit_turn_with_policies(
&self,
prompt: &str,
approval_policy: AskForApproval,
sandbox_policy: SandboxPolicy,
) -> Result<()> {
let session_model = self.session_configured.model.clone();
self.codex
@@ -160,7 +190,7 @@ impl TestCodex {
}],
final_output_json_schema: None,
cwd: self.cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
approval_policy,
sandbox_policy,
model: session_model,
effort: None,
@@ -258,6 +288,19 @@ impl TestCodexHarness {
.expect("output string")
.to_string()
}
pub async fn apply_patch_output(
&self,
call_id: &str,
output_type: ApplyPatchModelOutput,
) -> String {
match output_type {
ApplyPatchModelOutput::Freeform => self.custom_tool_call_output(call_id).await,
ApplyPatchModelOutput::Function
| ApplyPatchModelOutput::Shell
| ApplyPatchModelOutput::ShellViaHeredoc => self.function_call_stdout(call_id).await,
}
}
}
fn custom_tool_call_output<'a>(bodies: &'a [Value], call_id: &str) -> &'a Value {
codex-rs/core/tests/suite/apply_patch_cli.rs
+200 -76
View File
@@ -1,6 +1,8 @@
#![allow(clippy::expect_used)]
use anyhow::Result;
use core_test_support::responses::ev_apply_patch_call;
use core_test_support::test_codex::ApplyPatchModelOutput;
use pretty_assertions::assert_eq;
use std::fs;
@@ -25,6 +27,7 @@ use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodexHarness;
use core_test_support::wait_for_event;
use serde_json::json;
use test_case::test_case;
async fn apply_patch_harness() -> Result<TestCodexHarness> {
apply_patch_harness_with(|_| {}).await
@@ -45,19 +48,25 @@ async fn mount_apply_patch(
call_id: &str,
patch: &str,
assistant_msg: &str,
output_type: ApplyPatchModelOutput,
) {
mount_sse_sequence(
harness.server(),
apply_patch_responses(call_id, patch, assistant_msg),
apply_patch_responses(call_id, patch, assistant_msg, output_type),
)
.await;
}
fn apply_patch_responses(call_id: &str, patch: &str, assistant_msg: &str) -> Vec<String> {
fn apply_patch_responses(
call_id: &str,
patch: &str,
assistant_msg: &str,
output_type: ApplyPatchModelOutput,
) -> Vec<String> {
vec![
sse(vec![
ev_response_created("resp-1"),
ev_apply_patch_function_call(call_id, patch),
ev_apply_patch_call(call_id, patch, output_type),
ev_completed("resp-1"),
]),
sse(vec![
@@ -68,7 +77,13 @@ fn apply_patch_responses(call_id: &str, patch: &str, assistant_msg: &str) -> Vec
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_multiple_operations_integration() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_multiple_operations_integration(
output_type: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness_with(|config| {
@@ -86,11 +101,11 @@ async fn apply_patch_cli_multiple_operations_integration() -> Result<()> {
let patch = "*** Begin Patch\n*** Add File: nested/new.txt\n+created\n*** Delete File: delete.txt\n*** Update File: modify.txt\n@@\n-line2\n+changed\n*** End Patch";
let call_id = "apply-multi-ops";
mount_apply_patch(&harness, call_id, patch, "done").await;
mount_apply_patch(&harness, call_id, patch, "done", output_type).await;
harness.submit("please apply multi-ops patch").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, output_type).await;
let expected = r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
@@ -113,7 +128,11 @@ D delete.txt
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_multiple_chunks() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_multiple_chunks(model_output: ApplyPatchModelOutput) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -123,7 +142,7 @@ async fn apply_patch_cli_multiple_chunks() -> Result<()> {
let patch = "*** Begin Patch\n*** Update File: multi.txt\n@@\n-line2\n+changed2\n@@\n-line4\n+changed4\n*** End Patch";
let call_id = "apply-multi-chunks";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply multi-chunk patch").await?;
@@ -135,7 +154,13 @@ async fn apply_patch_cli_multiple_chunks() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_moves_file_to_new_directory() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_moves_file_to_new_directory(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -147,7 +172,7 @@ async fn apply_patch_cli_moves_file_to_new_directory() -> Result<()> {
let patch = "*** Begin Patch\n*** Update File: old/name.txt\n*** Move to: renamed/dir/name.txt\n@@\n-old content\n+new content\n*** End Patch";
let call_id = "apply-move";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply move patch").await?;
@@ -157,7 +182,13 @@ async fn apply_patch_cli_moves_file_to_new_directory() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_updates_file_appends_trailing_newline() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_updates_file_appends_trailing_newline(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -167,7 +198,7 @@ async fn apply_patch_cli_updates_file_appends_trailing_newline() -> Result<()> {
let patch = "*** Begin Patch\n*** Update File: no_newline.txt\n@@\n-no newline at end\n+first line\n+second line\n*** End Patch";
let call_id = "apply-append-nl";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply newline patch").await?;
@@ -178,7 +209,13 @@ async fn apply_patch_cli_updates_file_appends_trailing_newline() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_insert_only_hunk_modifies_file() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_insert_only_hunk_modifies_file(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -188,7 +225,7 @@ async fn apply_patch_cli_insert_only_hunk_modifies_file() -> Result<()> {
let patch = "*** Begin Patch\n*** Update File: insert_only.txt\n@@\n alpha\n+beta\n omega\n*** End Patch";
let call_id = "apply-insert-only";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("insert lines via apply_patch").await?;
@@ -197,7 +234,13 @@ async fn apply_patch_cli_insert_only_hunk_modifies_file() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_move_overwrites_existing_destination() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_move_overwrites_existing_destination(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -211,7 +254,7 @@ async fn apply_patch_cli_move_overwrites_existing_destination() -> Result<()> {
let patch = "*** Begin Patch\n*** Update File: old/name.txt\n*** Move to: renamed/dir/name.txt\n@@\n-from\n+new\n*** End Patch";
let call_id = "apply-move-overwrite";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply move overwrite patch").await?;
@@ -221,7 +264,13 @@ async fn apply_patch_cli_move_overwrites_existing_destination() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_move_without_content_change_has_no_turn_diff() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_move_without_content_change_has_no_turn_diff(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -236,7 +285,7 @@ async fn apply_patch_cli_move_without_content_change_has_no_turn_diff() -> Resul
let patch = "*** Begin Patch\n*** Update File: old/name.txt\n*** Move to: renamed/name.txt\n@@\n same\n*** End Patch";
let call_id = "apply-move-no-change";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
let model = test.session_configured.model.clone();
codex
@@ -272,7 +321,13 @@ async fn apply_patch_cli_move_without_content_change_has_no_turn_diff() -> Resul
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_add_overwrites_existing_file() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_add_overwrites_existing_file(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -282,7 +337,7 @@ async fn apply_patch_cli_add_overwrites_existing_file() -> Result<()> {
let patch = "*** Begin Patch\n*** Add File: duplicate.txt\n+new content\n*** End Patch";
let call_id = "apply-add-overwrite";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply add overwrite patch").await?;
@@ -291,18 +346,24 @@ async fn apply_patch_cli_add_overwrites_existing_file() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_rejects_invalid_hunk_header() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_rejects_invalid_hunk_header(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
let patch = "*** Begin Patch\n*** Frobnicate File: foo\n*** End Patch";
let call_id = "apply-invalid-header";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply invalid header patch").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains("apply_patch verification failed"),
@@ -316,7 +377,13 @@ async fn apply_patch_cli_rejects_invalid_hunk_header() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_reports_missing_context() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_reports_missing_context(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -327,11 +394,11 @@ async fn apply_patch_cli_reports_missing_context() -> Result<()> {
let patch =
"*** Begin Patch\n*** Update File: modify.txt\n@@\n-missing\n+changed\n*** End Patch";
let call_id = "apply-missing-context";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply missing context patch").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains("apply_patch verification failed"),
@@ -343,18 +410,24 @@ async fn apply_patch_cli_reports_missing_context() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_reports_missing_target_file() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_reports_missing_target_file(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
let patch = "*** Begin Patch\n*** Update File: missing.txt\n@@\n-nope\n+better\n*** End Patch";
let call_id = "apply-missing-file";
mount_apply_patch(&harness, call_id, patch, "fail").await;
mount_apply_patch(&harness, call_id, patch, "fail", model_output).await;
harness.submit("attempt to update a missing file").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains("apply_patch verification failed"),
"expected verification failure message"
@@ -372,18 +445,24 @@ async fn apply_patch_cli_reports_missing_target_file() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_delete_missing_file_reports_error() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_delete_missing_file_reports_error(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
let patch = "*** Begin Patch\n*** Delete File: missing.txt\n*** End Patch";
let call_id = "apply-delete-missing";
mount_apply_patch(&harness, call_id, patch, "fail").await;
mount_apply_patch(&harness, call_id, patch, "fail", model_output).await;
harness.submit("attempt to delete missing file").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains("apply_patch verification failed"),
@@ -402,18 +481,22 @@ async fn apply_patch_cli_delete_missing_file_reports_error() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_rejects_empty_patch() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_rejects_empty_patch(model_output: ApplyPatchModelOutput) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
let patch = "*** Begin Patch\n*** End Patch";
let call_id = "apply-empty";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply empty patch").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains("patch rejected: empty patch"),
"expected rejection for empty patch: {out}"
@@ -422,7 +505,13 @@ async fn apply_patch_cli_rejects_empty_patch() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_delete_directory_reports_verification_error() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_delete_directory_reports_verification_error(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -431,18 +520,24 @@ async fn apply_patch_cli_delete_directory_reports_verification_error() -> Result
let patch = "*** Begin Patch\n*** Delete File: dir\n*** End Patch";
let call_id = "apply-delete-dir";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("delete a directory via apply_patch").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(out.contains("apply_patch verification failed"));
assert!(out.contains("Failed to read"));
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_rejects_path_traversal_outside_workspace() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_rejects_path_traversal_outside_workspace(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -458,7 +553,7 @@ async fn apply_patch_cli_rejects_path_traversal_outside_workspace() -> Result<()
let patch = "*** Begin Patch\n*** Add File: ../escape.txt\n+outside\n*** End Patch";
let call_id = "apply-path-traversal";
mount_apply_patch(&harness, call_id, patch, "fail").await;
mount_apply_patch(&harness, call_id, patch, "fail", model_output).await;
let sandbox_policy = SandboxPolicy::WorkspaceWrite {
writable_roots: vec![],
@@ -473,7 +568,7 @@ async fn apply_patch_cli_rejects_path_traversal_outside_workspace() -> Result<()
)
.await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains(
"patch rejected: writing outside of the project; rejected by user approval settings"
@@ -488,7 +583,13 @@ async fn apply_patch_cli_rejects_path_traversal_outside_workspace() -> Result<()
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_rejects_move_path_traversal_outside_workspace() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_rejects_move_path_traversal_outside_workspace(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -507,7 +608,7 @@ async fn apply_patch_cli_rejects_move_path_traversal_outside_workspace() -> Resu
let patch = "*** Begin Patch\n*** Update File: stay.txt\n*** Move to: ../escape-move.txt\n@@\n-from\n+to\n*** End Patch";
let call_id = "apply-move-traversal";
mount_apply_patch(&harness, call_id, patch, "fail").await;
mount_apply_patch(&harness, call_id, patch, "fail", model_output).await;
let sandbox_policy = SandboxPolicy::WorkspaceWrite {
writable_roots: vec![],
@@ -519,7 +620,7 @@ async fn apply_patch_cli_rejects_move_path_traversal_outside_workspace() -> Resu
.submit_with_policy("attempt move traversal via apply_patch", sandbox_policy)
.await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(
out.contains(
"patch rejected: writing outside of the project; rejected by user approval settings"
@@ -535,7 +636,13 @@ async fn apply_patch_cli_rejects_move_path_traversal_outside_workspace() -> Resu
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_verification_failure_has_no_side_effects() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_verification_failure_has_no_side_effects(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness_with(|config| {
@@ -547,7 +654,7 @@ async fn apply_patch_cli_verification_failure_has_no_side_effects() -> Result<()
let call_id = "apply-partial-no-side-effects";
let patch = "*** Begin Patch\n*** Add File: created.txt\n+hello\n*** Update File: missing.txt\n@@\n-old\n+new\n*** End Patch";
mount_apply_patch(&harness, call_id, patch, "failed").await;
mount_apply_patch(&harness, call_id, patch, "failed", model_output).await;
harness.submit("attempt partial apply patch").await?;
@@ -700,7 +807,14 @@ async fn apply_patch_function_accepts_lenient_heredoc_wrapped_patch() -> Result<
format!("*** Begin Patch\n*** Add File: {file_name}\n+lenient\n*** End Patch\n");
let wrapped = format!("<<'EOF'\n{patch_inner}EOF\n");
let call_id = "apply-lenient";
mount_apply_patch(&harness, call_id, &wrapped, "ok").await;
mount_apply_patch(
&harness,
call_id,
wrapped.as_str(),
"ok",
ApplyPatchModelOutput::Function,
)
.await;
harness.submit("apply lenient heredoc patch").await?;
@@ -710,7 +824,11 @@ async fn apply_patch_function_accepts_lenient_heredoc_wrapped_patch() -> Result<
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_end_of_file_anchor() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_end_of_file_anchor(model_output: ApplyPatchModelOutput) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -720,7 +838,7 @@ async fn apply_patch_cli_end_of_file_anchor() -> Result<()> {
let patch = "*** Begin Patch\n*** Update File: tail.txt\n@@\n-last\n+end\n*** End of File\n*** End Patch";
let call_id = "apply-eof";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply EOF-anchored patch").await?;
assert_eq!(fs::read_to_string(&target)?, "alpha\nend\n");
@@ -728,7 +846,13 @@ async fn apply_patch_cli_end_of_file_anchor() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_cli_missing_second_chunk_context_rejected() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_cli_missing_second_chunk_context_rejected(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -740,11 +864,11 @@ async fn apply_patch_cli_missing_second_chunk_context_rejected() -> Result<()> {
let patch =
"*** Begin Patch\n*** Update File: two_chunks.txt\n@@\n-b\n+B\n\n-d\n+D\n*** End Patch";
let call_id = "apply-missing-ctx-2nd";
mount_apply_patch(&harness, call_id, patch, "fail").await;
mount_apply_patch(&harness, call_id, patch, "fail", model_output).await;
harness.submit("apply missing context second chunk").await?;
let out = harness.function_call_stdout(call_id).await;
let out = harness.apply_patch_output(call_id, model_output).await;
assert!(out.contains("apply_patch verification failed"));
assert!(
out.contains("Failed to find expected lines in"),
@@ -756,7 +880,13 @@ async fn apply_patch_cli_missing_second_chunk_context_rejected() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_emits_turn_diff_event_with_unified_diff() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_emits_turn_diff_event_with_unified_diff(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -767,16 +897,7 @@ async fn apply_patch_emits_turn_diff_event_with_unified_diff() -> Result<()> {
let call_id = "apply-diff-event";
let file = "udiff.txt";
let patch = format!("*** Begin Patch\n*** Add File: {file}\n+hello\n*** End Patch\n");
let first = sse(vec![
ev_response_created("resp-1"),
ev_apply_patch_function_call(call_id, &patch),
ev_completed("resp-1"),
]);
let second = sse(vec![
ev_assistant_message("msg-1", "ok"),
ev_completed("resp-2"),
]);
mount_sse_sequence(harness.server(), vec![first, second]).await;
mount_apply_patch(&harness, call_id, patch.as_str(), "ok", model_output).await;
let model = test.session_configured.model.clone();
codex
@@ -814,7 +935,13 @@ async fn apply_patch_emits_turn_diff_event_with_unified_diff() -> Result<()> {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_turn_diff_for_rename_with_content_change() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_turn_diff_for_rename_with_content_change(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -829,16 +956,7 @@ async fn apply_patch_turn_diff_for_rename_with_content_change() -> Result<()> {
// Patch: update + move
let call_id = "apply-rename-change";
let patch = "*** Begin Patch\n*** Update File: old.txt\n*** Move to: new.txt\n@@\n-old\n+new\n*** End Patch";
let first = sse(vec![
ev_response_created("resp-1"),
ev_apply_patch_function_call(call_id, patch),
ev_completed("resp-1"),
]);
let second = sse(vec![
ev_assistant_message("msg-1", "ok"),
ev_completed("resp-2"),
]);
mount_sse_sequence(harness.server(), vec![first, second]).await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
let model = test.session_configured.model.clone();
codex
@@ -1031,7 +1149,13 @@ async fn apply_patch_aggregates_diff_preserves_success_after_failure() -> Result
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_change_context_disambiguates_target() -> Result<()> {
#[test_case(ApplyPatchModelOutput::Freeform)]
#[test_case(ApplyPatchModelOutput::Function)]
#[test_case(ApplyPatchModelOutput::Shell)]
#[test_case(ApplyPatchModelOutput::ShellViaHeredoc)]
async fn apply_patch_change_context_disambiguates_target(
model_output: ApplyPatchModelOutput,
) -> Result<()> {
skip_if_no_network!(Ok(()));
let harness = apply_patch_harness().await?;
@@ -1042,7 +1166,7 @@ async fn apply_patch_change_context_disambiguates_target() -> Result<()> {
let patch =
"*** Begin Patch\n*** Update File: multi_ctx.txt\n@@ fn b\n-x=10\n+x=11\n*** End Patch";
let call_id = "apply-ctx";
mount_apply_patch(&harness, call_id, patch, "ok").await;
mount_apply_patch(&harness, call_id, patch, "ok", model_output).await;
harness.submit("apply with change_context").await?;
codex-rs/core/tests/suite/apply_patch_freeform.rs
-1001
View File
File diff suppressed because it is too large Load Diff
codex-rs/core/tests/suite/cli_stream.rs
+1 -6
View File
@@ -110,12 +110,7 @@ async fn exec_cli_applies_experimental_instructions_file() {
"data: {\"type\":\"response.created\",\"response\":{}}\n\n",
"data: {\"type\":\"response.completed\",\"response\":{\"id\":\"r1\"}}\n\n"
);
let resp_mock = core_test_support::responses::mount_sse_once_match(
&server,
path("/v1/responses"),
sse.to_string(),
)
.await;
let resp_mock = core_test_support::responses::mount_sse_once(&server, sse.to_string()).await;
// Create a temporary instructions file with a unique marker we can assert
// appears in the outbound request payload.
codex-rs/core/tests/suite/client.rs
+10 -30
View File
@@ -238,9 +238,7 @@ async fn resume_includes_initial_messages_and_sends_prior_items() {
// Mock server that will receive the resumed request
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
// Configure Codex to resume from our file
let model_provider = ModelProviderInfo {
@@ -381,9 +379,7 @@ async fn includes_base_instructions_override_in_request() {
skip_if_no_network!();
// Mock server
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
@@ -571,9 +567,7 @@ async fn includes_user_instructions_message_in_request() {
skip_if_no_network!();
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
@@ -631,9 +625,7 @@ async fn includes_configured_effort_in_request() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let TestCodex { codex, .. } = test_codex()
.with_model("gpt-5.1-codex")
.with_config(|config| {
@@ -672,9 +664,7 @@ async fn includes_no_effort_in_request() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let TestCodex { codex, .. } = test_codex()
.with_model("gpt-5.1-codex")
.build(&server)
@@ -711,9 +701,7 @@ async fn includes_default_reasoning_effort_in_request_when_defined_by_model_fami
skip_if_no_network!(Ok(()));
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let TestCodex { codex, .. } = test_codex().with_model("gpt-5.1").build(&server).await?;
codex
@@ -746,9 +734,7 @@ async fn includes_default_verbosity_in_request() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let TestCodex { codex, .. } = test_codex().with_model("gpt-5.1").build(&server).await?;
codex
@@ -781,9 +767,7 @@ async fn configured_verbosity_not_sent_for_models_without_support() -> anyhow::R
skip_if_no_network!(Ok(()));
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let TestCodex { codex, .. } = test_codex()
.with_model("gpt-5-codex")
.with_config(|config| {
@@ -821,9 +805,7 @@ async fn configured_verbosity_is_sent() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let TestCodex { codex, .. } = test_codex()
.with_model("gpt-5")
.with_config(|config| {
@@ -862,9 +844,7 @@ async fn includes_developer_instructions_message_in_request() {
skip_if_no_network!();
let server = MockServer::start().await;
let resp_mock =
responses::mount_sse_once_match(&server, path("/v1/responses"), sse_completed("resp1"))
.await;
let resp_mock = responses::mount_sse_once(&server, sse_completed("resp1")).await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
codex-rs/core/tests/suite/compact.rs
+67
View File
@@ -14,6 +14,7 @@ use codex_protocol::user_input::UserInput;
use core_test_support::load_default_config_for_test;
use core_test_support::skip_if_no_network;
use core_test_support::wait_for_event;
use core_test_support::wait_for_event_match;
use std::collections::VecDeque;
use tempfile::TempDir;
@@ -366,6 +367,72 @@ async fn manual_compact_uses_custom_prompt() {
assert!(!found_default_prompt, "default prompt should be replaced");
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn manual_compact_emits_estimated_token_usage_event() {
skip_if_no_network!();
let server = start_mock_server().await;
// Compact run where the API reports zero tokens in usage. Our local
// estimator should still compute a non-zero context size for the compacted
// history.
let sse_compact = sse(vec![
ev_assistant_message("m1", SUMMARY_TEXT),
ev_completed_with_tokens("r1", 0),
]);
mount_sse_once(&server, sse_compact).await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&home);
config.model_provider = model_provider;
set_test_compact_prompt(&mut config);
let conversation_manager = ConversationManager::with_auth(CodexAuth::from_api_key("dummy"));
let NewConversation {
conversation: codex,
..
} = conversation_manager.new_conversation(config).await.unwrap();
// Trigger manual compact and collect TokenCount events for the compact turn.
codex.submit(Op::Compact).await.unwrap();
// First TokenCount: from the compact API call (usage.total_tokens = 0).
let first = wait_for_event_match(&codex, |ev| match ev {
EventMsg::TokenCount(tc) => tc
.info
.as_ref()
.map(|info| info.last_token_usage.total_tokens),
_ => None,
})
.await;
// Second TokenCount: from the local post-compaction estimate.
let last = wait_for_event_match(&codex, |ev| match ev {
EventMsg::TokenCount(tc) => tc
.info
.as_ref()
.map(|info| info.last_token_usage.total_tokens),
_ => None,
})
.await;
// Ensure the compact task itself completes.
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
assert_eq!(
first, 0,
"expected first TokenCount from compact API usage to be zero"
);
assert!(
last > 0,
"second TokenCount should reflect a non-zero estimated context size after compaction"
);
}
// Windows CI only: bump to 4 workers to prevent SSE/event starvation and test timeouts.
#[cfg_attr(windows, tokio::test(flavor = "multi_thread", worker_threads = 4))]
#[cfg_attr(not(windows), tokio::test(flavor = "multi_thread", worker_threads = 2))]
codex-rs/core/tests/suite/grep_files.rs
+21 -108
View File
@@ -2,28 +2,14 @@
use anyhow::Result;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::user_input::UserInput;
use core_test_support::responses;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::sse;
use core_test_support::responses::mount_function_call_agent_response;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use serde_json::Value;
use std::collections::HashSet;
use std::path::Path;
use std::process::Command as StdCommand;
use wiremock::matchers::any;
const MODEL_WITH_TOOL: &str = "test-gpt-5-codex";
@@ -69,18 +55,22 @@ async fn grep_files_tool_collects_matches() -> Result<()> {
})
.to_string();
mount_tool_sequence(&server, call_id, &arguments, "grep_files").await;
submit_turn(&test, "please find uses of needle").await?;
let mocks =
mount_function_call_agent_response(&server, call_id, &arguments, "grep_files").await;
test.submit_turn("please find uses of needle").await?;
let bodies = recorded_bodies(&server).await?;
let tool_output = find_tool_output(&bodies, call_id).expect("tool output present");
let payload = tool_output.get("output").expect("output field present");
let (content_opt, success_opt) = extract_content_and_success(payload);
let req = mocks.completion.single_request();
let (content_opt, success_opt) = req
.function_call_output_content_and_success(call_id)
.expect("tool output present");
let content = content_opt.expect("content present");
let success = success_opt.unwrap_or(true);
assert!(success, "expected success for matches, got {payload:?}");
assert!(
success,
"expected success for matches, got content={content}"
);
let entries = collect_file_names(content);
let entries = collect_file_names(&content);
assert_eq!(entries.len(), 2, "content: {content}");
assert!(
entries.contains("alpha.rs"),
@@ -118,16 +108,17 @@ async fn grep_files_tool_reports_empty_results() -> Result<()> {
})
.to_string();
mount_tool_sequence(&server, call_id, &arguments, "grep_files").await;
submit_turn(&test, "search again").await?;
let mocks =
mount_function_call_agent_response(&server, call_id, &arguments, "grep_files").await;
test.submit_turn("search again").await?;
let bodies = recorded_bodies(&server).await?;
let tool_output = find_tool_output(&bodies, call_id).expect("tool output present");
let payload = tool_output.get("output").expect("output field present");
let (content_opt, success_opt) = extract_content_and_success(payload);
let req = mocks.completion.single_request();
let (content_opt, success_opt) = req
.function_call_output_content_and_success(call_id)
.expect("tool output present");
let content = content_opt.expect("content present");
if let Some(success) = success_opt {
assert!(!success, "expected success=false payload: {payload:?}");
assert!(!success, "expected success=false content={content}");
}
assert_eq!(content, "No matches found.");
@@ -144,73 +135,6 @@ async fn build_test_codex(server: &wiremock::MockServer) -> Result<TestCodex> {
builder.build(server).await
}
async fn submit_turn(test: &TestCodex, prompt: &str) -> Result<()> {
let session_model = test.session_configured.model.clone();
test.codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: prompt.into(),
}],
final_output_json_schema: None,
cwd: test.cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy: SandboxPolicy::DangerFullAccess,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
.await?;
wait_for_event(&test.codex, |event| {
matches!(event, EventMsg::TaskComplete(_))
})
.await;
Ok(())
}
async fn mount_tool_sequence(
server: &wiremock::MockServer,
call_id: &str,
arguments: &str,
tool_name: &str,
) {
let first_response = sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, tool_name, arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(server, any(), first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
responses::mount_sse_once_match(server, any(), second_response).await;
}
#[allow(clippy::expect_used)]
async fn recorded_bodies(server: &wiremock::MockServer) -> Result<Vec<Value>> {
let requests = server.received_requests().await.expect("requests recorded");
Ok(requests
.iter()
.map(|req| req.body_json::<Value>().expect("request json"))
.collect())
}
fn find_tool_output<'a>(requests: &'a [Value], call_id: &str) -> Option<&'a Value> {
requests.iter().find_map(|body| {
body.get("input")
.and_then(Value::as_array)
.and_then(|items| {
items.iter().find(|item| {
item.get("type").and_then(Value::as_str) == Some("function_call_output")
&& item.get("call_id").and_then(Value::as_str) == Some(call_id)
})
})
})
}
fn collect_file_names(content: &str) -> HashSet<String> {
content
.lines()
@@ -224,14 +148,3 @@ fn collect_file_names(content: &str) -> HashSet<String> {
})
.collect()
}
fn extract_content_and_success(value: &Value) -> (Option<&str>, Option<bool>) {
match value {
Value::String(text) => (Some(text.as_str()), None),
Value::Object(obj) => (
obj.get("content").and_then(Value::as_str),
obj.get("success").and_then(Value::as_bool),
),
_ => (None, None),
}
}
codex-rs/core/tests/suite/items.rs
+8 -9
View File
@@ -18,7 +18,7 @@ use core_test_support::responses::ev_reasoning_text_delta;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::ev_web_search_call_added;
use core_test_support::responses::ev_web_search_call_done;
use core_test_support::responses::mount_sse_once_match;
use core_test_support::responses::mount_sse_once;
use core_test_support::responses::sse;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
@@ -26,7 +26,6 @@ use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event_match;
use pretty_assertions::assert_eq;
use wiremock::matchers::any;
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn user_message_item_is_emitted() -> anyhow::Result<()> {
@@ -37,7 +36,7 @@ async fn user_message_item_is_emitted() -> anyhow::Result<()> {
let TestCodex { codex, .. } = test_codex().build(&server).await?;
let first_response = sse(vec![ev_response_created("resp-1"), ev_completed("resp-1")]);
mount_sse_once_match(&server, any(), first_response).await;
mount_sse_once(&server, first_response).await;
codex
.submit(Op::UserInput {
@@ -93,7 +92,7 @@ async fn assistant_message_item_is_emitted() -> anyhow::Result<()> {
ev_assistant_message("msg-1", "all done"),
ev_completed("resp-1"),
]);
mount_sse_once_match(&server, any(), first_response).await;
mount_sse_once(&server, first_response).await;
codex
.submit(Op::UserInput {
@@ -149,7 +148,7 @@ async fn reasoning_item_is_emitted() -> anyhow::Result<()> {
reasoning_item,
ev_completed("resp-1"),
]);
mount_sse_once_match(&server, any(), first_response).await;
mount_sse_once(&server, first_response).await;
codex
.submit(Op::UserInput {
@@ -207,7 +206,7 @@ async fn web_search_item_is_emitted() -> anyhow::Result<()> {
web_search_done,
ev_completed("resp-1"),
]);
mount_sse_once_match(&server, any(), first_response).await;
mount_sse_once(&server, first_response).await;
codex
.submit(Op::UserInput {
@@ -259,7 +258,7 @@ async fn agent_message_content_delta_has_item_metadata() -> anyhow::Result<()> {
ev_assistant_message("msg-1", "streamed response"),
ev_completed("resp-1"),
]);
mount_sse_once_match(&server, any(), stream).await;
mount_sse_once(&server, stream).await;
codex
.submit(Op::UserInput {
@@ -324,7 +323,7 @@ async fn reasoning_content_delta_has_item_metadata() -> anyhow::Result<()> {
ev_reasoning_item("reasoning-1", &["step one"], &[]),
ev_completed("resp-1"),
]);
mount_sse_once_match(&server, any(), stream).await;
mount_sse_once(&server, stream).await;
codex
.submit(Op::UserInput {
@@ -381,7 +380,7 @@ async fn reasoning_raw_content_delta_respects_flag() -> anyhow::Result<()> {
ev_reasoning_item("reasoning-raw", &["complete"], &["raw detail"]),
ev_completed("resp-1"),
]);
mount_sse_once_match(&server, any(), stream).await;
mount_sse_once(&server, stream).await;
codex
.submit(Op::UserInput {
codex-rs/core/tests/suite/list_dir.rs
+46 -339
View File
@@ -1,25 +1,11 @@
#![cfg(not(target_os = "windows"))]
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::user_input::UserInput;
use core_test_support::responses;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::sse;
use core_test_support::responses::mount_function_call_agent_response;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use pretty_assertions::assert_eq;
use serde_json::Value;
use wiremock::matchers::any;
use serde_json::json;
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
#[ignore = "disabled until we enable list_dir tool"]
@@ -27,99 +13,30 @@ async fn list_dir_tool_returns_entries() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let test = test_codex().build(&server).await?;
let TestCodex {
codex,
cwd,
session_configured,
..
} = test_codex().build(&server).await?;
let dir_path = cwd.path().join("sample_dir");
let dir_path = test.cwd.path().join("sample_dir");
std::fs::create_dir(&dir_path)?;
std::fs::write(dir_path.join("alpha.txt"), "first file")?;
std::fs::create_dir(dir_path.join("nested"))?;
let dir_path = dir_path.to_string_lossy().to_string();
let call_id = "list-dir-call";
let arguments = serde_json::json!({
let arguments = json!({
"dir_path": dir_path,
"offset": 1,
"limit": 2,
})
.to_string();
let first_response = sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, "list_dir", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
responses::mount_sse_once_match(&server, any(), second_response).await;
let session_model = session_configured.model.clone();
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: "list directory contents".into(),
}],
final_output_json_schema: None,
cwd: cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy: SandboxPolicy::DangerFullAccess,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.expect("recorded requests");
let request_bodies = requests
.iter()
.map(|req| req.body_json::<Value>().unwrap())
.collect::<Vec<_>>();
assert!(
!request_bodies.is_empty(),
"expected at least one request body"
);
let tool_output_item = request_bodies
.iter()
.find_map(|body| {
body.get("input")
.and_then(Value::as_array)
.and_then(|items| {
items.iter().find(|item| {
item.get("type").and_then(Value::as_str) == Some("function_call_output")
})
})
})
.unwrap_or_else(|| {
panic!("function_call_output item not found in requests: {request_bodies:#?}")
});
assert_eq!(
tool_output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = tool_output_item
.get("output")
.and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
.expect("output text present");
assert_eq!(output_text, "E1: [file] alpha.txt\nE2: [dir] nested");
let mocks = mount_function_call_agent_response(&server, call_id, &arguments, "list_dir").await;
test.submit_turn("list directory contents").await?;
let req = mocks.completion.single_request();
let (content_opt, _) = req
.function_call_output_content_and_success(call_id)
.expect("function_call_output present");
let output = content_opt.expect("output content present in tool output");
assert_eq!(output, "E1: [file] alpha.txt\nE2: [dir] nested");
Ok(())
}
@@ -130,15 +47,9 @@ async fn list_dir_tool_depth_one_omits_children() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let test = test_codex().build(&server).await?;
let TestCodex {
codex,
cwd,
session_configured,
..
} = test_codex().build(&server).await?;
let dir_path = cwd.path().join("depth_one");
let dir_path = test.cwd.path().join("depth_one");
std::fs::create_dir(&dir_path)?;
std::fs::write(dir_path.join("alpha.txt"), "alpha")?;
std::fs::create_dir(dir_path.join("nested"))?;
@@ -146,7 +57,7 @@ async fn list_dir_tool_depth_one_omits_children() -> anyhow::Result<()> {
let dir_path = dir_path.to_string_lossy().to_string();
let call_id = "list-dir-depth1";
let arguments = serde_json::json!({
let arguments = json!({
"dir_path": dir_path,
"offset": 1,
"limit": 10,
@@ -154,77 +65,15 @@ async fn list_dir_tool_depth_one_omits_children() -> anyhow::Result<()> {
})
.to_string();
let first_response = sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, "list_dir", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
responses::mount_sse_once_match(&server, any(), second_response).await;
let session_model = session_configured.model.clone();
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: "list directory contents depth one".into(),
}],
final_output_json_schema: None,
cwd: cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy: SandboxPolicy::DangerFullAccess,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
let mocks = mount_function_call_agent_response(&server, call_id, &arguments, "list_dir").await;
test.submit_turn("list directory contents depth one")
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.expect("recorded requests");
let request_bodies = requests
.iter()
.map(|req| req.body_json::<Value>().unwrap())
.collect::<Vec<_>>();
assert!(
!request_bodies.is_empty(),
"expected at least one request body"
);
let tool_output_item = request_bodies
.iter()
.find_map(|body| {
body.get("input")
.and_then(Value::as_array)
.and_then(|items| {
items.iter().find(|item| {
item.get("type").and_then(Value::as_str) == Some("function_call_output")
})
})
})
.unwrap_or_else(|| {
panic!("function_call_output item not found in requests: {request_bodies:#?}")
});
assert_eq!(
tool_output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = tool_output_item
.get("output")
.and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
.expect("output text present");
assert_eq!(output_text, "E1: [file] alpha.txt\nE2: [dir] nested");
let req = mocks.completion.single_request();
let (content_opt, _) = req
.function_call_output_content_and_success(call_id)
.expect("function_call_output present");
let output = content_opt.expect("output content present in tool output");
assert_eq!(output, "E1: [file] alpha.txt\nE2: [dir] nested");
Ok(())
}
@@ -235,15 +84,9 @@ async fn list_dir_tool_depth_two_includes_children_only() -> anyhow::Result<()>
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let test = test_codex().build(&server).await?;
let TestCodex {
codex,
cwd,
session_configured,
..
} = test_codex().build(&server).await?;
let dir_path = cwd.path().join("depth_two");
let dir_path = test.cwd.path().join("depth_two");
std::fs::create_dir(&dir_path)?;
std::fs::write(dir_path.join("alpha.txt"), "alpha")?;
let nested = dir_path.join("nested");
@@ -255,7 +98,7 @@ async fn list_dir_tool_depth_two_includes_children_only() -> anyhow::Result<()>
let dir_path_string = dir_path.to_string_lossy().to_string();
let call_id = "list-dir-depth2";
let arguments = serde_json::json!({
let arguments = json!({
"dir_path": dir_path_string,
"offset": 1,
"limit": 10,
@@ -263,81 +106,16 @@ async fn list_dir_tool_depth_two_includes_children_only() -> anyhow::Result<()>
})
.to_string();
let first_response = sse(vec![
serde_json::json!({
"type": "response.created",
"response": {"id": "resp-1"}
}),
ev_function_call(call_id, "list_dir", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
responses::mount_sse_once_match(&server, any(), second_response).await;
let session_model = session_configured.model.clone();
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: "list directory contents depth two".into(),
}],
final_output_json_schema: None,
cwd: cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy: SandboxPolicy::DangerFullAccess,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
let mocks = mount_function_call_agent_response(&server, call_id, &arguments, "list_dir").await;
test.submit_turn("list directory contents depth two")
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.expect("recorded requests");
let request_bodies = requests
.iter()
.map(|req| req.body_json::<Value>().unwrap())
.collect::<Vec<_>>();
assert!(
!request_bodies.is_empty(),
"expected at least one request body"
);
let tool_output_item = request_bodies
.iter()
.find_map(|body| {
body.get("input")
.and_then(Value::as_array)
.and_then(|items| {
items.iter().find(|item| {
item.get("type").and_then(Value::as_str) == Some("function_call_output")
})
})
})
.unwrap_or_else(|| {
panic!("function_call_output item not found in requests: {request_bodies:#?}")
});
let req = mocks.completion.single_request();
let (content_opt, _) = req
.function_call_output_content_and_success(call_id)
.expect("function_call_output present");
let output = content_opt.expect("output content present in tool output");
assert_eq!(
tool_output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = tool_output_item
.get("output")
.and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
.expect("output text present");
assert_eq!(
output_text,
output,
"E1: [file] alpha.txt\nE2: [dir] nested\nE3: [file] nested/beta.txt\nE4: [dir] nested/grand"
);
@@ -350,15 +128,9 @@ async fn list_dir_tool_depth_three_includes_grandchildren() -> anyhow::Result<()
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let test = test_codex().build(&server).await?;
let TestCodex {
codex,
cwd,
session_configured,
..
} = test_codex().build(&server).await?;
let dir_path = cwd.path().join("depth_three");
let dir_path = test.cwd.path().join("depth_three");
std::fs::create_dir(&dir_path)?;
std::fs::write(dir_path.join("alpha.txt"), "alpha")?;
let nested = dir_path.join("nested");
@@ -370,7 +142,7 @@ async fn list_dir_tool_depth_three_includes_grandchildren() -> anyhow::Result<()
let dir_path_string = dir_path.to_string_lossy().to_string();
let call_id = "list-dir-depth3";
let arguments = serde_json::json!({
let arguments = json!({
"dir_path": dir_path_string,
"offset": 1,
"limit": 10,
@@ -378,81 +150,16 @@ async fn list_dir_tool_depth_three_includes_grandchildren() -> anyhow::Result<()
})
.to_string();
let first_response = sse(vec![
serde_json::json!({
"type": "response.created",
"response": {"id": "resp-1"}
}),
ev_function_call(call_id, "list_dir", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
responses::mount_sse_once_match(&server, any(), second_response).await;
let session_model = session_configured.model.clone();
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: "list directory contents depth three".into(),
}],
final_output_json_schema: None,
cwd: cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy: SandboxPolicy::DangerFullAccess,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
let mocks = mount_function_call_agent_response(&server, call_id, &arguments, "list_dir").await;
test.submit_turn("list directory contents depth three")
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.expect("recorded requests");
let request_bodies = requests
.iter()
.map(|req| req.body_json::<Value>().unwrap())
.collect::<Vec<_>>();
assert!(
!request_bodies.is_empty(),
"expected at least one request body"
);
let tool_output_item = request_bodies
.iter()
.find_map(|body| {
body.get("input")
.and_then(Value::as_array)
.and_then(|items| {
items.iter().find(|item| {
item.get("type").and_then(Value::as_str) == Some("function_call_output")
})
})
})
.unwrap_or_else(|| {
panic!("function_call_output item not found in requests: {request_bodies:#?}")
});
let req = mocks.completion.single_request();
let (content_opt, _) = req
.function_call_output_content_and_success(call_id)
.expect("function_call_output present");
let output = content_opt.expect("output content present in tool output");
assert_eq!(
tool_output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = tool_output_item
.get("output")
.and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
.expect("output text present");
assert_eq!(
output_text,
output,
"E1: [file] alpha.txt\nE2: [dir] nested\nE3: [file] nested/beta.txt\nE4: [dir] nested/grand\nE5: [file] nested/grand/gamma.txt"
);
codex-rs/core/tests/suite/mod.rs
-2
View File
@@ -18,8 +18,6 @@ mod abort_tasks;
#[cfg(not(target_os = "windows"))]
mod apply_patch_cli;
#[cfg(not(target_os = "windows"))]
mod apply_patch_freeform;
#[cfg(not(target_os = "windows"))]
mod approvals;
mod auth_refresh;
mod cli_stream;
codex-rs/core/tests/suite/model_tools.rs
+1 -1
View File
@@ -42,7 +42,7 @@ async fn collect_tool_identifiers_for_model(model: &str) -> Vec<String> {
let server = MockServer::start().await;
let sse = sse_completed(model);
let resp_mock = responses::mount_sse_once_match(&server, wiremock::matchers::any(), sse).await;
let resp_mock = responses::mount_sse_once(&server, sse).await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
codex-rs/core/tests/suite/prompt_caching.rs
+135 -348
View File
@@ -1,9 +1,5 @@
#![allow(clippy::unwrap_used)]
use codex_core::CodexAuth;
use codex_core::ConversationManager;
use codex_core::ModelProviderInfo;
use codex_core::built_in_model_providers;
use codex_core::config::OPENAI_DEFAULT_MODEL;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
@@ -16,17 +12,15 @@ use codex_core::protocol_config_types::ReasoningSummary;
use codex_core::shell::Shell;
use codex_core::shell::default_user_shell;
use codex_protocol::user_input::UserInput;
use core_test_support::load_default_config_for_test;
use core_test_support::load_sse_fixture_with_id;
use core_test_support::responses::mount_sse_once;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use std::collections::HashMap;
use tempfile::TempDir;
use wiremock::Mock;
use wiremock::MockServer;
use wiremock::ResponseTemplate;
use wiremock::matchers::method;
use wiremock::matchers::path;
fn text_user_input(text: String) -> serde_json::Value {
serde_json::json!({
@@ -70,48 +64,24 @@ fn assert_tool_names(body: &serde_json::Value, expected_names: &[&str]) {
}
#[tokio::test(flavor = "multi_thread", worker_threads = 4)]
async fn codex_mini_latest_tools() {
skip_if_no_network!();
async fn codex_mini_latest_tools() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
// Expect two POSTs to /v1/responses
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
config.features.disable(Feature::ApplyPatchFreeform);
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
config.model = "codex-mini-latest".to_string();
config.model_family = find_family_for_model("codex-mini-latest").unwrap();
let codex = conversation_manager
.new_conversation(config)
.await
.expect("create new conversation")
.conversation;
let TestCodex { codex, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
config.features.disable(Feature::ApplyPatchFreeform);
config.model = "codex-mini-latest".to_string();
config.model_family = find_family_for_model("codex-mini-latest")
.expect("model family for codex-mini-latest");
})
.build(&server)
.await?;
codex
.submit(Op::UserInput {
@@ -119,8 +89,7 @@ async fn codex_mini_latest_tools() {
text: "hello 1".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
@@ -129,72 +98,45 @@ async fn codex_mini_latest_tools() {
text: "hello 2".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
let expected_instructions = [
include_str!("../../prompt.md"),
include_str!("../../../apply-patch/apply_patch_tool_instructions.md"),
]
.join("\n");
let body0 = requests[0].body_json::<serde_json::Value>().unwrap();
let body0 = req1.single_request().body_json();
assert_eq!(
body0["instructions"],
serde_json::json!(expected_instructions),
);
let body1 = requests[1].body_json::<serde_json::Value>().unwrap();
let body1 = req2.single_request().body_json();
assert_eq!(
body1["instructions"],
serde_json::json!(expected_instructions),
);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 4)]
async fn prompt_tools_are_consistent_across_requests() {
skip_if_no_network!();
async fn prompt_tools_are_consistent_across_requests() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
// Expect two POSTs to /v1/responses
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
let TestCodex { codex, config, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
})
.build(&server)
.await?;
let base_instructions = config.model_family.base_instructions.clone();
let codex = conversation_manager
.new_conversation(config)
.await
.expect("create new conversation")
.conversation;
codex
.submit(Op::UserInput {
@@ -202,8 +144,7 @@ async fn prompt_tools_are_consistent_across_requests() {
text: "hello 1".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
@@ -212,13 +153,9 @@ async fn prompt_tools_are_consistent_across_requests() {
text: "hello 2".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
// our internal implementation is responsible for keeping tools in sync
// with the OpenAI schema, so we just verify the tool presence here
let tools_by_model: HashMap<&'static str, Vec<&'static str>> = HashMap::from([
@@ -274,7 +211,7 @@ async fn prompt_tools_are_consistent_across_requests() {
.get(OPENAI_DEFAULT_MODEL)
.unwrap_or_else(|| panic!("expected tools to be defined for model {OPENAI_DEFAULT_MODEL}"))
.as_slice();
let body0 = requests[0].body_json::<serde_json::Value>().unwrap();
let body0 = req1.single_request().body_json();
let expected_instructions = if expected_tools_names.contains(&"apply_patch") {
base_instructions
@@ -292,54 +229,32 @@ async fn prompt_tools_are_consistent_across_requests() {
);
assert_tool_names(&body0, expected_tools_names);
let body1 = requests[1].body_json::<serde_json::Value>().unwrap();
let body1 = req2.single_request().body_json();
assert_eq!(
body1["instructions"],
serde_json::json!(expected_instructions),
);
assert_tool_names(&body1, expected_tools_names);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
#[ignore = "flaky on ubuntu-24.04-arm - aarch64-unknown-linux-gpu"]
async fn prefixes_context_and_instructions_once_and_consistently_across_requests() {
skip_if_no_network!();
async fn prefixes_context_and_instructions_once_and_consistently_across_requests()
-> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
// Expect two POSTs to /v1/responses
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
let codex = conversation_manager
.new_conversation(config)
.await
.expect("create new conversation")
.conversation;
let TestCodex { codex, config, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
})
.build(&server)
.await?;
codex
.submit(Op::UserInput {
@@ -347,8 +262,7 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests
text: "hello 1".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
@@ -357,31 +271,14 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests
text: "hello 2".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
let shell = default_user_shell().await;
let expected_env_text = format!(
r#"<environment_context>
<cwd>{}</cwd>
<approval_policy>on-request</approval_policy>
<sandbox_mode>read-only</sandbox_mode>
<network_access>restricted</network_access>
{}</environment_context>"#,
cwd.path().to_string_lossy(),
match shell.name() {
Some(name) => format!(" <shell>{name}</shell>\n"),
None => String::new(),
}
);
let cwd_str = config.cwd.to_string_lossy();
let expected_env_text = default_env_context_str(&cwd_str, &shell);
let expected_ui_text = format!(
"# AGENTS.md instructions for {}\n\n<INSTRUCTIONS>\nbe consistent and helpful\n</INSTRUCTIONS>",
cwd.path().to_string_lossy()
"# AGENTS.md instructions for {cwd_str}\n\n<INSTRUCTIONS>\nbe consistent and helpful\n</INSTRUCTIONS>"
);
let expected_env_msg = serde_json::json!({
@@ -400,7 +297,7 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests
"role": "user",
"content": [ { "type": "input_text", "text": "hello 1" } ]
});
let body1 = requests[0].body_json::<serde_json::Value>().unwrap();
let body1 = req1.single_request().body_json();
assert_eq!(
body1["input"],
serde_json::json!([expected_ui_msg, expected_env_msg, expected_user_message_1])
@@ -411,7 +308,7 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests
"role": "user",
"content": [ { "type": "input_text", "text": "hello 2" } ]
});
let body2 = requests[1].body_json::<serde_json::Value>().unwrap();
let body2 = req2.single_request().body_json();
let expected_body2 = serde_json::json!(
[
body1["input"].as_array().unwrap().as_slice(),
@@ -420,47 +317,25 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests
.concat()
);
assert_eq!(body2["input"], expected_body2);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
skip_if_no_network!();
async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
// Expect two POSTs to /v1/responses
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
let codex = conversation_manager
.new_conversation(config)
.await
.expect("create new conversation")
.conversation;
let TestCodex { codex, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
})
.build(&server)
.await?;
// First turn
codex
@@ -469,8 +344,7 @@ async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
text: "hello 1".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let writable = TempDir::new().unwrap();
@@ -488,8 +362,7 @@ async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
effort: Some(Some(ReasoningEffort::High)),
summary: Some(ReasoningSummary::Detailed),
})
.await
.unwrap();
.await?;
// Second turn after overrides
codex
@@ -498,16 +371,11 @@ async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
text: "hello 2".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
// Verify we issued exactly two requests, and the cached prefix stayed identical.
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
let body1 = requests[0].body_json::<serde_json::Value>().unwrap();
let body2 = requests[1].body_json::<serde_json::Value>().unwrap();
let body1 = req1.single_request().body_json();
let body2 = req2.single_request().body_json();
// prompt_cache_key should remain constant across overrides
assert_eq!(
body1["prompt_cache_key"], body2["prompt_cache_key"],
@@ -548,47 +416,25 @@ async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
.concat()
);
assert_eq!(body2["input"], expected_body2);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn per_turn_overrides_keep_cached_prefix_and_key_constant() {
skip_if_no_network!();
async fn per_turn_overrides_keep_cached_prefix_and_key_constant() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
// Expect two POSTs to /v1/responses
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
let codex = conversation_manager
.new_conversation(config)
.await
.expect("create new conversation")
.conversation;
let TestCodex { codex, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
})
.build(&server)
.await?;
// First turn
codex
@@ -597,8 +443,7 @@ async fn per_turn_overrides_keep_cached_prefix_and_key_constant() {
text: "hello 1".into(),
}],
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
// Second turn using per-turn overrides via UserTurn
@@ -622,16 +467,11 @@ async fn per_turn_overrides_keep_cached_prefix_and_key_constant() {
summary: ReasoningSummary::Detailed,
final_output_json_schema: None,
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
// Verify we issued exactly two requests, and the cached prefix stayed identical.
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
let body1 = requests[0].body_json::<serde_json::Value>().unwrap();
let body2 = requests[1].body_json::<serde_json::Value>().unwrap();
let body1 = req1.single_request().body_json();
let body2 = req2.single_request().body_json();
// prompt_cache_key should remain constant across per-turn overrides
assert_eq!(
@@ -672,38 +512,25 @@ async fn per_turn_overrides_keep_cached_prefix_and_key_constant() {
.concat()
);
assert_eq!(body2["input"], expected_body2);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn send_user_turn_with_no_changes_does_not_send_environment_context() {
skip_if_no_network!();
async fn send_user_turn_with_no_changes_does_not_send_environment_context() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
let TestCodex { codex, config, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
})
.build(&server)
.await?;
let default_cwd = config.cwd.clone();
let default_approval_policy = config.approval_policy;
@@ -712,14 +539,6 @@ async fn send_user_turn_with_no_changes_does_not_send_environment_context() {
let default_effort = config.model_reasoning_effort;
let default_summary = config.model_reasoning_summary;
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
let codex = conversation_manager
.new_conversation(config)
.await
.expect("create new conversation")
.conversation;
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
@@ -733,8 +552,7 @@ async fn send_user_turn_with_no_changes_does_not_send_environment_context() {
summary: default_summary,
final_output_json_schema: None,
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
@@ -750,27 +568,20 @@ async fn send_user_turn_with_no_changes_does_not_send_environment_context() {
summary: default_summary,
final_output_json_schema: None,
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
let body1 = requests[0].body_json::<serde_json::Value>().unwrap();
let body2 = requests[1].body_json::<serde_json::Value>().unwrap();
let body1 = req1.single_request().body_json();
let body2 = req2.single_request().body_json();
let shell = default_user_shell().await;
let default_cwd_lossy = default_cwd.to_string_lossy();
let expected_ui_text = format!(
"# AGENTS.md instructions for {}\n\n<INSTRUCTIONS>\nbe consistent and helpful\n</INSTRUCTIONS>",
default_cwd.to_string_lossy()
"# AGENTS.md instructions for {default_cwd_lossy}\n\n<INSTRUCTIONS>\nbe consistent and helpful\n</INSTRUCTIONS>"
);
let expected_ui_msg = text_user_input(expected_ui_text);
let expected_env_msg_1 = text_user_input(default_env_context_str(
&cwd.path().to_string_lossy(),
&shell,
));
let expected_env_msg_1 = text_user_input(default_env_context_str(&default_cwd_lossy, &shell));
let expected_user_message_1 = text_user_input("hello 1".to_string());
let expected_input_1 = serde_json::Value::Array(vec![
@@ -788,38 +599,25 @@ async fn send_user_turn_with_no_changes_does_not_send_environment_context() {
expected_user_message_2,
]);
assert_eq!(body2["input"], expected_input_2);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn send_user_turn_with_changes_sends_environment_context() {
skip_if_no_network!();
async fn send_user_turn_with_changes_sends_environment_context() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
use pretty_assertions::assert_eq;
let server = MockServer::start().await;
let server = start_mock_server().await;
let sse = sse_completed("resp");
let template = ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(sse, "text/event-stream");
Mock::given(method("POST"))
.and(path("/v1/responses"))
.respond_with(template)
.expect(2)
.mount(&server)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let cwd = TempDir::new().unwrap();
let codex_home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&codex_home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
let req1 = mount_sse_once(&server, sse_completed("resp-1")).await;
let req2 = mount_sse_once(&server, sse_completed("resp-2")).await;
let TestCodex { codex, config, .. } = test_codex()
.with_config(|config| {
config.user_instructions = Some("be consistent and helpful".to_string());
})
.build(&server)
.await?;
let default_cwd = config.cwd.clone();
let default_approval_policy = config.approval_policy;
@@ -828,14 +626,6 @@ async fn send_user_turn_with_changes_sends_environment_context() {
let default_effort = config.model_reasoning_effort;
let default_summary = config.model_reasoning_summary;
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
let codex = conversation_manager
.new_conversation(config.clone())
.await
.expect("create new conversation")
.conversation;
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
@@ -849,8 +639,7 @@ async fn send_user_turn_with_changes_sends_environment_context() {
summary: default_summary,
final_output_json_schema: None,
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
@@ -866,15 +655,11 @@ async fn send_user_turn_with_changes_sends_environment_context() {
summary: ReasoningSummary::Detailed,
final_output_json_schema: None,
})
.await
.unwrap();
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = server.received_requests().await.unwrap();
assert_eq!(requests.len(), 2, "expected two POST requests");
let body1 = requests[0].body_json::<serde_json::Value>().unwrap();
let body2 = requests[1].body_json::<serde_json::Value>().unwrap();
let body1 = req1.single_request().body_json();
let body2 = req2.single_request().body_json();
let shell = default_user_shell().await;
let expected_ui_text = format!(
@@ -913,4 +698,6 @@ async fn send_user_turn_with_changes_sends_environment_context() {
expected_user_message_2,
]);
assert_eq!(body2["input"], expected_input_2);
Ok(())
}
codex-rs/core/tests/suite/read_file.rs
+12 -69
View File
@@ -1,25 +1,11 @@
#![cfg(not(target_os = "windows"))]
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::user_input::UserInput;
use core_test_support::responses;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::sse;
use core_test_support::responses::mount_function_call_agent_response;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use pretty_assertions::assert_eq;
use serde_json::Value;
use wiremock::matchers::any;
use serde_json::json;
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
#[ignore = "disabled until we enable read_file tool"]
@@ -27,72 +13,29 @@ async fn read_file_tool_returns_requested_lines() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let test = test_codex().build(&server).await?;
let TestCodex {
codex,
cwd,
session_configured,
..
} = test_codex().build(&server).await?;
let file_path = cwd.path().join("sample.txt");
let file_path = test.cwd.path().join("sample.txt");
std::fs::write(&file_path, "first\nsecond\nthird\nfourth\n")?;
let file_path = file_path.to_string_lossy().to_string();
let call_id = "read-file-call";
let arguments = serde_json::json!({
let arguments = json!({
"file_path": file_path,
"offset": 2,
"limit": 2,
})
.to_string();
let first_response = sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, "read_file", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
let mocks = mount_function_call_agent_response(&server, call_id, &arguments, "read_file").await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
let second_mock = responses::mount_sse_once_match(&server, any(), second_response).await;
test.submit_turn("please inspect sample.txt").await?;
let session_model = session_configured.model.clone();
codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: "please inspect sample.txt".into(),
}],
final_output_json_schema: None,
cwd: cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy: SandboxPolicy::DangerFullAccess,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
.await?;
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let req = second_mock.single_request();
let tool_output_item = req.function_call_output(call_id);
assert_eq!(
tool_output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = tool_output_item
.get("output")
.and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
.expect("output text present");
let req = mocks.completion.single_request();
let (output_text_opt, _) = req
.function_call_output_content_and_success(call_id)
.expect("output present");
let output_text = output_text_opt.expect("output text present");
assert_eq!(output_text, "L2: second\nL3: third");
Ok(())
codex-rs/core/tests/suite/resume.rs
+3 -4
View File
@@ -6,14 +6,13 @@ use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_reasoning_item;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::mount_sse_once_match;
use core_test_support::responses::mount_sse_once;
use core_test_support::responses::sse;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use std::sync::Arc;
use wiremock::matchers::any;
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn resume_includes_initial_messages_from_rollout_events() -> Result<()> {
@@ -31,7 +30,7 @@ async fn resume_includes_initial_messages_from_rollout_events() -> Result<()> {
ev_assistant_message("msg-1", "Completed first turn"),
ev_completed("resp-initial"),
]);
mount_sse_once_match(&server, any(), initial_sse).await;
mount_sse_once(&server, initial_sse).await;
codex
.submit(Op::UserInput {
@@ -83,7 +82,7 @@ async fn resume_includes_initial_messages_from_reasoning_events() -> Result<()>
ev_assistant_message("msg-1", "Completed reasoning turn"),
ev_completed("resp-initial"),
]);
mount_sse_once_match(&server, any(), initial_sse).await;
mount_sse_once(&server, initial_sse).await;
codex
.submit(Op::UserInput {
codex-rs/core/tests/suite/rmcp_client.rs
+11 -22
View File
@@ -21,7 +21,7 @@ use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::user_input::UserInput;
use core_test_support::responses;
use core_test_support::responses::mount_sse_once_match;
use core_test_support::responses::mount_sse_once;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
@@ -36,7 +36,6 @@ use tokio::process::Child;
use tokio::process::Command;
use tokio::time::Instant;
use tokio::time::sleep;
use wiremock::matchers::any;
static OPENAI_PNG: &str = "data:image/png;base64,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";
@@ -51,9 +50,8 @@ async fn stdio_server_round_trip() -> anyhow::Result<()> {
let server_name = "rmcp";
let tool_name = format!("mcp__{server_name}__echo");
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, &tool_name, "{\"message\":\"ping\"}"),
@@ -61,9 +59,8 @@ async fn stdio_server_round_trip() -> anyhow::Result<()> {
]),
)
.await;
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_assistant_message("msg-1", "rmcp echo tool completed successfully."),
responses::ev_completed("resp-2"),
@@ -190,9 +187,8 @@ async fn stdio_image_responses_round_trip() -> anyhow::Result<()> {
let tool_name = format!("mcp__{server_name}__image");
// First stream: model decides to call the image tool.
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, &tool_name, "{}"),
@@ -201,9 +197,8 @@ async fn stdio_image_responses_round_trip() -> anyhow::Result<()> {
)
.await;
// Second stream: after tool execution, assistant emits a message and completes.
let final_mock = mount_sse_once_match(
let final_mock = mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_assistant_message("msg-1", "rmcp image tool completed successfully."),
responses::ev_completed("resp-2"),
@@ -531,9 +526,8 @@ async fn stdio_server_propagates_whitelisted_env_vars() -> anyhow::Result<()> {
let server_name = "rmcp_whitelist";
let tool_name = format!("mcp__{server_name}__echo");
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, &tool_name, "{\"message\":\"ping\"}"),
@@ -541,9 +535,8 @@ async fn stdio_server_propagates_whitelisted_env_vars() -> anyhow::Result<()> {
]),
)
.await;
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_assistant_message("msg-1", "rmcp echo tool completed successfully."),
responses::ev_completed("resp-2"),
@@ -666,9 +659,8 @@ async fn streamable_http_tool_call_round_trip() -> anyhow::Result<()> {
let server_name = "rmcp_http";
let tool_name = format!("mcp__{server_name}__echo");
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, &tool_name, "{\"message\":\"ping\"}"),
@@ -676,9 +668,8 @@ async fn streamable_http_tool_call_round_trip() -> anyhow::Result<()> {
]),
)
.await;
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_assistant_message(
"msg-1",
@@ -834,9 +825,8 @@ async fn streamable_http_with_oauth_round_trip() -> anyhow::Result<()> {
let server_name = "rmcp_http_oauth";
let tool_name = format!("mcp__{server_name}__echo");
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, &tool_name, "{\"message\":\"ping\"}"),
@@ -844,9 +834,8 @@ async fn streamable_http_with_oauth_round_trip() -> anyhow::Result<()> {
]),
)
.await;
mount_sse_once_match(
mount_sse_once(
&server,
any(),
responses::sse(vec![
responses::ev_assistant_message(
"msg-1",
codex-rs/core/tests/suite/shell_serialization.rs
+72 -197
View File
@@ -3,12 +3,7 @@
use anyhow::Result;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::user_input::UserInput;
use core_test_support::assert_regex_match;
use core_test_support::responses::ev_apply_patch_function_call;
use core_test_support::responses::ev_assistant_message;
@@ -21,9 +16,7 @@ use core_test_support::responses::mount_sse_sequence;
use core_test_support::responses::sse;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use pretty_assertions::assert_eq;
use regex_lite::Regex;
use serde_json::Value;
@@ -42,69 +35,6 @@ const FIXTURE_JSON: &str = r#"{
}
"#;
async fn submit_turn(test: &TestCodex, prompt: &str, sandbox_policy: SandboxPolicy) -> Result<()> {
let session_model = test.session_configured.model.clone();
test.codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: prompt.into(),
}],
final_output_json_schema: None,
cwd: test.cwd.path().to_path_buf(),
approval_policy: AskForApproval::Never,
sandbox_policy,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
.await?;
wait_for_event(&test.codex, |event| {
matches!(event, EventMsg::TaskComplete(_))
})
.await;
Ok(())
}
fn request_bodies(requests: &[wiremock::Request]) -> Result<Vec<Value>> {
requests
.iter()
.map(|req| Ok(serde_json::from_slice::<Value>(&req.body)?))
.collect()
}
fn find_function_call_output<'a>(bodies: &'a [Value], call_id: &str) -> Option<&'a Value> {
for body in bodies {
if let Some(items) = body.get("input").and_then(Value::as_array) {
for item in items {
if item.get("type").and_then(Value::as_str) == Some("function_call_output")
&& item.get("call_id").and_then(Value::as_str) == Some(call_id)
{
return Some(item);
}
}
}
}
None
}
fn find_custom_tool_call_output<'a>(bodies: &'a [Value], call_id: &str) -> Option<&'a Value> {
for body in bodies {
if let Some(items) = body.get("input").and_then(Value::as_array) {
for item in items {
if item.get("type").and_then(Value::as_str) == Some("custom_tool_call_output")
&& item.get("call_id").and_then(Value::as_str) == Some(call_id)
{
return Some(item);
}
}
}
}
None
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn shell_output_stays_json_without_freeform_apply_patch() -> Result<()> {
skip_if_no_network!(Ok(()));
@@ -133,21 +63,16 @@ async fn shell_output_stays_json_without_freeform_apply_patch() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"run the json shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item = find_function_call_output(&bodies, call_id).expect("shell output present");
let req = mock.last_request().expect("shell output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -202,22 +127,18 @@ async fn shell_output_is_structured_with_freeform_apply_patch() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"run the structured shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("structured output present");
let req = mock
.last_request()
.expect("structured shell output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -269,21 +190,16 @@ async fn shell_output_preserves_fixture_json_without_serialization() -> Result<(
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"read the fixture JSON with sed",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item = find_function_call_output(&bodies, call_id).expect("shell output present");
let req = mock.last_request().expect("shell output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -345,22 +261,18 @@ async fn shell_output_structures_fixture_with_serialization() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"read the fixture JSON with structured output",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("structured output present");
let req = mock
.last_request()
.expect("structured output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -420,22 +332,18 @@ async fn shell_output_for_freeform_tool_records_duration() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"run the structured shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("structured output present");
let req = mock
.last_request()
.expect("structured output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -490,22 +398,18 @@ async fn shell_output_reserializes_truncated_content() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"run the truncation shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("truncated output present");
let req = mock
.last_request()
.expect("truncated output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -570,22 +474,18 @@ async fn apply_patch_custom_tool_output_is_structured() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"apply the patch via custom tool",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let req = mock
.last_request()
.expect("apply_patch output request recorded");
let output_item = req.custom_tool_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -630,22 +530,18 @@ async fn apply_patch_custom_tool_call_creates_file() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"apply the patch via custom tool to create a file",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let req = mock
.last_request()
.expect("apply_patch output request recorded");
let output_item = req.custom_tool_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -699,22 +595,18 @@ async fn apply_patch_custom_tool_call_updates_existing_file() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"apply the patch via custom tool to update a file",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let req = mock
.last_request()
.expect("apply_patch output request recorded");
let output_item = req.custom_tool_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -762,22 +654,18 @@ async fn apply_patch_custom_tool_call_reports_failure_output() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"attempt a failing apply_patch via custom tool",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let req = mock
.last_request()
.expect("apply_patch output request recorded");
let output_item = req.custom_tool_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -817,22 +705,18 @@ async fn apply_patch_function_call_output_is_structured() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"apply the patch via function-call apply_patch",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("apply_patch function output present");
let req = mock
.last_request()
.expect("apply_patch function output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -880,21 +764,16 @@ async fn shell_output_is_structured_for_nonzero_exit() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"run the failing shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item = find_function_call_output(&bodies, call_id).expect("shell output present");
let req = mock.last_request().expect("shell output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
@@ -934,22 +813,18 @@ async fn local_shell_call_output_is_structured() -> Result<()> {
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
let mock = mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
test.submit_turn_with_policy(
"run the local shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("local shell output present");
let req = mock
.last_request()
.expect("local shell output request recorded");
let output_item = req.function_call_output(call_id);
let output = output_item
.get("output")
.and_then(Value::as_str)
codex-rs/core/tests/suite/tool_harness.rs
+36 -57
View File
@@ -14,6 +14,7 @@ use codex_protocol::plan_tool::StepStatus;
use codex_protocol::user_input::UserInput;
use core_test_support::assert_regex_match;
use core_test_support::responses;
use core_test_support::responses::ResponsesRequest;
use core_test_support::responses::ev_apply_patch_function_call;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
@@ -28,14 +29,23 @@ use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use serde_json::Value;
use serde_json::json;
use wiremock::matchers::any;
fn extract_output_text(item: &Value) -> Option<&str> {
item.get("output").and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
fn call_output(req: &ResponsesRequest, call_id: &str) -> (String, Option<bool>) {
let raw = req.function_call_output(call_id);
assert_eq!(
raw.get("call_id").and_then(Value::as_str),
Some(call_id),
"mismatched call_id in function_call_output"
);
let (content_opt, success) = match req.function_call_output_content_and_success(call_id) {
Some(values) => values,
None => panic!("function_call_output present"),
};
let content = match content_opt {
Some(c) => c,
None => panic!("function_call_output content present"),
};
(content, success)
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
@@ -62,13 +72,13 @@ async fn shell_tool_executes_command_and_streams_output() -> anyhow::Result<()>
ev_local_shell_call(call_id, "completed", command),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "all done"),
ev_completed("resp-2"),
]);
let second_mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let second_mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -90,9 +100,8 @@ async fn shell_tool_executes_command_and_streams_output() -> anyhow::Result<()>
wait_for_event(&codex, |event| matches!(event, EventMsg::TaskComplete(_))).await;
let req = second_mock.single_request();
let output_item = req.function_call_output(call_id);
let output_text = extract_output_text(&output_item).expect("output text present");
let exec_output: Value = serde_json::from_str(output_text)?;
let (output_text, _) = call_output(&req, call_id);
let exec_output: Value = serde_json::from_str(&output_text)?;
assert_eq!(exec_output["metadata"]["exit_code"], 0);
let stdout = exec_output["output"].as_str().expect("stdout field");
assert_regex_match(r"(?s)^tool harness\n?$", stdout);
@@ -129,13 +138,13 @@ async fn update_plan_tool_emits_plan_update_event() -> anyhow::Result<()> {
ev_function_call(call_id, "update_plan", &plan_args),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "plan acknowledged"),
ev_completed("resp-2"),
]);
let second_mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let second_mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -174,12 +183,7 @@ async fn update_plan_tool_emits_plan_update_event() -> anyhow::Result<()> {
assert!(saw_plan_update, "expected PlanUpdate event");
let req = second_mock.single_request();
let output_item = req.function_call_output(call_id);
assert_eq!(
output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = extract_output_text(&output_item).expect("output text present");
let (output_text, _success_flag) = call_output(&req, call_id);
assert_eq!(output_text, "Plan updated");
Ok(())
@@ -210,13 +214,13 @@ async fn update_plan_tool_rejects_malformed_payload() -> anyhow::Result<()> {
ev_function_call(call_id, "update_plan", &invalid_args),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "malformed plan payload"),
ev_completed("resp-2"),
]);
let second_mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let second_mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -252,22 +256,12 @@ async fn update_plan_tool_rejects_malformed_payload() -> anyhow::Result<()> {
);
let req = second_mock.single_request();
let output_item = req.function_call_output(call_id);
assert_eq!(
output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = extract_output_text(&output_item).expect("output text present");
let (output_text, success_flag) = call_output(&req, call_id);
assert!(
output_text.contains("failed to parse function arguments"),
"expected parse error message in output text, got {output_text:?}"
);
if let Some(success_flag) = output_item
.get("output")
.and_then(|value| value.as_object())
.and_then(|obj| obj.get("success"))
.and_then(serde_json::Value::as_bool)
{
if let Some(success_flag) = success_flag {
assert!(
!success_flag,
"expected tool output to mark success=false for malformed payload"
@@ -308,13 +302,13 @@ async fn apply_patch_tool_executes_and_emits_patch_events() -> anyhow::Result<()
ev_apply_patch_function_call(call_id, &patch_content),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "patch complete"),
ev_completed("resp-2"),
]);
let second_mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let second_mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -357,12 +351,7 @@ async fn apply_patch_tool_executes_and_emits_patch_events() -> anyhow::Result<()
assert!(patch_end_success);
let req = second_mock.single_request();
let output_item = req.function_call_output(call_id);
assert_eq!(
output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = extract_output_text(&output_item).expect("output text present");
let (output_text, _success_flag) = call_output(&req, call_id);
let expected_pattern = format!(
r"(?s)^Exit code: 0
@@ -372,7 +361,7 @@ Success. Updated the following files:
A {file_name}
?$"
);
assert_regex_match(&expected_pattern, output_text);
assert_regex_match(&expected_pattern, &output_text);
let updated_contents = fs::read_to_string(file_path)?;
assert_eq!(
@@ -409,13 +398,13 @@ async fn apply_patch_reports_parse_diagnostics() -> anyhow::Result<()> {
ev_apply_patch_function_call(call_id, patch_content),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "failed"),
ev_completed("resp-2"),
]);
let second_mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let second_mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -437,12 +426,7 @@ async fn apply_patch_reports_parse_diagnostics() -> anyhow::Result<()> {
wait_for_event(&codex, |event| matches!(event, EventMsg::TaskComplete(_))).await;
let req = second_mock.single_request();
let output_item = req.function_call_output(call_id);
assert_eq!(
output_item.get("call_id").and_then(Value::as_str),
Some(call_id)
);
let output_text = extract_output_text(&output_item).expect("output text present");
let (output_text, success_flag) = call_output(&req, call_id);
assert!(
output_text.contains("apply_patch verification failed"),
@@ -453,12 +437,7 @@ async fn apply_patch_reports_parse_diagnostics() -> anyhow::Result<()> {
"expected parse diagnostics in output text, got {output_text:?}"
);
if let Some(success_flag) = output_item
.get("output")
.and_then(|value| value.as_object())
.and_then(|obj| obj.get("success"))
.and_then(serde_json::Value::as_bool)
{
if let Some(success_flag) = success_flag {
assert!(
!success_flag,
"expected tool output to mark success=false for parse failures"
codex-rs/core/tests/suite/tools.rs
+16 -61
View File
@@ -10,11 +10,7 @@ use anyhow::Result;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::user_input::UserInput;
use core_test_support::assert_regex_match;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
@@ -26,44 +22,11 @@ use core_test_support::responses::mount_sse_sequence;
use core_test_support::responses::sse;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use regex_lite::Regex;
use serde_json::Value;
use serde_json::json;
async fn submit_turn(
test: &TestCodex,
prompt: &str,
approval_policy: AskForApproval,
sandbox_policy: SandboxPolicy,
) -> Result<()> {
let session_model = test.session_configured.model.clone();
test.codex
.submit(Op::UserTurn {
items: vec![UserInput::Text {
text: prompt.into(),
}],
final_output_json_schema: None,
cwd: test.cwd.path().to_path_buf(),
approval_policy,
sandbox_policy,
model: session_model,
effort: None,
summary: ReasoningSummary::Auto,
})
.await?;
wait_for_event(&test.codex, |event| {
matches!(event, EventMsg::TaskComplete(_))
})
.await;
Ok(())
}
fn tool_names(body: &Value) -> Vec<String> {
body.get("tools")
.and_then(Value::as_array)
@@ -110,8 +73,7 @@ async fn custom_tool_unknown_returns_custom_output_error() -> Result<()> {
)
.await;
submit_turn(
&test,
test.submit_turn_with_policies(
"invoke custom tool",
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
@@ -189,8 +151,7 @@ async fn shell_escalated_permissions_rejected_then_ok() -> Result<()> {
)
.await;
submit_turn(
&test,
test.submit_turn_with_policies(
"run the shell command",
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
@@ -202,24 +163,22 @@ async fn shell_escalated_permissions_rejected_then_ok() -> Result<()> {
"approval policy is {policy:?}; reject command — you should not ask for escalated permissions if the approval policy is {policy:?}"
);
let blocked_item = second_mock
let blocked_output = second_mock
.single_request()
.function_call_output(call_id_blocked);
.function_call_output_content_and_success(call_id_blocked)
.and_then(|(content, _)| content)
.expect("blocked output string");
assert_eq!(
blocked_item.get("output").and_then(Value::as_str),
Some(expected_message.as_str()),
blocked_output, expected_message,
"unexpected rejection message"
);
let success_item = third_mock
let success_output = third_mock
.single_request()
.function_call_output(call_id_success);
let output_json: Value = serde_json::from_str(
success_item
.get("output")
.and_then(Value::as_str)
.expect("success output string"),
)?;
.function_call_output_content_and_success(call_id_success)
.and_then(|(content, _)| content)
.expect("success output string");
let output_json: Value = serde_json::from_str(&success_output)?;
assert_eq!(
output_json["metadata"]["exit_code"].as_i64(),
Some(0),
@@ -348,8 +307,7 @@ async fn collect_tools(use_unified_exec: bool) -> Result<Vec<String>> {
});
let test = builder.build(&server).await?;
submit_turn(
&test,
test.submit_turn_with_policies(
"list tools",
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
@@ -423,8 +381,7 @@ async fn shell_timeout_includes_timeout_prefix_and_metadata() -> Result<()> {
)
.await;
submit_turn(
&test,
test.submit_turn_with_policies(
"run a long command",
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
@@ -515,8 +472,7 @@ time.sleep(60)
let start = Instant::now();
let output_str = tokio::time::timeout(Duration::from_secs(10), async {
submit_turn(
&test,
test.submit_turn_with_policies(
"run a command with a detached grandchild",
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
@@ -600,8 +556,7 @@ async fn shell_spawn_failure_truncates_exec_error() -> Result<()> {
)
.await;
submit_turn(
&test,
test.submit_turn_with_policies(
"spawn a missing binary",
AskForApproval::Never,
SandboxPolicy::DangerFullAccess,
codex-rs/core/tests/suite/truncation.rs
+93 -20
View File
@@ -19,7 +19,7 @@ use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::mount_sse_once_match;
use core_test_support::responses::mount_sse_once;
use core_test_support::responses::mount_sse_sequence;
use core_test_support::responses::sse;
use core_test_support::responses::start_mock_server;
@@ -32,7 +32,6 @@ use serde_json::Value;
use serde_json::json;
use std::collections::HashMap;
use std::time::Duration;
use wiremock::matchers::any;
// Verifies byte-truncation formatting for function error output (RespondToModel errors)
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
@@ -86,7 +85,7 @@ async fn truncate_function_error_trims_respond_to_model() -> Result<()> {
serde_json::from_str::<serde_json::Value>(&output).is_err(),
"expected error output to be plain text",
);
let truncated_pattern = r#"(?s)^Total output lines: 1\s+.*\[\.\.\. output truncated to fit 10240 bytes \.\.\.\]\s*$"#;
let truncated_pattern = r#"(?s)^Total output lines: 1\s+.*\[\.\.\. output truncated to fit 11264 bytes \.\.\.\]\s*$"#;
assert_regex_match(truncated_pattern, &output);
assert!(
!output.contains("omitted"),
@@ -113,15 +112,25 @@ async fn tool_call_output_exceeds_limit_truncated_for_model() -> Result<()> {
let fixture = builder.build(&server).await?;
let call_id = "shell-too-large";
let args = serde_json::json!({
"command": ["/bin/sh", "-c", "seq 1 400"],
"timeout_ms": 5_000,
});
let args = if cfg!(windows) {
serde_json::json!({
"command": [
"powershell",
"-Command",
"for ($i=1; $i -le 400; $i++) { Write-Output $i }"
],
"timeout_ms": 5_000,
})
} else {
serde_json::json!({
"command": ["/bin/sh", "-c", "seq 1 400"],
"timeout_ms": 5_000,
})
};
// First response: model tells us to run the tool; second: complete the turn.
mount_sse_once_match(
mount_sse_once(
&server,
any(),
sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, "shell", &serde_json::to_string(&args)?),
@@ -129,9 +138,8 @@ async fn tool_call_output_exceeds_limit_truncated_for_model() -> Result<()> {
]),
)
.await;
let mock2 = mount_sse_once_match(
let mock2 = mount_sse_once(
&server,
any(),
sse(vec![
responses::ev_assistant_message("msg-1", "done"),
responses::ev_completed("resp-2"),
@@ -149,6 +157,7 @@ async fn tool_call_output_exceeds_limit_truncated_for_model() -> Result<()> {
.single_request()
.function_call_output_text(call_id)
.context("function_call_output present for shell call")?;
let output = output.replace("\r\n", "\n");
// Expect plain text (not JSON) with truncation markers and line elision.
assert!(
@@ -180,6 +189,73 @@ $"#;
Ok(())
}
// Ensures shell tool outputs that exceed the line limit are truncated only once.
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn tool_call_output_truncated_only_once() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.model = "gpt-5-codex".to_string();
config.model_family =
find_family_for_model("gpt-5-codex").expect("gpt-5-codex is a model family");
});
let fixture = builder.build(&server).await?;
let call_id = "shell-single-truncation";
let args = if cfg!(windows) {
serde_json::json!({
"command": [
"powershell",
"-Command",
"for ($i=1; $i -le 2000; $i++) { Write-Output $i }"
],
"timeout_ms": 5_000,
})
} else {
serde_json::json!({
"command": ["/bin/sh", "-c", "seq 1 2000"],
"timeout_ms": 5_000,
})
};
mount_sse_once(
&server,
sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, "shell", &serde_json::to_string(&args)?),
responses::ev_completed("resp-1"),
]),
)
.await;
let mock2 = mount_sse_once(
&server,
sse(vec![
responses::ev_assistant_message("msg-1", "done"),
responses::ev_completed("resp-2"),
]),
)
.await;
fixture
.submit_turn_with_policy("trigger big shell output", SandboxPolicy::DangerFullAccess)
.await?;
let output = mock2
.single_request()
.function_call_output_text(call_id)
.context("function_call_output present for shell call")?;
let total_line_headers = output.matches("Total output lines:").count();
assert_eq!(
total_line_headers, 1,
"shell output should carry only one truncation header: {output}"
);
Ok(())
}
// Verifies that an MCP tool call result exceeding the model formatting limits
// is truncated before being sent back to the model.
#[tokio::test(flavor = "multi_thread", worker_threads = 1)]
@@ -196,9 +272,8 @@ async fn mcp_tool_call_output_exceeds_limit_truncated_for_model() -> Result<()>
let large_msg = "long-message-with-newlines-".repeat(600);
let args_json = serde_json::json!({ "message": large_msg });
mount_sse_once_match(
mount_sse_once(
&server,
any(),
sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, &tool_name, &args_json.to_string()),
@@ -206,9 +281,8 @@ async fn mcp_tool_call_output_exceeds_limit_truncated_for_model() -> Result<()>
]),
)
.await;
let mock2 = mount_sse_once_match(
let mock2 = mount_sse_once(
&server,
any(),
sse(vec![
responses::ev_assistant_message("msg-1", "rmcp echo tool completed."),
responses::ev_completed("resp-2"),
@@ -269,7 +343,8 @@ async fn mcp_tool_call_output_exceeds_limit_truncated_for_model() -> Result<()>
output.starts_with("Total output lines: 1\n\n{"),
"expected total line header and JSON head, got: {output}"
);
let byte_marker = Regex::new(r"\[\.\.\. output truncated to fit 10240 bytes \.\.\.\]")
let byte_marker = Regex::new(r"\[\.\.\. output truncated to fit 11264 bytes \.\.\.\]")
.expect("compile regex");
assert!(
byte_marker.is_match(&output),
@@ -291,9 +366,8 @@ async fn mcp_image_output_preserves_image_and_no_text_summary() -> Result<()> {
let server_name = "rmcp";
let tool_name = format!("mcp__{server_name}__image");
mount_sse_once_match(
mount_sse_once(
&server,
any(),
sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, &tool_name, "{}"),
@@ -301,9 +375,8 @@ async fn mcp_image_output_preserves_image_and_no_text_summary() -> Result<()> {
]),
)
.await;
let final_mock = mount_sse_once_match(
let final_mock = mount_sse_once(
&server,
any(),
sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
codex-rs/core/tests/suite/unified_exec.rs
+13 -2
View File
@@ -214,7 +214,11 @@ async fn unified_exec_emits_exec_command_begin_event() -> Result<()> {
assert_eq!(
begin_event.command,
vec!["/bin/echo hello unified exec".to_string()]
vec![
"/bin/bash".to_string(),
"-lc".to_string(),
"/bin/echo hello unified exec".to_string()
]
);
assert_eq!(begin_event.cwd, cwd.path());
@@ -654,7 +658,14 @@ async fn unified_exec_skips_begin_event_for_empty_input() -> Result<()> {
"expected only the initial command to emit begin event"
);
assert_eq!(begin_events[0].call_id, open_call_id);
assert_eq!(begin_events[0].command[0], "/bin/sh -c echo ready");
assert_eq!(
begin_events[0].command,
vec![
"/bin/bash".to_string(),
"-lc".to_string(),
"/bin/sh -c echo ready".to_string()
]
);
Ok(())
}
codex-rs/core/tests/suite/user_notification.rs
+1 -2
View File
@@ -15,7 +15,6 @@ use pretty_assertions::assert_eq;
use serde_json::Value;
use serde_json::json;
use tempfile::TempDir;
use wiremock::matchers::any;
use responses::ev_assistant_message;
use responses::ev_completed;
@@ -35,7 +34,7 @@ async fn summarize_context_three_requests_and_instructions() -> anyhow::Result<(
let sse1 = sse(vec![ev_assistant_message("m1", "Done"), ev_completed("r1")]);
responses::mount_sse_once_match(&server, any(), sse1).await;
responses::mount_sse_once(&server, sse1).await;
let notify_dir = TempDir::new()?;
// write a script to the notify that touches a file next to it
codex-rs/core/tests/suite/user_shell_cmd.rs
+79
View File
@@ -1,13 +1,25 @@
use anyhow::Context;
use codex_core::ConversationManager;
use codex_core::NewConversation;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::EventMsg;
use codex_core::protocol::ExecCommandEndEvent;
use codex_core::protocol::ExecOutputStream;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_core::protocol::TurnAbortReason;
use core_test_support::assert_regex_match;
use core_test_support::load_default_config_for_test;
use core_test_support::responses;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::mount_sse_once;
use core_test_support::responses::sse;
use core_test_support::responses::start_mock_server;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use core_test_support::wait_for_event_match;
use regex_lite::escape;
@@ -249,3 +261,70 @@ async fn user_shell_command_output_is_truncated_in_history() -> anyhow::Result<(
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn user_shell_command_is_truncated_only_once() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.model = "gpt-5-codex".to_string();
config.model_family =
find_family_for_model("gpt-5-codex").expect("gpt-5-codex is a model family");
});
let fixture = builder.build(&server).await?;
let call_id = "user-shell-double-truncation";
let args = if cfg!(windows) {
serde_json::json!({
"command": [
"powershell",
"-Command",
"for ($i=1; $i -le 2000; $i++) { Write-Output $i }"
],
"timeout_ms": 5_000,
})
} else {
serde_json::json!({
"command": ["/bin/sh", "-c", "seq 1 2000"],
"timeout_ms": 5_000,
})
};
mount_sse_once(
&server,
sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, "shell", &serde_json::to_string(&args)?),
ev_completed("resp-1"),
]),
)
.await;
let mock2 = mount_sse_once(
&server,
sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]),
)
.await;
fixture
.submit_turn_with_policy("trigger big shell output", SandboxPolicy::DangerFullAccess)
.await?;
let output = mock2
.single_request()
.function_call_output_text(call_id)
.context("function_call_output present for shell call")?;
let truncation_headers = output.matches("Total output lines:").count();
assert_eq!(
truncation_headers, 1,
"shell output should carry only one truncation header: {output}"
);
Ok(())
}
codex-rs/core/tests/suite/view_image.rs
+32 -30
View File
@@ -24,7 +24,6 @@ use image::ImageBuffer;
use image::Rgba;
use image::load_from_memory;
use serde_json::Value;
use wiremock::matchers::any;
fn find_image_message(body: &Value) -> Option<&Value> {
body.get("input")
@@ -45,14 +44,6 @@ fn find_image_message(body: &Value) -> Option<&Value> {
})
}
fn extract_output_text(item: &Value) -> Option<&str> {
item.get("output").and_then(|value| match value {
Value::String(text) => Some(text.as_str()),
Value::Object(obj) => obj.get("content").and_then(Value::as_str),
_ => None,
})
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn user_turn_with_local_image_attaches_image() -> anyhow::Result<()> {
skip_if_no_network!(Ok(()));
@@ -79,7 +70,7 @@ async fn user_turn_with_local_image_attaches_image() -> anyhow::Result<()> {
ev_assistant_message("msg-1", "done"),
ev_completed("resp-1"),
]);
let mock = responses::mount_sse_once_match(&server, any(), response).await;
let mock = responses::mount_sse_once(&server, response).await;
let session_model = session_configured.model.clone();
@@ -164,13 +155,13 @@ async fn view_image_tool_attaches_local_image() -> anyhow::Result<()> {
ev_function_call(call_id, "view_image", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
let mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -207,10 +198,12 @@ async fn view_image_tool_attaches_local_image() -> anyhow::Result<()> {
assert_eq!(tool_event.call_id, call_id);
assert_eq!(tool_event.path, abs_path);
let body = mock.single_request().body_json();
let output_item = mock.single_request().function_call_output(call_id);
let output_text = extract_output_text(&output_item).expect("output text present");
let req = mock.single_request();
let body = req.body_json();
let output_text = req
.function_call_output_content_and_success(call_id)
.and_then(|(content, _)| content)
.expect("output text present");
assert_eq!(output_text, "attached local image path");
let image_message =
@@ -272,13 +265,13 @@ async fn view_image_tool_errors_when_path_is_directory() -> anyhow::Result<()> {
ev_function_call(call_id, "view_image", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
let mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -299,9 +292,12 @@ async fn view_image_tool_errors_when_path_is_directory() -> anyhow::Result<()> {
wait_for_event(&codex, |event| matches!(event, EventMsg::TaskComplete(_))).await;
let body_with_tool_output = mock.single_request().body_json();
let output_item = mock.single_request().function_call_output(call_id);
let output_text = extract_output_text(&output_item).expect("output text present");
let req = mock.single_request();
let body_with_tool_output = req.body_json();
let output_text = req
.function_call_output_content_and_success(call_id)
.and_then(|(content, _)| content)
.expect("output text present");
let expected_message = format!("image path `{}` is not a file", abs_path.display());
assert_eq!(output_text, expected_message);
@@ -341,13 +337,13 @@ async fn view_image_tool_placeholder_for_non_image_files() -> anyhow::Result<()>
ev_function_call(call_id, "view_image", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
let mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -398,8 +394,11 @@ async fn view_image_tool_placeholder_for_non_image_files() -> anyhow::Result<()>
"placeholder should mention path: {placeholder}"
);
let output_item = mock.single_request().function_call_output(call_id);
let output_text = extract_output_text(&output_item).expect("output text present");
let output_text = mock
.single_request()
.function_call_output_content_and_success(call_id)
.and_then(|(content, _)| content)
.expect("output text present");
assert_eq!(output_text, "attached local image path");
Ok(())
@@ -429,13 +428,13 @@ async fn view_image_tool_errors_when_file_missing() -> anyhow::Result<()> {
ev_function_call(call_id, "view_image", &arguments),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
responses::mount_sse_once(&server, first_response).await;
let second_response = sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]);
let mock = responses::mount_sse_once_match(&server, any(), second_response).await;
let mock = responses::mount_sse_once(&server, second_response).await;
let session_model = session_configured.model.clone();
@@ -456,9 +455,12 @@ async fn view_image_tool_errors_when_file_missing() -> anyhow::Result<()> {
wait_for_event(&codex, |event| matches!(event, EventMsg::TaskComplete(_))).await;
let body_with_tool_output = mock.single_request().body_json();
let output_item = mock.single_request().function_call_output(call_id);
let output_text = extract_output_text(&output_item).expect("output text present");
let req = mock.single_request();
let body_with_tool_output = req.body_json();
let output_text = req
.function_call_output_content_and_success(call_id)
.and_then(|(content, _)| content)
.expect("output text present");
let expected_prefix = format!("unable to locate image at `{}`:", abs_path.display());
assert!(
output_text.starts_with(&expected_prefix),
codex-rs/exec/src/cli.rs
+4
View File
@@ -52,6 +52,10 @@ pub struct Cli {
#[arg(long = "skip-git-repo-check", default_value_t = false)]
pub skip_git_repo_check: bool,
/// Additional directories that should be writable alongside the primary workspace.
#[arg(long = "add-dir", value_name = "DIR", value_hint = clap::ValueHint::DirPath)]
pub add_dir: Vec<PathBuf>,
/// Path to a JSON Schema file describing the model's final response shape.
#[arg(long = "output-schema", value_name = "FILE")]
pub output_schema: Option<PathBuf>,
codex-rs/exec/src/lib.rs
+2 -1
View File
@@ -61,6 +61,7 @@ pub async fn run_main(cli: Cli, codex_linux_sandbox_exe: Option<PathBuf>) -> any
dangerously_bypass_approvals_and_sandbox,
cwd,
skip_git_repo_check,
add_dir,
color,
last_message_file,
json: json_mode,
@@ -179,7 +180,7 @@ pub async fn run_main(cli: Cli, codex_linux_sandbox_exe: Option<PathBuf>) -> any
show_raw_agent_reasoning: oss.then_some(true),
tools_web_search_request: None,
experimental_sandbox_command_assessment: None,
additional_writable_roots: Vec::new(),
additional_writable_roots: add_dir,
};
// Parse `-c` overrides.
let cli_kv_overrides = match config_overrides.parse_overrides() {
codex-rs/exec/tests/suite/add_dir.rs
+72
View File
@@ -0,0 +1,72 @@
#![cfg(not(target_os = "windows"))]
#![allow(clippy::expect_used, clippy::unwrap_used)]
use core_test_support::responses;
use core_test_support::test_codex_exec::test_codex_exec;
/// Verify that the --add-dir flag is accepted and the command runs successfully.
/// This test confirms the CLI argument is properly wired up.
#[tokio::test(flavor = "multi_thread", worker_threads = 1)]
async fn accepts_add_dir_flag() -> anyhow::Result<()> {
let test = test_codex_exec();
let server = responses::start_mock_server().await;
let body = responses::sse(vec![
responses::ev_response_created("response_1"),
responses::ev_assistant_message("response_1", "Task completed"),
responses::ev_completed("response_1"),
]);
responses::mount_sse_once(&server, body).await;
// Create temporary directories to use with --add-dir
let temp_dir1 = tempfile::tempdir()?;
let temp_dir2 = tempfile::tempdir()?;
test.cmd_with_server(&server)
.arg("--skip-git-repo-check")
.arg("--sandbox")
.arg("workspace-write")
.arg("--add-dir")
.arg(temp_dir1.path())
.arg("--add-dir")
.arg(temp_dir2.path())
.arg("test with additional directories")
.assert()
.code(0);
Ok(())
}
/// Verify that multiple --add-dir flags can be specified.
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn accepts_multiple_add_dir_flags() -> anyhow::Result<()> {
let test = test_codex_exec();
let server = responses::start_mock_server().await;
let body = responses::sse(vec![
responses::ev_response_created("response_1"),
responses::ev_assistant_message("response_1", "Multiple directories accepted"),
responses::ev_completed("response_1"),
]);
responses::mount_sse_once(&server, body).await;
let temp_dir1 = tempfile::tempdir()?;
let temp_dir2 = tempfile::tempdir()?;
let temp_dir3 = tempfile::tempdir()?;
test.cmd_with_server(&server)
.arg("--skip-git-repo-check")
.arg("--sandbox")
.arg("workspace-write")
.arg("--add-dir")
.arg(temp_dir1.path())
.arg("--add-dir")
.arg(temp_dir2.path())
.arg("--add-dir")
.arg(temp_dir3.path())
.arg("test with three directories")
.assert()
.code(0);
Ok(())
}
codex-rs/exec/tests/suite/mod.rs
+1
View File
@@ -1,4 +1,5 @@
// Aggregates all former standalone integration tests as modules.
mod add_dir;
mod apply_patch;
mod auth_env;
mod originator;
codex-rs/exec/tests/suite/output_schema.rs
+1 -2
View File
@@ -4,7 +4,6 @@
use core_test_support::responses;
use core_test_support::test_codex_exec::test_codex_exec;
use serde_json::Value;
use wiremock::matchers::any;
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn exec_includes_output_schema_in_request() -> anyhow::Result<()> {
@@ -28,7 +27,7 @@ async fn exec_includes_output_schema_in_request() -> anyhow::Result<()> {
responses::ev_assistant_message("m1", "fixture hello"),
responses::ev_completed("resp1"),
]);
let response_mock = responses::mount_sse_once_match(&server, any(), body).await;
let response_mock = responses::mount_sse_once(&server, body).await;
test.cmd_with_server(&server)
.arg("--skip-git-repo-check")
codex-rs/exec/tests/suite/server_error_exit.rs
+1 -2
View File
@@ -3,7 +3,6 @@
use core_test_support::responses;
use core_test_support::test_codex_exec::test_codex_exec;
use wiremock::matchers::any;
/// Verify that when the server reports an error, `codex-exec` exits with a
/// non-zero status code so automation can detect failures.
@@ -21,7 +20,7 @@ async fn exits_non_zero_when_server_reports_error() -> anyhow::Result<()> {
"error": {"code": "rate_limit_exceeded", "message": "synthetic server error"}
}
})]);
responses::mount_sse_once_match(&server, any(), body).await;
responses::mount_sse_once(&server, body).await;
test.cmd_with_server(&server)
.arg("--skip-git-repo-check")
codex-rs/protocol/src/items.rs
+4
View File
@@ -11,6 +11,8 @@ use serde::Serialize;
use ts_rs::TS;
#[derive(Debug, Clone, Deserialize, Serialize, TS, JsonSchema)]
#[serde(tag = "type")]
#[ts(tag = "type")]
pub enum TurnItem {
UserMessage(UserMessageItem),
AgentMessage(AgentMessageItem),
@@ -25,6 +27,8 @@ pub struct UserMessageItem {
}
#[derive(Debug, Clone, Deserialize, Serialize, TS, JsonSchema)]
#[serde(tag = "type")]
#[ts(tag = "type")]
pub enum AgentMessageContent {
Text { text: String },
}
codex-rs/protocol/src/protocol.rs
+17 -3
View File
@@ -241,7 +241,7 @@ pub enum AskForApproval {
/// Determines execution restrictions for model shell commands.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Display, JsonSchema, TS)]
#[strum(serialize_all = "kebab-case")]
#[serde(tag = "mode", rename_all = "kebab-case")]
#[serde(tag = "type", rename_all = "kebab-case")]
pub enum SandboxPolicy {
/// No restrictions whatsoever. Use with caution.
#[serde(rename = "danger-full-access")]
@@ -432,6 +432,7 @@ pub struct Event {
/// NOTE: Make sure none of these values have optional types, as it will mess up the extension code-gen.
#[derive(Debug, Clone, Deserialize, Serialize, Display, JsonSchema, TS)]
#[serde(tag = "type", rename_all = "snake_case")]
#[ts(tag = "type")]
#[strum(serialize_all = "snake_case")]
pub enum EventMsg {
/// Error while executing a submission
@@ -617,6 +618,9 @@ pub struct ReasoningContentDeltaEvent {
pub turn_id: String,
pub item_id: String,
pub delta: String,
// load with default value so it's backward compatible with the old format.
#[serde(default)]
pub summary_index: i64,
}
impl HasLegacyEvent for ReasoningContentDeltaEvent {
@@ -633,6 +637,9 @@ pub struct ReasoningRawContentDeltaEvent {
pub turn_id: String,
pub item_id: String,
pub delta: String,
// load with default value so it's backward compatible with the old format.
#[serde(default)]
pub content_index: i64,
}
impl HasLegacyEvent for ReasoningRawContentDeltaEvent {
@@ -923,7 +930,13 @@ pub struct AgentReasoningRawContentDeltaEvent {
}
#[derive(Debug, Clone, Deserialize, Serialize, JsonSchema, TS)]
pub struct AgentReasoningSectionBreakEvent {}
pub struct AgentReasoningSectionBreakEvent {
// load with default value so it's backward compatible with the old format.
#[serde(default)]
pub item_id: String,
#[serde(default)]
pub summary_index: i64,
}
#[derive(Debug, Clone, Deserialize, Serialize, JsonSchema, TS)]
pub struct AgentReasoningDeltaEvent {
@@ -1431,7 +1444,8 @@ pub enum ReviewDecision {
}
#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "snake_case")]
#[serde(tag = "type", rename_all = "snake_case")]
#[ts(tag = "type")]
pub enum FileChange {
Add {
content: String,
codex-rs/tui/src/bottom_pane/list_selection_view.rs
+183 -10
View File
@@ -250,6 +250,10 @@ impl ListSelectionView {
pub(crate) fn take_last_selected_index(&mut self) -> Option<usize> {
self.last_selected_actual_idx.take()
}
fn rows_width(total_width: u16) -> u16 {
total_width.saturating_sub(2)
}
}
impl BottomPaneView for ListSelectionView {
@@ -327,8 +331,13 @@ impl Renderable for ListSelectionView {
// Measure wrapped height for up to MAX_POPUP_ROWS items at the given width.
// Build the same display rows used by the renderer so wrapping math matches.
let rows = self.build_rows();
let rows_height = measure_rows_height(&rows, &self.state, MAX_POPUP_ROWS, width);
let rows_width = Self::rows_width(width);
let rows_height = measure_rows_height(
&rows,
&self.state,
MAX_POPUP_ROWS,
rows_width.saturating_add(1),
);
// Subtract 4 for the padding on the left and right of the header.
let mut height = self.header.desired_height(width.saturating_sub(4));
@@ -362,8 +371,13 @@ impl Renderable for ListSelectionView {
// Subtract 4 for the padding on the left and right of the header.
.desired_height(content_area.width.saturating_sub(4));
let rows = self.build_rows();
let rows_height =
measure_rows_height(&rows, &self.state, MAX_POPUP_ROWS, content_area.width);
let rows_width = Self::rows_width(content_area.width);
let rows_height = measure_rows_height(
&rows,
&self.state,
MAX_POPUP_ROWS,
rows_width.saturating_add(1),
);
let [header_area, _, search_area, list_area] = Layout::vertical([
Constraint::Max(header_height),
Constraint::Max(1),
@@ -398,18 +412,18 @@ impl Renderable for ListSelectionView {
}
if list_area.height > 0 {
let list_area = Rect {
x: list_area.x - 2,
let render_area = Rect {
x: list_area.x.saturating_sub(2),
y: list_area.y,
width: list_area.width + 2,
width: rows_width.max(1),
height: list_area.height,
};
render_rows(
list_area,
render_area,
buf,
&rows,
&self.state,
list_area.height as usize,
render_area.height as usize,
"no matches",
);
}
@@ -467,7 +481,10 @@ mod tests {
}
fn render_lines(view: &ListSelectionView) -> String {
let width = 48;
render_lines_with_width(view, 48)
}
fn render_lines_with_width(view: &ListSelectionView, width: u16) -> String {
let height = view.desired_height(width);
let area = Rect::new(0, 0, width, height);
let mut buf = Buffer::empty(area);
@@ -535,4 +552,160 @@ mod tests {
"expected search query line to include rendered query, got {lines:?}"
);
}
#[test]
fn width_changes_do_not_hide_rows() {
let (tx_raw, _rx) = unbounded_channel::<AppEvent>();
let tx = AppEventSender::new(tx_raw);
let items = vec![
SelectionItem {
name: "gpt-5.1-codex".to_string(),
description: Some(
"Optimized for Codex. Balance of reasoning quality and coding ability."
.to_string(),
),
is_current: true,
dismiss_on_select: true,
..Default::default()
},
SelectionItem {
name: "gpt-5.1-codex-mini".to_string(),
description: Some(
"Optimized for Codex. Cheaper, faster, but less capable.".to_string(),
),
dismiss_on_select: true,
..Default::default()
},
SelectionItem {
name: "gpt-4.1-codex".to_string(),
description: Some(
"Legacy model. Use when you need compatibility with older automations."
.to_string(),
),
dismiss_on_select: true,
..Default::default()
},
];
let view = ListSelectionView::new(
SelectionViewParams {
title: Some("Select Model and Effort".to_string()),
items,
..Default::default()
},
tx,
);
let mut missing: Vec<u16> = Vec::new();
for width in 60..=90 {
let rendered = render_lines_with_width(&view, width);
if !rendered.contains("3.") {
missing.push(width);
}
}
assert!(
missing.is_empty(),
"third option missing at widths {missing:?}"
);
}
#[test]
fn narrow_width_keeps_all_rows_visible() {
let (tx_raw, _rx) = unbounded_channel::<AppEvent>();
let tx = AppEventSender::new(tx_raw);
let desc = "x".repeat(10);
let items: Vec<SelectionItem> = (1..=3)
.map(|idx| SelectionItem {
name: format!("Item {idx}"),
description: Some(desc.clone()),
dismiss_on_select: true,
..Default::default()
})
.collect();
let view = ListSelectionView::new(
SelectionViewParams {
title: Some("Debug".to_string()),
items,
..Default::default()
},
tx,
);
let rendered = render_lines_with_width(&view, 24);
assert!(
rendered.contains("3."),
"third option missing for width 24:\n{rendered}"
);
}
#[test]
fn snapshot_model_picker_width_80() {
let (tx_raw, _rx) = unbounded_channel::<AppEvent>();
let tx = AppEventSender::new(tx_raw);
let items = vec![
SelectionItem {
name: "gpt-5.1-codex".to_string(),
description: Some(
"Optimized for Codex. Balance of reasoning quality and coding ability."
.to_string(),
),
is_current: true,
dismiss_on_select: true,
..Default::default()
},
SelectionItem {
name: "gpt-5.1-codex-mini".to_string(),
description: Some(
"Optimized for Codex. Cheaper, faster, but less capable.".to_string(),
),
dismiss_on_select: true,
..Default::default()
},
SelectionItem {
name: "gpt-4.1-codex".to_string(),
description: Some(
"Legacy model. Use when you need compatibility with older automations."
.to_string(),
),
dismiss_on_select: true,
..Default::default()
},
];
let view = ListSelectionView::new(
SelectionViewParams {
title: Some("Select Model and Effort".to_string()),
items,
..Default::default()
},
tx,
);
assert_snapshot!(
"list_selection_model_picker_width_80",
render_lines_with_width(&view, 80)
);
}
#[test]
fn snapshot_narrow_width_preserves_third_option() {
let (tx_raw, _rx) = unbounded_channel::<AppEvent>();
let tx = AppEventSender::new(tx_raw);
let desc = "x".repeat(10);
let items: Vec<SelectionItem> = (1..=3)
.map(|idx| SelectionItem {
name: format!("Item {idx}"),
description: Some(desc.clone()),
dismiss_on_select: true,
..Default::default()
})
.collect();
let view = ListSelectionView::new(
SelectionViewParams {
title: Some("Debug".to_string()),
items,
..Default::default()
},
tx,
);
assert_snapshot!(
"list_selection_narrow_width_preserves_rows",
render_lines_with_width(&view, 24)
);
}
}
codex-rs/tui/src/bottom_pane/snapshots/codex_tui__bottom_pane__list_selection_view__tests__list_selection_model_picker_width_80.snap
+13
View File
@@ -0,0 +1,13 @@
---
source: tui/src/bottom_pane/list_selection_view.rs
expression: "render_lines_with_width(&view, 80)"
---
Select Model and Effort
› 1. gpt-5.1-codex (current) Optimized for Codex. Balance of reasoning
quality and coding ability.
2. gpt-5.1-codex-mini Optimized for Codex. Cheaper, faster, but less
capable.
3. gpt-4.1-codex Legacy model. Use when you need compatibility
with older automations.
codex-rs/tui/src/bottom_pane/snapshots/codex_tui__bottom_pane__list_selection_view__tests__list_selection_narrow_width_preserves_rows.snap
+16
View File
@@ -0,0 +1,16 @@
---
source: tui/src/bottom_pane/list_selection_view.rs
expression: "render_lines_with_width(&view, 24)"
---
Debug
› 1. Item 1
xxxxxxxxx
x
2. Item 2
xxxxxxxxx
x
3. Item 3
xxxxxxxxx
x
codex-rs/tui/src/chatwidget/tests.rs
+31
View File
@@ -100,6 +100,37 @@ fn upgrade_event_payload_for_tests(mut payload: serde_json::Value) -> serde_json
serde_json::Value::String(aggregated),
);
}
} else if ty == "patch_apply_begin"
&& let Some(changes) = m.get_mut("changes").and_then(|v| v.as_object_mut())
{
for change in changes.values_mut() {
if change.get("type").is_some() {
continue;
}
if let Some(change_obj) = change.as_object_mut()
&& change_obj.len() == 1
&& let Some((legacy_kind, legacy_value)) = change_obj
.iter()
.next()
.map(|(k, v)| (k.clone(), v.clone()))
{
change_obj.clear();
change_obj.insert(
"type".to_string(),
serde_json::Value::String(legacy_kind.clone()),
);
match legacy_value {
serde_json::Value::Object(payload) => {
for (k, v) in payload {
change_obj.insert(k, v);
}
}
other => {
change_obj.insert("content".to_string(), other);
}
}
}
}
}
}
payload
sdk/typescript/eslint.config.js
+7
View File
@@ -10,5 +10,12 @@ export default defineConfig(eslint.configs.recommended, tseslint.configs.recomme
rules: {
"node-import/prefer-node-protocol": 2,
"@typescript-eslint/no-unused-vars": [
"error",
{
argsIgnorePattern: "^_",
varsIgnorePattern: "^_",
},
],
},
});
sdk/typescript/src/exec.ts
+15 -1
View File
@@ -18,12 +18,16 @@ export type CodexExecArgs = {
sandboxMode?: SandboxMode;
// --cd
workingDirectory?: string;
// --add-dir
additionalDirectories?: string[];
// --skip-git-repo-check
skipGitRepoCheck?: boolean;
// --output-schema
outputSchemaFile?: string;
// --config model_reasoning_effort
modelReasoningEffort?: ModelReasoningEffort;
// AbortSignal to cancel the execution
signal?: AbortSignal;
// --config sandbox_workspace_write.network_access
networkAccessEnabled?: boolean;
// --config features.web_search_request
@@ -56,6 +60,12 @@ export class CodexExec {
commandArgs.push("--cd", args.workingDirectory);
}
if (args.additionalDirectories?.length) {
for (const dir of args.additionalDirectories) {
commandArgs.push("--add-dir", dir);
}
}
if (args.skipGitRepoCheck) {
commandArgs.push("--skip-git-repo-check");
}
@@ -69,7 +79,10 @@ export class CodexExec {
}
if (args.networkAccessEnabled !== undefined) {
commandArgs.push("--config", `sandbox_workspace_write.network_access=${args.networkAccessEnabled}`);
commandArgs.push(
"--config",
`sandbox_workspace_write.network_access=${args.networkAccessEnabled}`,
);
}
if (args.webSearchEnabled !== undefined) {
@@ -105,6 +118,7 @@ export class CodexExec {
const child = spawn(this.executablePath, commandArgs, {
env,
signal: args.signal,
});
let spawnError: unknown | null = null;
sdk/typescript/src/thread.ts
+2
View File
@@ -86,9 +86,11 @@ export class Thread {
skipGitRepoCheck: options?.skipGitRepoCheck,
outputSchemaFile: schemaPath,
modelReasoningEffort: options?.modelReasoningEffort,
signal: turnOptions.signal,
networkAccessEnabled: options?.networkAccessEnabled,
webSearchEnabled: options?.webSearchEnabled,
approvalPolicy: options?.approvalPolicy,
additionalDirectories: options?.additionalDirectories,
});
try {
for await (const item of generator) {
sdk/typescript/src/threadOptions.ts
+1
View File
@@ -13,4 +13,5 @@ export type ThreadOptions = {
networkAccessEnabled?: boolean;
webSearchEnabled?: boolean;
approvalPolicy?: ApprovalMode;
additionalDirectories?: string[];
};
sdk/typescript/src/turnOptions.ts
+2
View File
@@ -1,4 +1,6 @@
export type TurnOptions = {
/** JSON schema describing the expected agent output. */
outputSchema?: unknown;
/** AbortSignal to cancel the turn. */
signal?: AbortSignal;
};
sdk/typescript/tests/abort.test.ts
+165
View File
@@ -0,0 +1,165 @@
import path from "node:path";
import { describe, expect, it } from "@jest/globals";
import { Codex } from "../src/codex";
import {
assistantMessage,
responseCompleted,
responseStarted,
shell_call as shellCall,
sse,
SseResponseBody,
startResponsesTestProxy,
} from "./responsesProxy";
const codexExecPath = path.join(process.cwd(), "..", "..", "codex-rs", "target", "debug", "codex");
function* infiniteShellCall(): Generator<SseResponseBody> {
while (true) {
yield sse(responseStarted(), shellCall(), responseCompleted());
}
}
describe("AbortSignal support", () => {
it("aborts run() when signal is aborted", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: infiniteShellCall(),
});
try {
const client = new Codex({ codexPathOverride: codexExecPath, baseUrl: url, apiKey: "test" });
const thread = client.startThread();
// Create an abort controller and abort it immediately
const controller = new AbortController();
controller.abort("Test abort");
// The operation should fail because the signal is already aborted
await expect(thread.run("Hello, world!", { signal: controller.signal })).rejects.toThrow();
} finally {
await close();
}
});
it("aborts runStreamed() when signal is aborted", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: infiniteShellCall(),
});
try {
const client = new Codex({ codexPathOverride: codexExecPath, baseUrl: url, apiKey: "test" });
const thread = client.startThread();
// Create an abort controller and abort it immediately
const controller = new AbortController();
controller.abort("Test abort");
const { events } = await thread.runStreamed("Hello, world!", { signal: controller.signal });
// Attempting to iterate should fail
let iterationStarted = false;
try {
for await (const event of events) {
iterationStarted = true;
// Should not get here
expect(event).toBeUndefined();
}
// If we get here, the test should fail
throw new Error(
"Expected iteration to throw due to aborted signal, but it completed successfully",
);
} catch (error) {
// We expect an error to be thrown
expect(iterationStarted).toBe(false); // Should fail before any iteration
expect(error).toBeDefined();
}
} finally {
await close();
}
});
it("aborts run() when signal is aborted during execution", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: infiniteShellCall(),
});
try {
const client = new Codex({ codexPathOverride: codexExecPath, baseUrl: url, apiKey: "test" });
const thread = client.startThread();
const controller = new AbortController();
// Start the operation and abort it immediately after
const runPromise = thread.run("Hello, world!", { signal: controller.signal });
// Abort after a tiny delay to simulate aborting during execution
setTimeout(() => controller.abort("Aborted during execution"), 10);
// The operation should fail
await expect(runPromise).rejects.toThrow();
} finally {
await close();
}
});
it("aborts runStreamed() when signal is aborted during iteration", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: infiniteShellCall(),
});
try {
const client = new Codex({ codexPathOverride: codexExecPath, baseUrl: url, apiKey: "test" });
const thread = client.startThread();
const controller = new AbortController();
const { events } = await thread.runStreamed("Hello, world!", { signal: controller.signal });
// Abort during iteration
let eventCount = 0;
await expect(
(async () => {
for await (const event of events) {
void event; // Consume the event
eventCount++;
// Abort after first event
if (eventCount === 5) {
controller.abort("Aborted during iteration");
}
// Continue iterating - should eventually throw
}
})(),
).rejects.toThrow();
} finally {
await close();
}
});
it("completes normally when signal is not aborted", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: [sse(responseStarted(), assistantMessage("Hi!"), responseCompleted())],
});
try {
const client = new Codex({ codexPathOverride: codexExecPath, baseUrl: url, apiKey: "test" });
const thread = client.startThread();
const controller = new AbortController();
// Don't abort - should complete successfully
const result = await thread.run("Hello, world!", { signal: controller.signal });
expect(result.finalResponse).toBe("Hi!");
expect(result.items).toHaveLength(1);
} finally {
await close();
}
});
});
sdk/typescript/tests/responsesProxy.ts
+29 -10
View File
@@ -19,18 +19,18 @@ const DEFAULT_COMPLETED_USAGE: ResponseCompletedUsage = {
total_tokens: 47,
};
type SseEvent = {
export type SseEvent = {
type: string;
[key: string]: unknown;
};
type SseResponseBody = {
export type SseResponseBody = {
kind: "sse";
events: SseEvent[];
};
export type ResponsesProxyOptions = {
responseBodies: SseResponseBody[];
responseBodies: Generator<SseResponseBody> | SseResponseBody[];
statusCode?: number;
};
@@ -64,11 +64,17 @@ function formatSseEvent(event: SseEvent): string {
export async function startResponsesTestProxy(
options: ResponsesProxyOptions,
): Promise<ResponsesProxy> {
const responseBodies = options.responseBodies;
if (responseBodies.length === 0) {
throw new Error("responseBodies is required");
function* createGenerator(array: SseResponseBody[]): Generator<SseResponseBody> {
for (const elem of array) {
yield elem;
}
throw new Error("not enough responses provided");
}
const responseBodies: Generator<SseResponseBody> = Array.isArray(options.responseBodies)
? createGenerator(options.responseBodies)
: options.responseBodies;
const requests: RecordedRequest[] = [];
function readRequestBody(req: http.IncomingMessage): Promise<string> {
@@ -84,8 +90,6 @@ export async function startResponsesTestProxy(
});
}
let responseIndex = 0;
const server = http.createServer((req, res) => {
async function handle(): Promise<void> {
if (req.method === "POST" && req.url === "/responses") {
@@ -97,8 +101,7 @@ export async function startResponsesTestProxy(
res.statusCode = status;
res.setHeader("content-type", "text/event-stream");
const responseBody = responseBodies[Math.min(responseIndex, responseBodies.length - 1)]!;
responseIndex += 1;
const responseBody = responseBodies.next().value;
for (const event of responseBody.events) {
res.write(formatSseEvent(event));
}
@@ -177,6 +180,22 @@ export function assistantMessage(text: string, itemId: string = DEFAULT_MESSAGE_
};
}
export function shell_call(): SseEvent {
const command = ["bash", "-lc", "echo 'Hello, world!'"];
return {
type: "response.output_item.done",
item: {
type: "function_call",
call_id: `call_id${Math.random().toString(36).slice(2)}`,
name: "shell",
arguments: JSON.stringify({
command,
timeout_ms: 100,
}),
},
};
}
export function responseFailed(errorMessage: string): SseEvent {
return {
type: "error",
sdk/typescript/tests/run.test.ts
+49 -4
View File
@@ -14,6 +14,7 @@ import {
sse,
responseFailed,
startResponsesTestProxy,
SseResponseBody,
} from "./responsesProxy";
const codexExecPath = path.join(process.cwd(), "..", "..", "codex-rs", "target", "debug", "codex");
@@ -347,6 +348,48 @@ describe("Codex", () => {
}
});
it("passes additionalDirectories as repeated flags", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: [
sse(
responseStarted("response_1"),
assistantMessage("Additional directories applied", "item_1"),
responseCompleted("response_1"),
),
],
});
const { args: spawnArgs, restore } = codexExecSpy();
try {
const client = new Codex({ codexPathOverride: codexExecPath, baseUrl: url, apiKey: "test" });
const thread = client.startThread({
additionalDirectories: ["../backend", "/tmp/shared"],
});
await thread.run("test additional dirs");
const commandArgs = spawnArgs[0];
expect(commandArgs).toBeDefined();
if (!commandArgs) {
throw new Error("Command args missing");
}
// Find the --add-dir flags
const addDirArgs: string[] = [];
for (let i = 0; i < commandArgs.length; i += 1) {
if (commandArgs[i] === "--add-dir") {
addDirArgs.push(commandArgs[i + 1] ?? "");
}
}
expect(addDirArgs).toEqual(["../backend", "/tmp/shared"]);
} finally {
restore();
await close();
}
});
it("writes output schema to a temporary file and forwards it", async () => {
const { url, close, requests } = await startResponsesTestProxy({
statusCode: 200,
@@ -572,10 +615,12 @@ describe("Codex", () => {
it("throws ThreadRunError on turn failures", async () => {
const { url, close } = await startResponsesTestProxy({
statusCode: 200,
responseBodies: [
sse(responseStarted("response_1")),
sse(responseFailed("rate limit exceeded")),
],
responseBodies: (function* (): Generator<SseResponseBody> {
yield sse(responseStarted("response_1"));
while (true) {
yield sse(responseFailed("rate limit exceeded"));
}
})(),
});
try {