Compare commits

..

1 Commits

Author SHA1 Message Date
Michael Bolin
0a68853d8b feat: build the standalone codex-exec binary as part of the GitHub Release 2025-09-29 09:54:54 -07:00
288 changed files with 4084 additions and 26825 deletions

View File

@@ -1,6 +1,6 @@
[codespell]
# Ref: https://github.com/codespell-project/codespell#using-a-config-file
skip = .git*,vendor,*-lock.yaml,*.lock,.codespellrc,*test.ts,*.jsonl,frame*.txt
skip = .git*,vendor,*-lock.yaml,*.lock,.codespellrc,*test.ts,*.jsonl
check-hidden = true
ignore-regex = ^\s*"image/\S+": ".*|\b(afterAll)\b
ignore-words-list = ratatui,ser

View File

@@ -27,7 +27,7 @@ jobs:
- name: Install dependencies
run: pnpm install --frozen-lockfile
# stage_npm_packages.py requires DotSlash when staging releases.
# build_npm_package.py requires DotSlash when staging releases.
- uses: facebook/install-dotslash@v2
- name: Stage npm package
@@ -37,12 +37,10 @@ jobs:
run: |
set -euo pipefail
CODEX_VERSION=0.40.0
OUTPUT_DIR="${RUNNER_TEMP}"
python3 ./scripts/stage_npm_packages.py \
PACK_OUTPUT="${RUNNER_TEMP}/codex-npm.tgz"
python3 ./codex-cli/scripts/build_npm_package.py \
--release-version "$CODEX_VERSION" \
--package codex \
--output-dir "$OUTPUT_DIR"
PACK_OUTPUT="${OUTPUT_DIR}/codex-npm-${CODEX_VERSION}.tgz"
--pack-output "$PACK_OUTPUT"
echo "pack_output=$PACK_OUTPUT" >> "$GITHUB_OUTPUT"
- name: Upload staged npm package artifact

View File

@@ -25,3 +25,4 @@ jobs:
uses: codespell-project/actions-codespell@406322ec52dd7b488e48c1c4b82e2a8b3a1bf630 # v2.1
with:
ignore_words_file: .codespellignore
skip: frame*.txt

View File

@@ -97,7 +97,7 @@ jobs:
sudo apt install -y musl-tools pkg-config
- name: Cargo build
run: cargo build --target ${{ matrix.target }} --release --bin codex --bin codex-responses-api-proxy
run: cargo build --target ${{ matrix.target }} --release --bin codex --bin codex-responses-api-proxy --bin codex-exec
- name: Stage artifacts
shell: bash
@@ -216,30 +216,31 @@ jobs:
echo "npm_tag=" >> "$GITHUB_OUTPUT"
fi
- name: Setup pnpm
uses: pnpm/action-setup@v4
with:
run_install: false
- name: Setup Node.js for npm packaging
uses: actions/setup-node@v5
with:
node-version: 22
- name: Install dependencies
run: pnpm install --frozen-lockfile
# stage_npm_packages.py requires DotSlash when staging releases.
# build_npm_package.py requires DotSlash when staging releases.
- uses: facebook/install-dotslash@v2
- name: Stage npm packages
- name: Stage codex CLI npm package
env:
GH_TOKEN: ${{ github.token }}
run: |
./scripts/stage_npm_packages.py \
--release-version "${{ steps.release_name.outputs.name }}" \
set -euo pipefail
TMP_DIR="${RUNNER_TEMP}/npm-stage"
./codex-cli/scripts/build_npm_package.py \
--package codex \
--release-version "${{ steps.release_name.outputs.name }}" \
--staging-dir "${TMP_DIR}" \
--pack-output "${GITHUB_WORKSPACE}/dist/npm/codex-npm-${{ steps.release_name.outputs.name }}.tgz"
- name: Stage responses API proxy npm package
env:
GH_TOKEN: ${{ github.token }}
run: |
set -euo pipefail
TMP_DIR="${RUNNER_TEMP}/npm-stage-responses"
./codex-cli/scripts/build_npm_package.py \
--package codex-responses-api-proxy \
--package codex-sdk
--release-version "${{ steps.release_name.outputs.name }}" \
--staging-dir "${TMP_DIR}" \
--pack-output "${GITHUB_WORKSPACE}/dist/npm/codex-responses-api-proxy-npm-${{ steps.release_name.outputs.name }}.tgz"
- name: Create GitHub Release
uses: softprops/action-gh-release@v2
@@ -299,10 +300,6 @@ jobs:
--repo "${GITHUB_REPOSITORY}" \
--pattern "codex-responses-api-proxy-npm-${version}.tgz" \
--dir dist/npm
gh release download "$tag" \
--repo "${GITHUB_REPOSITORY}" \
--pattern "codex-sdk-npm-${version}.tgz" \
--dir dist/npm
# No NODE_AUTH_TOKEN needed because we use OIDC.
- name: Publish to npm
@@ -319,7 +316,6 @@ jobs:
tarballs=(
"codex-npm-${VERSION}.tgz"
"codex-responses-api-proxy-npm-${VERSION}.tgz"
"codex-sdk-npm-${VERSION}.tgz"
)
for tarball in "${tarballs[@]}"; do

View File

@@ -1,43 +0,0 @@
name: sdk
on:
push:
branches: [main]
pull_request: {}
jobs:
sdks:
runs-on: ubuntu-latest
timeout-minutes: 10
steps:
- name: Checkout repository
uses: actions/checkout@v5
- name: Setup pnpm
uses: pnpm/action-setup@v4
with:
run_install: false
- name: Setup Node.js
uses: actions/setup-node@v5
with:
node-version: 22
cache: pnpm
- uses: dtolnay/rust-toolchain@1.90
- name: build codex
run: cargo build --bin codex
working-directory: codex-rs
- name: Install dependencies
run: pnpm install --frozen-lockfile
- name: Build SDK packages
run: pnpm -r --filter ./sdk/typescript run build
- name: Lint SDK packages
run: pnpm -r --filter ./sdk/typescript run lint
- name: Test SDK packages
run: pnpm -r --filter ./sdk/typescript run test

View File

@@ -1,19 +1,11 @@
# npm releases
Use the staging helper in the repo root to generate npm tarballs for a release. For
example, to stage the CLI, responses proxy, and SDK packages for version `0.6.0`:
Run the following:
To build the 0.2.x or later version of the npm module, which runs the Rust version of the CLI, build it as follows:
```bash
./scripts/stage_npm_packages.py \
--release-version 0.6.0 \
--package codex \
--package codex-responses-api-proxy \
--package codex-sdk
./codex-cli/scripts/build_npm_package.py --release-version 0.6.0
```
This downloads the native artifacts once, hydrates `vendor/` for each package, and writes
tarballs to `dist/npm/`.
If you need to invoke `build_npm_package.py` directly, run
`codex-cli/scripts/install_native_deps.py` first and pass `--vendor-src` pointing to the
directory that contains the populated `vendor/` tree.
Note this will create `./codex-cli/vendor/` as a side-effect.

View File

@@ -3,6 +3,7 @@
import argparse
import json
import re
import shutil
import subprocess
import sys
@@ -13,25 +14,19 @@ SCRIPT_DIR = Path(__file__).resolve().parent
CODEX_CLI_ROOT = SCRIPT_DIR.parent
REPO_ROOT = CODEX_CLI_ROOT.parent
RESPONSES_API_PROXY_NPM_ROOT = REPO_ROOT / "codex-rs" / "responses-api-proxy" / "npm"
CODEX_SDK_ROOT = REPO_ROOT / "sdk" / "typescript"
GITHUB_REPO = "openai/codex"
PACKAGE_NATIVE_COMPONENTS: dict[str, list[str]] = {
"codex": ["codex", "rg"],
"codex-responses-api-proxy": ["codex-responses-api-proxy"],
"codex-sdk": ["codex"],
}
COMPONENT_DEST_DIR: dict[str, str] = {
"codex": "codex",
"codex-responses-api-proxy": "codex-responses-api-proxy",
"rg": "path",
}
# The docs are not clear on what the expected value/format of
# workflow/workflowName is:
# https://cli.github.com/manual/gh_run_list
WORKFLOW_NAME = ".github/workflows/rust-release.yml"
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser(description="Build or stage the Codex CLI npm package.")
parser.add_argument(
"--package",
choices=("codex", "codex-responses-api-proxy", "codex-sdk"),
choices=("codex", "codex-responses-api-proxy"),
default="codex",
help="Which npm package to stage (default: codex).",
)
@@ -42,9 +37,14 @@ def parse_args() -> argparse.Namespace:
parser.add_argument(
"--release-version",
help=(
"Version to stage for npm release."
"Version to stage for npm release. When provided, the script also resolves the "
"matching rust-release workflow unless --workflow-url is supplied."
),
)
parser.add_argument(
"--workflow-url",
help="Optional GitHub Actions workflow run URL used to download native binaries.",
)
parser.add_argument(
"--staging-dir",
type=Path,
@@ -64,11 +64,6 @@ def parse_args() -> argparse.Namespace:
type=Path,
help="Path where the generated npm tarball should be written.",
)
parser.add_argument(
"--vendor-src",
type=Path,
help="Directory containing pre-installed native binaries to bundle (vendor root).",
)
return parser.parse_args()
@@ -91,19 +86,29 @@ def main() -> int:
try:
stage_sources(staging_dir, version, package)
vendor_src = args.vendor_src.resolve() if args.vendor_src else None
native_components = PACKAGE_NATIVE_COMPONENTS.get(package, [])
workflow_url = args.workflow_url
resolved_head_sha: str | None = None
if not workflow_url:
if release_version:
workflow = resolve_release_workflow(version)
workflow_url = workflow["url"]
resolved_head_sha = workflow.get("headSha")
else:
workflow_url = resolve_latest_alpha_workflow_url()
elif release_version:
try:
workflow = resolve_release_workflow(version)
resolved_head_sha = workflow.get("headSha")
except Exception:
resolved_head_sha = None
if native_components:
if vendor_src is None:
components_str = ", ".join(native_components)
raise RuntimeError(
"Native components "
f"({components_str}) required for package '{package}'. Provide --vendor-src "
"pointing to a directory containing pre-installed binaries."
)
if release_version and resolved_head_sha:
print(f"should `git checkout {resolved_head_sha}`")
copy_native_binaries(vendor_src, staging_dir, native_components)
if not workflow_url:
raise RuntimeError("Unable to determine workflow URL for native binaries.")
install_native_binaries(staging_dir, workflow_url, package)
if release_version:
staging_dir_str = str(staging_dir)
@@ -114,20 +119,12 @@ def main() -> int:
f" node {staging_dir_str}/bin/codex.js --version\n"
f" node {staging_dir_str}/bin/codex.js --help\n\n"
)
elif package == "codex-responses-api-proxy":
else:
print(
f"Staged version {version} for release in {staging_dir_str}\n\n"
"Verify the responses API proxy:\n"
f" node {staging_dir_str}/bin/codex-responses-api-proxy.js --help\n\n"
)
else:
print(
f"Staged version {version} for release in {staging_dir_str}\n\n"
"Verify the SDK contents:\n"
f" ls {staging_dir_str}/dist\n"
f" ls {staging_dir_str}/vendor\n"
" node -e \"import('./dist/index.js').then(() => console.log('ok'))\"\n\n"
)
else:
print(f"Staged package in {staging_dir}")
@@ -155,9 +152,10 @@ def prepare_staging_dir(staging_dir: Path | None) -> tuple[Path, bool]:
def stage_sources(staging_dir: Path, version: str, package: str) -> None:
bin_dir = staging_dir / "bin"
bin_dir.mkdir(parents=True, exist_ok=True)
if package == "codex":
bin_dir = staging_dir / "bin"
bin_dir.mkdir(parents=True, exist_ok=True)
shutil.copy2(CODEX_CLI_ROOT / "bin" / "codex.js", bin_dir / "codex.js")
rg_manifest = CODEX_CLI_ROOT / "bin" / "rg"
if rg_manifest.exists():
@@ -169,8 +167,6 @@ def stage_sources(staging_dir: Path, version: str, package: str) -> None:
package_json_path = CODEX_CLI_ROOT / "package.json"
elif package == "codex-responses-api-proxy":
bin_dir = staging_dir / "bin"
bin_dir.mkdir(parents=True, exist_ok=True)
launcher_src = RESPONSES_API_PROXY_NPM_ROOT / "bin" / "codex-responses-api-proxy.js"
shutil.copy2(launcher_src, bin_dir / "codex-responses-api-proxy.js")
@@ -179,9 +175,6 @@ def stage_sources(staging_dir: Path, version: str, package: str) -> None:
shutil.copy2(readme_src, staging_dir / "README.md")
package_json_path = RESPONSES_API_PROXY_NPM_ROOT / "package.json"
elif package == "codex-sdk":
package_json_path = CODEX_SDK_ROOT / "package.json"
stage_codex_sdk_sources(staging_dir)
else:
raise RuntimeError(f"Unknown package '{package}'.")
@@ -189,85 +182,91 @@ def stage_sources(staging_dir: Path, version: str, package: str) -> None:
package_json = json.load(fh)
package_json["version"] = version
if package == "codex-sdk":
scripts = package_json.get("scripts")
if isinstance(scripts, dict):
scripts.pop("prepare", None)
files = package_json.get("files")
if isinstance(files, list):
if "vendor" not in files:
files.append("vendor")
else:
package_json["files"] = ["dist", "vendor"]
with open(staging_dir / "package.json", "w", encoding="utf-8") as out:
json.dump(package_json, out, indent=2)
out.write("\n")
def run_command(cmd: list[str], cwd: Path | None = None) -> None:
print("+", " ".join(cmd))
subprocess.run(cmd, cwd=cwd, check=True)
def install_native_binaries(staging_dir: Path, workflow_url: str, package: str) -> None:
package_components = {
"codex": ["codex", "rg"],
"codex-responses-api-proxy": ["codex-responses-api-proxy"],
}
components = package_components.get(package)
if components is None:
raise RuntimeError(f"Unknown package '{package}'.")
cmd = ["./scripts/install_native_deps.py", "--workflow-url", workflow_url]
for component in components:
cmd.extend(["--component", component])
cmd.append(str(staging_dir))
subprocess.check_call(cmd, cwd=CODEX_CLI_ROOT)
def stage_codex_sdk_sources(staging_dir: Path) -> None:
package_root = CODEX_SDK_ROOT
run_command(["pnpm", "install", "--frozen-lockfile"], cwd=package_root)
run_command(["pnpm", "run", "build"], cwd=package_root)
dist_src = package_root / "dist"
if not dist_src.exists():
raise RuntimeError("codex-sdk build did not produce a dist directory.")
shutil.copytree(dist_src, staging_dir / "dist")
readme_src = package_root / "README.md"
if readme_src.exists():
shutil.copy2(readme_src, staging_dir / "README.md")
license_src = REPO_ROOT / "LICENSE"
if license_src.exists():
shutil.copy2(license_src, staging_dir / "LICENSE")
def resolve_latest_alpha_workflow_url() -> str:
version = determine_latest_alpha_version()
workflow = resolve_release_workflow(version)
return workflow["url"]
def copy_native_binaries(vendor_src: Path, staging_dir: Path, components: list[str]) -> None:
vendor_src = vendor_src.resolve()
if not vendor_src.exists():
raise RuntimeError(f"Vendor source directory not found: {vendor_src}")
components_set = {component for component in components if component in COMPONENT_DEST_DIR}
if not components_set:
return
vendor_dest = staging_dir / "vendor"
if vendor_dest.exists():
shutil.rmtree(vendor_dest)
vendor_dest.mkdir(parents=True, exist_ok=True)
for target_dir in vendor_src.iterdir():
if not target_dir.is_dir():
def determine_latest_alpha_version() -> str:
releases = list_releases()
best_key: tuple[int, int, int, int] | None = None
best_version: str | None = None
pattern = re.compile(r"^rust-v(\d+)\.(\d+)\.(\d+)-alpha\.(\d+)$")
for release in releases:
tag = release.get("tag_name", "")
match = pattern.match(tag)
if not match:
continue
key = tuple(int(match.group(i)) for i in range(1, 5))
if best_key is None or key > best_key:
best_key = key
best_version = (
f"{match.group(1)}.{match.group(2)}.{match.group(3)}-alpha.{match.group(4)}"
)
dest_target_dir = vendor_dest / target_dir.name
dest_target_dir.mkdir(parents=True, exist_ok=True)
if best_version is None:
raise RuntimeError("No alpha releases found when resolving workflow URL.")
return best_version
for component in components_set:
dest_dir_name = COMPONENT_DEST_DIR.get(component)
if dest_dir_name is None:
continue
src_component_dir = target_dir / dest_dir_name
if not src_component_dir.exists():
raise RuntimeError(
f"Missing native component '{component}' in vendor source: {src_component_dir}"
)
def list_releases() -> list[dict]:
stdout = subprocess.check_output(
["gh", "api", f"/repos/{GITHUB_REPO}/releases?per_page=100"],
text=True,
)
try:
releases = json.loads(stdout or "[]")
except json.JSONDecodeError as exc:
raise RuntimeError("Unable to parse releases JSON.") from exc
if not isinstance(releases, list):
raise RuntimeError("Unexpected response when listing releases.")
return releases
dest_component_dir = dest_target_dir / dest_dir_name
if dest_component_dir.exists():
shutil.rmtree(dest_component_dir)
shutil.copytree(src_component_dir, dest_component_dir)
def resolve_release_workflow(version: str) -> dict:
stdout = subprocess.check_output(
[
"gh",
"run",
"list",
"--branch",
f"rust-v{version}",
"--json",
"workflowName,url,headSha",
"--workflow",
WORKFLOW_NAME,
"--jq",
"first(.[])",
],
text=True,
)
workflow = json.loads(stdout or "[]")
if not workflow:
raise RuntimeError(f"Unable to find rust-release workflow for version {version}.")
return workflow
def run_npm_pack(staging_dir: Path, output_path: Path) -> Path:

1267
codex-rs/Cargo.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -1,14 +1,8 @@
[workspace]
members = [
"backend-client",
"ansi-escape",
"app-server",
"app-server-protocol",
"apply-patch",
"arg0",
"codex-backend-openapi-models",
"cloud-tasks",
"cloud-tasks-client",
"cli",
"common",
"core",
@@ -27,10 +21,7 @@ members = [
"protocol-ts",
"rmcp-client",
"responses-api-proxy",
"otel",
"tui",
"git-apply",
"utils/json-to-toml",
"utils/readiness",
]
resolver = "2"
@@ -45,10 +36,7 @@ edition = "2024"
[workspace.dependencies]
# Internal
app_test_support = { path = "app-server/tests/common" }
codex-ansi-escape = { path = "ansi-escape" }
codex-app-server = { path = "app-server" }
codex-app-server-protocol = { path = "app-server-protocol" }
codex-apply-patch = { path = "apply-patch" }
codex-arg0 = { path = "arg0" }
codex-chatgpt = { path = "chatgpt" }
@@ -62,14 +50,11 @@ codex-login = { path = "login" }
codex-mcp-client = { path = "mcp-client" }
codex-mcp-server = { path = "mcp-server" }
codex-ollama = { path = "ollama" }
codex-otel = { path = "otel" }
codex-process-hardening = { path = "process-hardening" }
codex-protocol = { path = "protocol" }
codex-protocol-ts = { path = "protocol-ts" }
codex-responses-api-proxy = { path = "responses-api-proxy" }
codex-rmcp-client = { path = "rmcp-client" }
codex-tui = { path = "tui" }
codex-utils-json-to-toml = { path = "utils/json-to-toml" }
codex-utils-readiness = { path = "utils/readiness" }
core_test_support = { path = "core/tests/common" }
mcp-types = { path = "mcp-types" }
@@ -97,7 +82,6 @@ derive_more = "2"
diffy = "0.4.2"
dirs = "6"
dotenvy = "0.15.7"
dunce = "1.0.4"
env-flags = "0.1.1"
env_logger = "0.11.5"
escargot = "0.5"
@@ -119,14 +103,8 @@ mime_guess = "2.0.5"
multimap = "0.10.0"
nucleo-matcher = "0.3.1"
openssl-sys = "*"
opentelemetry = "0.30.0"
opentelemetry-appender-tracing = "0.30.0"
opentelemetry-otlp = "0.30.0"
opentelemetry-semantic-conventions = "0.30.0"
opentelemetry_sdk = "0.30.0"
os_info = "3.12.0"
owo-colors = "4.2.0"
paste = "1.0.15"
path-absolutize = "3.1.1"
path-clean = "1.0.1"
pathdiff = "0.2"
@@ -163,11 +141,9 @@ tokio-test = "0.4"
tokio-util = "0.7.16"
toml = "0.9.5"
toml_edit = "0.23.4"
tonic = "0.13.1"
tracing = "0.1.41"
tracing-appender = "0.2.3"
tracing-subscriber = "0.3.20"
tracing-test = "0.2.5"
tree-sitter = "0.25.9"
tree-sitter-bash = "0.25.0"
ts-rs = "11"

View File

@@ -25,14 +25,12 @@ Codex supports a rich set of configuration options. Note that the Rust CLI uses
Codex CLI functions as an MCP client that can connect to MCP servers on startup. See the [`mcp_servers`](../docs/config.md#mcp_servers) section in the configuration documentation for details.
It is still experimental, but you can also launch Codex as an MCP _server_ by running `codex mcp-server`. Use the [`@modelcontextprotocol/inspector`](https://github.com/modelcontextprotocol/inspector) to try it out:
It is still experimental, but you can also launch Codex as an MCP _server_ by running `codex mcp`. Use the [`@modelcontextprotocol/inspector`](https://github.com/modelcontextprotocol/inspector) to try it out:
```shell
npx @modelcontextprotocol/inspector codex mcp-server
npx @modelcontextprotocol/inspector codex mcp
```
Use `codex mcp` to add/list/get/remove MCP server launchers defined in `config.toml`, and `codex mcp-server` to run the MCP server directly.
### Notifications
You can enable notifications by configuring a script that is run whenever the agent finishes a turn. The [notify documentation](../docs/config.md#notify) includes a detailed example that explains how to get desktop notifications via [terminal-notifier](https://github.com/julienXX/terminal-notifier) on macOS.

View File

@@ -1,24 +0,0 @@
[package]
edition = "2024"
name = "codex-app-server-protocol"
version = { workspace = true }
[lib]
name = "codex_app_server_protocol"
path = "src/lib.rs"
[lints]
workspace = true
[dependencies]
codex-protocol = { workspace = true }
paste = { workspace = true }
serde = { workspace = true, features = ["derive"] }
serde_json = { workspace = true }
strum_macros = { workspace = true }
ts-rs = { workspace = true }
uuid = { workspace = true, features = ["serde", "v7"] }
[dev-dependencies]
anyhow = { workspace = true }
pretty_assertions = { workspace = true }

View File

@@ -1,67 +0,0 @@
//! We do not do true JSON-RPC 2.0, as we neither send nor expect the
//! "jsonrpc": "2.0" field.
use serde::Deserialize;
use serde::Serialize;
use ts_rs::TS;
pub const JSONRPC_VERSION: &str = "2.0";
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, Hash, Eq, TS)]
#[serde(untagged)]
pub enum RequestId {
String(String),
#[ts(type = "number")]
Integer(i64),
}
pub type Result = serde_json::Value;
/// Refers to any valid JSON-RPC object that can be decoded off the wire, or encoded to be sent.
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, TS)]
#[serde(untagged)]
pub enum JSONRPCMessage {
Request(JSONRPCRequest),
Notification(JSONRPCNotification),
Response(JSONRPCResponse),
Error(JSONRPCError),
}
/// A request that expects a response.
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, TS)]
pub struct JSONRPCRequest {
pub id: RequestId,
pub method: String,
#[serde(default, skip_serializing_if = "Option::is_none")]
pub params: Option<serde_json::Value>,
}
/// A notification which does not expect a response.
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, TS)]
pub struct JSONRPCNotification {
pub method: String,
#[serde(default, skip_serializing_if = "Option::is_none")]
pub params: Option<serde_json::Value>,
}
/// A successful (non-error) response to a request.
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, TS)]
pub struct JSONRPCResponse {
pub id: RequestId,
pub result: Result,
}
/// A response to a request that indicates an error occurred.
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, TS)]
pub struct JSONRPCError {
pub error: JSONRPCErrorError,
pub id: RequestId,
}
#[derive(Debug, Clone, PartialEq, Deserialize, Serialize, TS)]
pub struct JSONRPCErrorError {
pub code: i64,
#[serde(default, skip_serializing_if = "Option::is_none")]
pub data: Option<serde_json::Value>,
pub message: String,
}

View File

@@ -1,5 +0,0 @@
mod jsonrpc_lite;
mod protocol;
pub use jsonrpc_lite::*;
pub use protocol::*;

View File

@@ -1,49 +0,0 @@
[package]
edition = "2024"
name = "codex-app-server"
version = { workspace = true }
[[bin]]
name = "codex-app-server"
path = "src/main.rs"
[lib]
name = "codex_app_server"
path = "src/lib.rs"
[lints]
workspace = true
[dependencies]
anyhow = { workspace = true }
codex-arg0 = { workspace = true }
codex-common = { workspace = true, features = ["cli"] }
codex-core = { workspace = true }
codex-file-search = { workspace = true }
codex-login = { workspace = true }
codex-protocol = { workspace = true }
codex-app-server-protocol = { workspace = true }
codex-utils-json-to-toml = { workspace = true }
serde = { workspace = true, features = ["derive"] }
serde_json = { workspace = true }
tokio = { workspace = true, features = [
"io-std",
"macros",
"process",
"rt-multi-thread",
"signal",
] }
tracing = { workspace = true, features = ["log"] }
tracing-subscriber = { workspace = true, features = ["env-filter", "fmt"] }
uuid = { workspace = true, features = ["serde", "v7"] }
[dev-dependencies]
app_test_support = { workspace = true }
assert_cmd = { workspace = true }
base64 = { workspace = true }
core_test_support = { workspace = true }
os_info = { workspace = true }
pretty_assertions = { workspace = true }
tempfile = { workspace = true }
toml = { workspace = true }
wiremock = { workspace = true }

View File

@@ -1,2 +0,0 @@
pub(crate) const INVALID_REQUEST_ERROR_CODE: i64 = -32600;
pub(crate) const INTERNAL_ERROR_CODE: i64 = -32603;

View File

@@ -1,84 +0,0 @@
use std::num::NonZero;
use std::num::NonZeroUsize;
use std::path::PathBuf;
use std::sync::Arc;
use std::sync::atomic::AtomicBool;
use codex_app_server_protocol::FuzzyFileSearchResult;
use codex_file_search as file_search;
use tokio::task::JoinSet;
use tracing::warn;
const LIMIT_PER_ROOT: usize = 50;
const MAX_THREADS: usize = 12;
const COMPUTE_INDICES: bool = true;
pub(crate) async fn run_fuzzy_file_search(
query: String,
roots: Vec<String>,
cancellation_flag: Arc<AtomicBool>,
) -> Vec<FuzzyFileSearchResult> {
#[expect(clippy::expect_used)]
let limit_per_root =
NonZero::new(LIMIT_PER_ROOT).expect("LIMIT_PER_ROOT should be a valid non-zero usize");
let cores = std::thread::available_parallelism()
.map(std::num::NonZero::get)
.unwrap_or(1);
let threads = cores.min(MAX_THREADS);
let threads_per_root = (threads / roots.len()).max(1);
let threads = NonZero::new(threads_per_root).unwrap_or(NonZeroUsize::MIN);
let mut files: Vec<FuzzyFileSearchResult> = Vec::new();
let mut join_set = JoinSet::new();
for root in roots {
let search_dir = PathBuf::from(&root);
let query = query.clone();
let cancel_flag = cancellation_flag.clone();
join_set.spawn_blocking(move || {
match file_search::run(
query.as_str(),
limit_per_root,
&search_dir,
Vec::new(),
threads,
cancel_flag,
COMPUTE_INDICES,
) {
Ok(res) => Ok((root, res)),
Err(err) => Err((root, err)),
}
});
}
while let Some(res) = join_set.join_next().await {
match res {
Ok(Ok((root, res))) => {
for m in res.matches {
let result = FuzzyFileSearchResult {
root: root.clone(),
path: m.path,
score: m.score,
indices: m.indices,
};
files.push(result);
}
}
Ok(Err((root, err))) => {
warn!("fuzzy-file-search in dir '{root}' failed: {err}");
}
Err(err) => {
warn!("fuzzy-file-search join_next failed: {err}");
}
}
}
files.sort_by(file_search::cmp_by_score_desc_then_path_asc::<
FuzzyFileSearchResult,
_,
_,
>(|f| f.score, |f| f.path.as_str()));
files
}

View File

@@ -1,139 +0,0 @@
#![deny(clippy::print_stdout, clippy::print_stderr)]
use std::io::ErrorKind;
use std::io::Result as IoResult;
use std::path::PathBuf;
use codex_common::CliConfigOverrides;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
use codex_app_server_protocol::JSONRPCMessage;
use tokio::io::AsyncBufReadExt;
use tokio::io::AsyncWriteExt;
use tokio::io::BufReader;
use tokio::io::{self};
use tokio::sync::mpsc;
use tracing::debug;
use tracing::error;
use tracing::info;
use tracing_subscriber::EnvFilter;
use crate::message_processor::MessageProcessor;
use crate::outgoing_message::OutgoingMessage;
use crate::outgoing_message::OutgoingMessageSender;
mod codex_message_processor;
mod error_code;
mod fuzzy_file_search;
mod message_processor;
mod outgoing_message;
/// Size of the bounded channels used to communicate between tasks. The value
/// is a balance between throughput and memory usage 128 messages should be
/// plenty for an interactive CLI.
const CHANNEL_CAPACITY: usize = 128;
pub async fn run_main(
codex_linux_sandbox_exe: Option<PathBuf>,
cli_config_overrides: CliConfigOverrides,
) -> IoResult<()> {
// Install a simple subscriber so `tracing` output is visible. Users can
// control the log level with `RUST_LOG`.
tracing_subscriber::fmt()
.with_writer(std::io::stderr)
.with_env_filter(EnvFilter::from_default_env())
.init();
// Set up channels.
let (incoming_tx, mut incoming_rx) = mpsc::channel::<JSONRPCMessage>(CHANNEL_CAPACITY);
let (outgoing_tx, mut outgoing_rx) = mpsc::unbounded_channel::<OutgoingMessage>();
// Task: read from stdin, push to `incoming_tx`.
let stdin_reader_handle = tokio::spawn({
async move {
let stdin = io::stdin();
let reader = BufReader::new(stdin);
let mut lines = reader.lines();
while let Some(line) = lines.next_line().await.unwrap_or_default() {
match serde_json::from_str::<JSONRPCMessage>(&line) {
Ok(msg) => {
if incoming_tx.send(msg).await.is_err() {
// Receiver gone nothing left to do.
break;
}
}
Err(e) => error!("Failed to deserialize JSONRPCMessage: {e}"),
}
}
debug!("stdin reader finished (EOF)");
}
});
// Parse CLI overrides once and derive the base Config eagerly so later
// components do not need to work with raw TOML values.
let cli_kv_overrides = cli_config_overrides.parse_overrides().map_err(|e| {
std::io::Error::new(
ErrorKind::InvalidInput,
format!("error parsing -c overrides: {e}"),
)
})?;
let config = Config::load_with_cli_overrides(cli_kv_overrides, ConfigOverrides::default())
.map_err(|e| {
std::io::Error::new(ErrorKind::InvalidData, format!("error loading config: {e}"))
})?;
// Task: process incoming messages.
let processor_handle = tokio::spawn({
let outgoing_message_sender = OutgoingMessageSender::new(outgoing_tx);
let mut processor = MessageProcessor::new(
outgoing_message_sender,
codex_linux_sandbox_exe,
std::sync::Arc::new(config),
);
async move {
while let Some(msg) = incoming_rx.recv().await {
match msg {
JSONRPCMessage::Request(r) => processor.process_request(r).await,
JSONRPCMessage::Response(r) => processor.process_response(r).await,
JSONRPCMessage::Notification(n) => processor.process_notification(n).await,
JSONRPCMessage::Error(e) => processor.process_error(e),
}
}
info!("processor task exited (channel closed)");
}
});
// Task: write outgoing messages to stdout.
let stdout_writer_handle = tokio::spawn(async move {
let mut stdout = io::stdout();
while let Some(outgoing_message) = outgoing_rx.recv().await {
let Ok(value) = serde_json::to_value(outgoing_message) else {
error!("Failed to convert OutgoingMessage to JSON value");
continue;
};
match serde_json::to_string(&value) {
Ok(mut json) => {
json.push('\n');
if let Err(e) = stdout.write_all(json.as_bytes()).await {
error!("Failed to write to stdout: {e}");
break;
}
}
Err(e) => error!("Failed to serialize JSONRPCMessage: {e}"),
}
}
info!("stdout writer exited (channel closed)");
});
// Wait for all tasks to finish. The typical exit path is the stdin reader
// hitting EOF which, once it drops `incoming_tx`, propagates shutdown to
// the processor and then to the stdout task.
let _ = tokio::join!(stdin_reader_handle, processor_handle, stdout_writer_handle);
Ok(())
}

View File

@@ -1,10 +0,0 @@
use codex_app_server::run_main;
use codex_arg0::arg0_dispatch_or_else;
use codex_common::CliConfigOverrides;
fn main() -> anyhow::Result<()> {
arg0_dispatch_or_else(|codex_linux_sandbox_exe| async move {
run_main(codex_linux_sandbox_exe, CliConfigOverrides::default()).await?;
Ok(())
})
}

View File

@@ -1,133 +0,0 @@
use std::path::PathBuf;
use crate::codex_message_processor::CodexMessageProcessor;
use crate::error_code::INVALID_REQUEST_ERROR_CODE;
use crate::outgoing_message::OutgoingMessageSender;
use codex_app_server_protocol::ClientInfo;
use codex_app_server_protocol::ClientRequest;
use codex_app_server_protocol::InitializeResponse;
use codex_app_server_protocol::JSONRPCError;
use codex_app_server_protocol::JSONRPCErrorError;
use codex_app_server_protocol::JSONRPCNotification;
use codex_app_server_protocol::JSONRPCRequest;
use codex_app_server_protocol::JSONRPCResponse;
use codex_core::AuthManager;
use codex_core::ConversationManager;
use codex_core::config::Config;
use codex_core::default_client::USER_AGENT_SUFFIX;
use codex_core::default_client::get_codex_user_agent;
use std::sync::Arc;
pub(crate) struct MessageProcessor {
outgoing: Arc<OutgoingMessageSender>,
codex_message_processor: CodexMessageProcessor,
initialized: bool,
}
impl MessageProcessor {
/// Create a new `MessageProcessor`, retaining a handle to the outgoing
/// `Sender` so handlers can enqueue messages to be written to stdout.
pub(crate) fn new(
outgoing: OutgoingMessageSender,
codex_linux_sandbox_exe: Option<PathBuf>,
config: Arc<Config>,
) -> Self {
let outgoing = Arc::new(outgoing);
let auth_manager = AuthManager::shared(config.codex_home.clone());
let conversation_manager = Arc::new(ConversationManager::new(auth_manager.clone()));
let codex_message_processor = CodexMessageProcessor::new(
auth_manager,
conversation_manager,
outgoing.clone(),
codex_linux_sandbox_exe,
config,
);
Self {
outgoing,
codex_message_processor,
initialized: false,
}
}
pub(crate) async fn process_request(&mut self, request: JSONRPCRequest) {
let request_id = request.id.clone();
if let Ok(request_json) = serde_json::to_value(request)
&& let Ok(codex_request) = serde_json::from_value::<ClientRequest>(request_json)
{
match codex_request {
// Handle Initialize internally so CodexMessageProcessor does not have to concern
// itself with the `initialized` bool.
ClientRequest::Initialize { request_id, params } => {
if self.initialized {
let error = JSONRPCErrorError {
code: INVALID_REQUEST_ERROR_CODE,
message: "Already initialized".to_string(),
data: None,
};
self.outgoing.send_error(request_id, error).await;
return;
} else {
let ClientInfo {
name,
title: _title,
version,
} = params.client_info;
let user_agent_suffix = format!("{name}; {version}");
if let Ok(mut suffix) = USER_AGENT_SUFFIX.lock() {
*suffix = Some(user_agent_suffix);
}
let user_agent = get_codex_user_agent();
let response = InitializeResponse { user_agent };
self.outgoing.send_response(request_id, response).await;
self.initialized = true;
return;
}
}
_ => {
if !self.initialized {
let error = JSONRPCErrorError {
code: INVALID_REQUEST_ERROR_CODE,
message: "Not initialized".to_string(),
data: None,
};
self.outgoing.send_error(request_id, error).await;
return;
}
}
}
self.codex_message_processor
.process_request(codex_request)
.await;
} else {
let error = JSONRPCErrorError {
code: INVALID_REQUEST_ERROR_CODE,
message: "Invalid request".to_string(),
data: None,
};
self.outgoing.send_error(request_id, error).await;
}
}
pub(crate) async fn process_notification(&self, notification: JSONRPCNotification) {
// Currently, we do not expect to receive any notifications from the
// client, so we just log them.
tracing::info!("<- notification: {:?}", notification);
}
/// Handle a standalone JSON-RPC response originating from the peer.
pub(crate) async fn process_response(&mut self, response: JSONRPCResponse) {
tracing::info!("<- response: {:?}", response);
let JSONRPCResponse { id, result, .. } = response;
self.outgoing.notify_client_response(id, result).await
}
/// Handle an error object received from the peer.
pub(crate) fn process_error(&mut self, err: JSONRPCError) {
tracing::error!("<- error: {:?}", err);
}
}

View File

@@ -1,174 +0,0 @@
use std::collections::HashMap;
use std::sync::atomic::AtomicI64;
use std::sync::atomic::Ordering;
use codex_app_server_protocol::JSONRPCErrorError;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::Result;
use codex_app_server_protocol::ServerNotification;
use codex_app_server_protocol::ServerRequest;
use codex_app_server_protocol::ServerRequestPayload;
use serde::Serialize;
use tokio::sync::Mutex;
use tokio::sync::mpsc;
use tokio::sync::oneshot;
use tracing::warn;
use crate::error_code::INTERNAL_ERROR_CODE;
/// Sends messages to the client and manages request callbacks.
pub(crate) struct OutgoingMessageSender {
next_request_id: AtomicI64,
sender: mpsc::UnboundedSender<OutgoingMessage>,
request_id_to_callback: Mutex<HashMap<RequestId, oneshot::Sender<Result>>>,
}
impl OutgoingMessageSender {
pub(crate) fn new(sender: mpsc::UnboundedSender<OutgoingMessage>) -> Self {
Self {
next_request_id: AtomicI64::new(0),
sender,
request_id_to_callback: Mutex::new(HashMap::new()),
}
}
pub(crate) async fn send_request(
&self,
request: ServerRequestPayload,
) -> oneshot::Receiver<Result> {
let id = RequestId::Integer(self.next_request_id.fetch_add(1, Ordering::Relaxed));
let outgoing_message_id = id.clone();
let (tx_approve, rx_approve) = oneshot::channel();
{
let mut request_id_to_callback = self.request_id_to_callback.lock().await;
request_id_to_callback.insert(id, tx_approve);
}
let outgoing_message =
OutgoingMessage::Request(request.request_with_id(outgoing_message_id));
let _ = self.sender.send(outgoing_message);
rx_approve
}
pub(crate) async fn notify_client_response(&self, id: RequestId, result: Result) {
let entry = {
let mut request_id_to_callback = self.request_id_to_callback.lock().await;
request_id_to_callback.remove_entry(&id)
};
match entry {
Some((id, sender)) => {
if let Err(err) = sender.send(result) {
warn!("could not notify callback for {id:?} due to: {err:?}");
}
}
None => {
warn!("could not find callback for {id:?}");
}
}
}
pub(crate) async fn send_response<T: Serialize>(&self, id: RequestId, response: T) {
match serde_json::to_value(response) {
Ok(result) => {
let outgoing_message = OutgoingMessage::Response(OutgoingResponse { id, result });
let _ = self.sender.send(outgoing_message);
}
Err(err) => {
self.send_error(
id,
JSONRPCErrorError {
code: INTERNAL_ERROR_CODE,
message: format!("failed to serialize response: {err}"),
data: None,
},
)
.await;
}
}
}
pub(crate) async fn send_server_notification(&self, notification: ServerNotification) {
let _ = self
.sender
.send(OutgoingMessage::AppServerNotification(notification));
}
/// All notifications should be migrated to [`ServerNotification`] and
/// [`OutgoingMessage::Notification`] should be removed.
pub(crate) async fn send_notification(&self, notification: OutgoingNotification) {
let outgoing_message = OutgoingMessage::Notification(notification);
let _ = self.sender.send(outgoing_message);
}
pub(crate) async fn send_error(&self, id: RequestId, error: JSONRPCErrorError) {
let outgoing_message = OutgoingMessage::Error(OutgoingError { id, error });
let _ = self.sender.send(outgoing_message);
}
}
/// Outgoing message from the server to the client.
#[derive(Debug, Clone, Serialize)]
#[serde(untagged)]
pub(crate) enum OutgoingMessage {
Request(ServerRequest),
Notification(OutgoingNotification),
/// AppServerNotification is specific to the case where this is run as an
/// "app server" as opposed to an MCP server.
AppServerNotification(ServerNotification),
Response(OutgoingResponse),
Error(OutgoingError),
}
#[derive(Debug, Clone, PartialEq, Serialize)]
pub(crate) struct OutgoingNotification {
pub method: String,
#[serde(default, skip_serializing_if = "Option::is_none")]
pub params: Option<serde_json::Value>,
}
#[derive(Debug, Clone, PartialEq, Serialize)]
pub(crate) struct OutgoingResponse {
pub id: RequestId,
pub result: Result,
}
#[derive(Debug, Clone, PartialEq, Serialize)]
pub(crate) struct OutgoingError {
pub error: JSONRPCErrorError,
pub id: RequestId,
}
#[cfg(test)]
mod tests {
use codex_app_server_protocol::LoginChatGptCompleteNotification;
use pretty_assertions::assert_eq;
use serde_json::json;
use uuid::Uuid;
use super::*;
#[test]
fn verify_server_notification_serialization() {
let notification =
ServerNotification::LoginChatGptComplete(LoginChatGptCompleteNotification {
login_id: Uuid::nil(),
success: true,
error: None,
});
let jsonrpc_notification = OutgoingMessage::AppServerNotification(notification);
assert_eq!(
json!({
"method": "loginChatGptComplete",
"params": {
"loginId": Uuid::nil(),
"success": true,
},
}),
serde_json::to_value(jsonrpc_notification)
.expect("ensure the strum macros serialize the method field correctly"),
"ensure the strum macros serialize the method field correctly"
);
}
}

View File

@@ -1,3 +0,0 @@
// Single integration test binary that aggregates all test modules.
// The submodules live in `tests/suite/`.
mod suite;

View File

@@ -1,21 +0,0 @@
[package]
edition = "2024"
name = "app_test_support"
version = { workspace = true }
[lib]
path = "lib.rs"
[dependencies]
anyhow = { workspace = true }
assert_cmd = { workspace = true }
codex-app-server-protocol = { workspace = true }
serde = { workspace = true }
serde_json = { workspace = true }
tokio = { workspace = true, features = [
"io-std",
"macros",
"process",
"rt-multi-thread",
] }
wiremock = { workspace = true }

View File

@@ -1,17 +0,0 @@
mod mcp_process;
mod mock_model_server;
mod responses;
use codex_app_server_protocol::JSONRPCResponse;
pub use mcp_process::McpProcess;
pub use mock_model_server::create_mock_chat_completions_server;
pub use responses::create_apply_patch_sse_response;
pub use responses::create_final_assistant_message_sse_response;
pub use responses::create_shell_sse_response;
use serde::de::DeserializeOwned;
pub fn to_response<T: DeserializeOwned>(response: JSONRPCResponse) -> anyhow::Result<T> {
let value = serde_json::to_value(response.result)?;
let codex_response = serde_json::from_value(value)?;
Ok(codex_response)
}

View File

@@ -1,502 +0,0 @@
use std::collections::VecDeque;
use std::path::Path;
use std::process::Stdio;
use std::sync::atomic::AtomicI64;
use std::sync::atomic::Ordering;
use tokio::io::AsyncBufReadExt;
use tokio::io::AsyncWriteExt;
use tokio::io::BufReader;
use tokio::process::Child;
use tokio::process::ChildStdin;
use tokio::process::ChildStdout;
use anyhow::Context;
use assert_cmd::prelude::*;
use codex_app_server_protocol::AddConversationListenerParams;
use codex_app_server_protocol::ArchiveConversationParams;
use codex_app_server_protocol::CancelLoginChatGptParams;
use codex_app_server_protocol::ClientInfo;
use codex_app_server_protocol::ClientNotification;
use codex_app_server_protocol::GetAuthStatusParams;
use codex_app_server_protocol::InitializeParams;
use codex_app_server_protocol::InterruptConversationParams;
use codex_app_server_protocol::ListConversationsParams;
use codex_app_server_protocol::LoginApiKeyParams;
use codex_app_server_protocol::NewConversationParams;
use codex_app_server_protocol::RemoveConversationListenerParams;
use codex_app_server_protocol::ResumeConversationParams;
use codex_app_server_protocol::SendUserMessageParams;
use codex_app_server_protocol::SendUserTurnParams;
use codex_app_server_protocol::ServerRequest;
use codex_app_server_protocol::SetDefaultModelParams;
use codex_app_server_protocol::JSONRPCError;
use codex_app_server_protocol::JSONRPCMessage;
use codex_app_server_protocol::JSONRPCNotification;
use codex_app_server_protocol::JSONRPCRequest;
use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use std::process::Command as StdCommand;
use tokio::process::Command;
pub struct McpProcess {
next_request_id: AtomicI64,
/// Retain this child process until the client is dropped. The Tokio runtime
/// will make a "best effort" to reap the process after it exits, but it is
/// not a guarantee. See the `kill_on_drop` documentation for details.
#[allow(dead_code)]
process: Child,
stdin: ChildStdin,
stdout: BufReader<ChildStdout>,
pending_user_messages: VecDeque<JSONRPCNotification>,
}
impl McpProcess {
pub async fn new(codex_home: &Path) -> anyhow::Result<Self> {
Self::new_with_env(codex_home, &[]).await
}
/// Creates a new MCP process, allowing tests to override or remove
/// specific environment variables for the child process only.
///
/// Pass a tuple of (key, Some(value)) to set/override, or (key, None) to
/// remove a variable from the child's environment.
pub async fn new_with_env(
codex_home: &Path,
env_overrides: &[(&str, Option<&str>)],
) -> anyhow::Result<Self> {
// Use assert_cmd to locate the binary path and then switch to tokio::process::Command
let std_cmd = StdCommand::cargo_bin("codex-app-server")
.context("should find binary for codex-mcp-server")?;
let program = std_cmd.get_program().to_owned();
let mut cmd = Command::new(program);
cmd.stdin(Stdio::piped());
cmd.stdout(Stdio::piped());
cmd.stderr(Stdio::piped());
cmd.env("CODEX_HOME", codex_home);
cmd.env("RUST_LOG", "debug");
for (k, v) in env_overrides {
match v {
Some(val) => {
cmd.env(k, val);
}
None => {
cmd.env_remove(k);
}
}
}
let mut process = cmd
.kill_on_drop(true)
.spawn()
.context("codex-mcp-server proc should start")?;
let stdin = process
.stdin
.take()
.ok_or_else(|| anyhow::format_err!("mcp should have stdin fd"))?;
let stdout = process
.stdout
.take()
.ok_or_else(|| anyhow::format_err!("mcp should have stdout fd"))?;
let stdout = BufReader::new(stdout);
// Forward child's stderr to our stderr so failures are visible even
// when stdout/stderr are captured by the test harness.
if let Some(stderr) = process.stderr.take() {
let mut stderr_reader = BufReader::new(stderr).lines();
tokio::spawn(async move {
while let Ok(Some(line)) = stderr_reader.next_line().await {
eprintln!("[mcp stderr] {line}");
}
});
}
Ok(Self {
next_request_id: AtomicI64::new(0),
process,
stdin,
stdout,
pending_user_messages: VecDeque::new(),
})
}
/// Performs the initialization handshake with the MCP server.
pub async fn initialize(&mut self) -> anyhow::Result<()> {
let params = Some(serde_json::to_value(InitializeParams {
client_info: ClientInfo {
name: "codex-app-server-tests".to_string(),
title: None,
version: "0.1.0".to_string(),
},
})?);
let req_id = self.send_request("initialize", params).await?;
let initialized = self.read_jsonrpc_message().await?;
let JSONRPCMessage::Response(response) = initialized else {
unreachable!("expected JSONRPCMessage::Response for initialize, got {initialized:?}");
};
if response.id != RequestId::Integer(req_id) {
anyhow::bail!(
"initialize response id mismatch: expected {}, got {:?}",
req_id,
response.id
);
}
// Send notifications/initialized to ack the response.
self.send_notification(ClientNotification::Initialized)
.await?;
Ok(())
}
/// Send a `newConversation` JSON-RPC request.
pub async fn send_new_conversation_request(
&mut self,
params: NewConversationParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("newConversation", params).await
}
/// Send an `archiveConversation` JSON-RPC request.
pub async fn send_archive_conversation_request(
&mut self,
params: ArchiveConversationParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("archiveConversation", params).await
}
/// Send an `addConversationListener` JSON-RPC request.
pub async fn send_add_conversation_listener_request(
&mut self,
params: AddConversationListenerParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("addConversationListener", params).await
}
/// Send a `sendUserMessage` JSON-RPC request with a single text item.
pub async fn send_send_user_message_request(
&mut self,
params: SendUserMessageParams,
) -> anyhow::Result<i64> {
// Wire format expects variants in camelCase; text item uses external tagging.
let params = Some(serde_json::to_value(params)?);
self.send_request("sendUserMessage", params).await
}
/// Send a `removeConversationListener` JSON-RPC request.
pub async fn send_remove_conversation_listener_request(
&mut self,
params: RemoveConversationListenerParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("removeConversationListener", params)
.await
}
/// Send a `sendUserTurn` JSON-RPC request.
pub async fn send_send_user_turn_request(
&mut self,
params: SendUserTurnParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("sendUserTurn", params).await
}
/// Send a `interruptConversation` JSON-RPC request.
pub async fn send_interrupt_conversation_request(
&mut self,
params: InterruptConversationParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("interruptConversation", params).await
}
/// Send a `getAuthStatus` JSON-RPC request.
pub async fn send_get_auth_status_request(
&mut self,
params: GetAuthStatusParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("getAuthStatus", params).await
}
/// Send a `getUserSavedConfig` JSON-RPC request.
pub async fn send_get_user_saved_config_request(&mut self) -> anyhow::Result<i64> {
self.send_request("getUserSavedConfig", None).await
}
/// Send a `getUserAgent` JSON-RPC request.
pub async fn send_get_user_agent_request(&mut self) -> anyhow::Result<i64> {
self.send_request("getUserAgent", None).await
}
/// Send a `userInfo` JSON-RPC request.
pub async fn send_user_info_request(&mut self) -> anyhow::Result<i64> {
self.send_request("userInfo", None).await
}
/// Send a `setDefaultModel` JSON-RPC request.
pub async fn send_set_default_model_request(
&mut self,
params: SetDefaultModelParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("setDefaultModel", params).await
}
/// Send a `listConversations` JSON-RPC request.
pub async fn send_list_conversations_request(
&mut self,
params: ListConversationsParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("listConversations", params).await
}
/// Send a `resumeConversation` JSON-RPC request.
pub async fn send_resume_conversation_request(
&mut self,
params: ResumeConversationParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("resumeConversation", params).await
}
/// Send a `loginApiKey` JSON-RPC request.
pub async fn send_login_api_key_request(
&mut self,
params: LoginApiKeyParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("loginApiKey", params).await
}
/// Send a `loginChatGpt` JSON-RPC request.
pub async fn send_login_chat_gpt_request(&mut self) -> anyhow::Result<i64> {
self.send_request("loginChatGpt", None).await
}
/// Send a `cancelLoginChatGpt` JSON-RPC request.
pub async fn send_cancel_login_chat_gpt_request(
&mut self,
params: CancelLoginChatGptParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("cancelLoginChatGpt", params).await
}
/// Send a `logoutChatGpt` JSON-RPC request.
pub async fn send_logout_chat_gpt_request(&mut self) -> anyhow::Result<i64> {
self.send_request("logoutChatGpt", None).await
}
/// Send a `fuzzyFileSearch` JSON-RPC request.
pub async fn send_fuzzy_file_search_request(
&mut self,
query: &str,
roots: Vec<String>,
cancellation_token: Option<String>,
) -> anyhow::Result<i64> {
let mut params = serde_json::json!({
"query": query,
"roots": roots,
});
if let Some(token) = cancellation_token {
params["cancellationToken"] = serde_json::json!(token);
}
self.send_request("fuzzyFileSearch", Some(params)).await
}
async fn send_request(
&mut self,
method: &str,
params: Option<serde_json::Value>,
) -> anyhow::Result<i64> {
let request_id = self.next_request_id.fetch_add(1, Ordering::Relaxed);
let message = JSONRPCMessage::Request(JSONRPCRequest {
id: RequestId::Integer(request_id),
method: method.to_string(),
params,
});
self.send_jsonrpc_message(message).await?;
Ok(request_id)
}
pub async fn send_response(
&mut self,
id: RequestId,
result: serde_json::Value,
) -> anyhow::Result<()> {
self.send_jsonrpc_message(JSONRPCMessage::Response(JSONRPCResponse { id, result }))
.await
}
pub async fn send_notification(
&mut self,
notification: ClientNotification,
) -> anyhow::Result<()> {
let value = serde_json::to_value(notification)?;
self.send_jsonrpc_message(JSONRPCMessage::Notification(JSONRPCNotification {
method: value
.get("method")
.and_then(|m| m.as_str())
.ok_or_else(|| anyhow::format_err!("notification missing method field"))?
.to_string(),
params: value.get("params").cloned(),
}))
.await
}
async fn send_jsonrpc_message(&mut self, message: JSONRPCMessage) -> anyhow::Result<()> {
eprintln!("writing message to stdin: {message:?}");
let payload = serde_json::to_string(&message)?;
self.stdin.write_all(payload.as_bytes()).await?;
self.stdin.write_all(b"\n").await?;
self.stdin.flush().await?;
Ok(())
}
async fn read_jsonrpc_message(&mut self) -> anyhow::Result<JSONRPCMessage> {
let mut line = String::new();
self.stdout.read_line(&mut line).await?;
let message = serde_json::from_str::<JSONRPCMessage>(&line)?;
eprintln!("read message from stdout: {message:?}");
Ok(message)
}
pub async fn read_stream_until_request_message(&mut self) -> anyhow::Result<ServerRequest> {
eprintln!("in read_stream_until_request_message()");
loop {
let message = self.read_jsonrpc_message().await?;
match message {
JSONRPCMessage::Notification(notification) => {
eprintln!("notification: {notification:?}");
self.enqueue_user_message(notification);
}
JSONRPCMessage::Request(jsonrpc_request) => {
return jsonrpc_request.try_into().with_context(
|| "failed to deserialize ServerRequest from JSONRPCRequest",
);
}
JSONRPCMessage::Error(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Error: {message:?}");
}
JSONRPCMessage::Response(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Response: {message:?}");
}
}
}
}
pub async fn read_stream_until_response_message(
&mut self,
request_id: RequestId,
) -> anyhow::Result<JSONRPCResponse> {
eprintln!("in read_stream_until_response_message({request_id:?})");
loop {
let message = self.read_jsonrpc_message().await?;
match message {
JSONRPCMessage::Notification(notification) => {
eprintln!("notification: {notification:?}");
self.enqueue_user_message(notification);
}
JSONRPCMessage::Request(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Request: {message:?}");
}
JSONRPCMessage::Error(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Error: {message:?}");
}
JSONRPCMessage::Response(jsonrpc_response) => {
if jsonrpc_response.id == request_id {
return Ok(jsonrpc_response);
}
}
}
}
}
pub async fn read_stream_until_error_message(
&mut self,
request_id: RequestId,
) -> anyhow::Result<JSONRPCError> {
loop {
let message = self.read_jsonrpc_message().await?;
match message {
JSONRPCMessage::Notification(notification) => {
eprintln!("notification: {notification:?}");
self.enqueue_user_message(notification);
}
JSONRPCMessage::Request(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Request: {message:?}");
}
JSONRPCMessage::Response(_) => {
// Keep scanning; we're waiting for an error with matching id.
}
JSONRPCMessage::Error(err) => {
if err.id == request_id {
return Ok(err);
}
}
}
}
}
pub async fn read_stream_until_notification_message(
&mut self,
method: &str,
) -> anyhow::Result<JSONRPCNotification> {
eprintln!("in read_stream_until_notification_message({method})");
if let Some(notification) = self.take_pending_notification_by_method(method) {
return Ok(notification);
}
loop {
let message = self.read_jsonrpc_message().await?;
match message {
JSONRPCMessage::Notification(notification) => {
if notification.method == method {
return Ok(notification);
}
self.enqueue_user_message(notification);
}
JSONRPCMessage::Request(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Request: {message:?}");
}
JSONRPCMessage::Error(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Error: {message:?}");
}
JSONRPCMessage::Response(_) => {
anyhow::bail!("unexpected JSONRPCMessage::Response: {message:?}");
}
}
}
}
fn take_pending_notification_by_method(&mut self, method: &str) -> Option<JSONRPCNotification> {
if let Some(pos) = self
.pending_user_messages
.iter()
.position(|notification| notification.method == method)
{
return self.pending_user_messages.remove(pos);
}
None
}
fn enqueue_user_message(&mut self, notification: JSONRPCNotification) {
if notification.method == "codex/event/user_message" {
self.pending_user_messages.push_back(notification);
}
}
}

View File

@@ -1,47 +0,0 @@
use std::sync::atomic::AtomicUsize;
use std::sync::atomic::Ordering;
use wiremock::Mock;
use wiremock::MockServer;
use wiremock::Respond;
use wiremock::ResponseTemplate;
use wiremock::matchers::method;
use wiremock::matchers::path;
/// Create a mock server that will provide the responses, in order, for
/// requests to the `/v1/chat/completions` endpoint.
pub async fn create_mock_chat_completions_server(responses: Vec<String>) -> MockServer {
let server = MockServer::start().await;
let num_calls = responses.len();
let seq_responder = SeqResponder {
num_calls: AtomicUsize::new(0),
responses,
};
Mock::given(method("POST"))
.and(path("/v1/chat/completions"))
.respond_with(seq_responder)
.expect(num_calls as u64)
.mount(&server)
.await;
server
}
struct SeqResponder {
num_calls: AtomicUsize,
responses: Vec<String>,
}
impl Respond for SeqResponder {
fn respond(&self, _: &wiremock::Request) -> ResponseTemplate {
let call_num = self.num_calls.fetch_add(1, Ordering::SeqCst);
match self.responses.get(call_num) {
Some(response) => ResponseTemplate::new(200)
.insert_header("content-type", "text/event-stream")
.set_body_raw(response.clone(), "text/event-stream"),
None => panic!("no response for {call_num}"),
}
}
}

View File

@@ -1,95 +0,0 @@
use serde_json::json;
use std::path::Path;
pub fn create_shell_sse_response(
command: Vec<String>,
workdir: Option<&Path>,
timeout_ms: Option<u64>,
call_id: &str,
) -> anyhow::Result<String> {
// The `arguments`` for the `shell` tool is a serialized JSON object.
let tool_call_arguments = serde_json::to_string(&json!({
"command": command,
"workdir": workdir.map(|w| w.to_string_lossy()),
"timeout": timeout_ms
}))?;
let tool_call = json!({
"choices": [
{
"delta": {
"tool_calls": [
{
"id": call_id,
"function": {
"name": "shell",
"arguments": tool_call_arguments
}
}
]
},
"finish_reason": "tool_calls"
}
]
});
let sse = format!(
"data: {}\n\ndata: DONE\n\n",
serde_json::to_string(&tool_call)?
);
Ok(sse)
}
pub fn create_final_assistant_message_sse_response(message: &str) -> anyhow::Result<String> {
let assistant_message = json!({
"choices": [
{
"delta": {
"content": message
},
"finish_reason": "stop"
}
]
});
let sse = format!(
"data: {}\n\ndata: DONE\n\n",
serde_json::to_string(&assistant_message)?
);
Ok(sse)
}
pub fn create_apply_patch_sse_response(
patch_content: &str,
call_id: &str,
) -> anyhow::Result<String> {
// Use shell command to call apply_patch with heredoc format
let shell_command = format!("apply_patch <<'EOF'\n{patch_content}\nEOF");
let tool_call_arguments = serde_json::to_string(&json!({
"command": ["bash", "-lc", shell_command]
}))?;
let tool_call = json!({
"choices": [
{
"delta": {
"tool_calls": [
{
"id": call_id,
"function": {
"name": "shell",
"arguments": tool_call_arguments
}
}
]
},
"finish_reason": "tool_calls"
}
]
});
let sse = format!(
"data: {}\n\ndata: DONE\n\n",
serde_json::to_string(&tool_call)?
);
Ok(sse)
}

View File

@@ -1,104 +0,0 @@
use app_test_support::McpProcess;
use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use pretty_assertions::assert_eq;
use serde_json::json;
use tempfile::TempDir;
use tokio::time::timeout;
const DEFAULT_READ_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(10);
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn test_fuzzy_file_search_sorts_and_includes_indices() {
// Prepare a temporary Codex home and a separate root with test files.
let codex_home = TempDir::new().expect("create temp codex home");
let root = TempDir::new().expect("create temp search root");
// Create files designed to have deterministic ordering for query "abc".
std::fs::write(root.path().join("abc"), "x").expect("write file abc");
std::fs::write(root.path().join("abcde"), "x").expect("write file abcx");
std::fs::write(root.path().join("abexy"), "x").expect("write file abcx");
std::fs::write(root.path().join("zzz.txt"), "x").expect("write file zzz");
// Start MCP server and initialize.
let mut mcp = McpProcess::new(codex_home.path()).await.expect("spawn mcp");
timeout(DEFAULT_READ_TIMEOUT, mcp.initialize())
.await
.expect("init timeout")
.expect("init failed");
let root_path = root.path().to_string_lossy().to_string();
// Send fuzzyFileSearch request.
let request_id = mcp
.send_fuzzy_file_search_request("abe", vec![root_path.clone()], None)
.await
.expect("send fuzzyFileSearch");
// Read response and verify shape and ordering.
let resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(request_id)),
)
.await
.expect("fuzzyFileSearch timeout")
.expect("fuzzyFileSearch resp");
let value = resp.result;
assert_eq!(
value,
json!({
"files": [
{ "root": root_path.clone(), "path": "abexy", "score": 88, "indices": [0, 1, 2] },
{ "root": root_path.clone(), "path": "abcde", "score": 74, "indices": [0, 1, 4] },
]
})
);
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn test_fuzzy_file_search_accepts_cancellation_token() {
let codex_home = TempDir::new().expect("create temp codex home");
let root = TempDir::new().expect("create temp search root");
std::fs::write(root.path().join("alpha.txt"), "contents").expect("write alpha");
let mut mcp = McpProcess::new(codex_home.path()).await.expect("spawn mcp");
timeout(DEFAULT_READ_TIMEOUT, mcp.initialize())
.await
.expect("init timeout")
.expect("init failed");
let root_path = root.path().to_string_lossy().to_string();
let request_id = mcp
.send_fuzzy_file_search_request("alp", vec![root_path.clone()], None)
.await
.expect("send fuzzyFileSearch");
let request_id_2 = mcp
.send_fuzzy_file_search_request(
"alp",
vec![root_path.clone()],
Some(request_id.to_string()),
)
.await
.expect("send fuzzyFileSearch");
let resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(request_id_2)),
)
.await
.expect("fuzzyFileSearch timeout")
.expect("fuzzyFileSearch resp");
let files = resp
.result
.get("files")
.and_then(|value| value.as_array())
.cloned()
.expect("files array");
assert_eq!(files.len(), 1);
assert_eq!(files[0]["root"], root_path);
assert_eq!(files[0]["path"], "alpha.txt");
}

View File

@@ -1,13 +0,0 @@
mod archive_conversation;
mod auth;
mod codex_message_processor_flow;
mod config;
mod create_conversation;
mod fuzzy_file_search;
mod interrupt;
mod list_resume;
mod login;
mod send_message;
mod set_default_model;
mod user_agent;
mod user_info;

View File

@@ -1,18 +0,0 @@
[package]
name = "codex-backend-client"
version = "0.0.0"
edition = "2024"
publish = false
[lib]
path = "src/lib.rs"
[dependencies]
anyhow = "1"
serde = { version = "1", features = ["derive"] }
serde_json = "1"
reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls"] }
codex-backend-openapi-models = { path = "../codex-backend-openapi-models" }
[dev-dependencies]
pretty_assertions = "1"

View File

@@ -1,244 +0,0 @@
use crate::types::CodeTaskDetailsResponse;
use crate::types::PaginatedListTaskListItem;
use crate::types::TurnAttemptsSiblingTurnsResponse;
use anyhow::Result;
use reqwest::header::AUTHORIZATION;
use reqwest::header::CONTENT_TYPE;
use reqwest::header::HeaderMap;
use reqwest::header::HeaderName;
use reqwest::header::HeaderValue;
use reqwest::header::USER_AGENT;
use serde::de::DeserializeOwned;
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum PathStyle {
/// /api/codex/…
CodexApi,
/// /wham/…
ChatGptApi,
}
impl PathStyle {
pub fn from_base_url(base_url: &str) -> Self {
if base_url.contains("/backend-api") {
PathStyle::ChatGptApi
} else {
PathStyle::CodexApi
}
}
}
#[derive(Clone, Debug)]
pub struct Client {
base_url: String,
http: reqwest::Client,
bearer_token: Option<String>,
user_agent: Option<HeaderValue>,
chatgpt_account_id: Option<String>,
path_style: PathStyle,
}
impl Client {
pub fn new(base_url: impl Into<String>) -> Result<Self> {
let mut base_url = base_url.into();
// Normalize common ChatGPT hostnames to include /backend-api so we hit the WHAM paths.
// Also trim trailing slashes for consistent URL building.
while base_url.ends_with('/') {
base_url.pop();
}
if (base_url.starts_with("https://chatgpt.com")
|| base_url.starts_with("https://chat.openai.com"))
&& !base_url.contains("/backend-api")
{
base_url = format!("{base_url}/backend-api");
}
let http = reqwest::Client::builder().build()?;
let path_style = PathStyle::from_base_url(&base_url);
Ok(Self {
base_url,
http,
bearer_token: None,
user_agent: None,
chatgpt_account_id: None,
path_style,
})
}
pub fn with_bearer_token(mut self, token: impl Into<String>) -> Self {
self.bearer_token = Some(token.into());
self
}
pub fn with_user_agent(mut self, ua: impl Into<String>) -> Self {
if let Ok(hv) = HeaderValue::from_str(&ua.into()) {
self.user_agent = Some(hv);
}
self
}
pub fn with_chatgpt_account_id(mut self, account_id: impl Into<String>) -> Self {
self.chatgpt_account_id = Some(account_id.into());
self
}
pub fn with_path_style(mut self, style: PathStyle) -> Self {
self.path_style = style;
self
}
fn headers(&self) -> HeaderMap {
let mut h = HeaderMap::new();
if let Some(ua) = &self.user_agent {
h.insert(USER_AGENT, ua.clone());
} else {
h.insert(USER_AGENT, HeaderValue::from_static("codex-cli"));
}
if let Some(token) = &self.bearer_token {
let value = format!("Bearer {token}");
if let Ok(hv) = HeaderValue::from_str(&value) {
h.insert(AUTHORIZATION, hv);
}
}
if let Some(acc) = &self.chatgpt_account_id
&& let Ok(name) = HeaderName::from_bytes(b"ChatGPT-Account-Id")
&& let Ok(hv) = HeaderValue::from_str(acc)
{
h.insert(name, hv);
}
h
}
async fn exec_request(
&self,
req: reqwest::RequestBuilder,
method: &str,
url: &str,
) -> Result<(String, String)> {
let res = req.send().await?;
let status = res.status();
let ct = res
.headers()
.get(CONTENT_TYPE)
.and_then(|v| v.to_str().ok())
.unwrap_or("")
.to_string();
let body = res.text().await.unwrap_or_default();
if !status.is_success() {
anyhow::bail!("{method} {url} failed: {status}; content-type={ct}; body={body}");
}
Ok((body, ct))
}
fn decode_json<T: DeserializeOwned>(&self, url: &str, ct: &str, body: &str) -> Result<T> {
match serde_json::from_str::<T>(body) {
Ok(v) => Ok(v),
Err(e) => {
anyhow::bail!("Decode error for {url}: {e}; content-type={ct}; body={body}");
}
}
}
pub async fn list_tasks(
&self,
limit: Option<i32>,
task_filter: Option<&str>,
environment_id: Option<&str>,
) -> Result<PaginatedListTaskListItem> {
let url = match self.path_style {
PathStyle::CodexApi => format!("{}/api/codex/tasks/list", self.base_url),
PathStyle::ChatGptApi => format!("{}/wham/tasks/list", self.base_url),
};
let req = self.http.get(&url).headers(self.headers());
let req = if let Some(lim) = limit {
req.query(&[("limit", lim)])
} else {
req
};
let req = if let Some(tf) = task_filter {
req.query(&[("task_filter", tf)])
} else {
req
};
let req = if let Some(id) = environment_id {
req.query(&[("environment_id", id)])
} else {
req
};
let (body, ct) = self.exec_request(req, "GET", &url).await?;
self.decode_json::<PaginatedListTaskListItem>(&url, &ct, &body)
}
pub async fn get_task_details(&self, task_id: &str) -> Result<CodeTaskDetailsResponse> {
let (parsed, _body, _ct) = self.get_task_details_with_body(task_id).await?;
Ok(parsed)
}
pub async fn get_task_details_with_body(
&self,
task_id: &str,
) -> Result<(CodeTaskDetailsResponse, String, String)> {
let url = match self.path_style {
PathStyle::CodexApi => format!("{}/api/codex/tasks/{}", self.base_url, task_id),
PathStyle::ChatGptApi => format!("{}/wham/tasks/{}", self.base_url, task_id),
};
let req = self.http.get(&url).headers(self.headers());
let (body, ct) = self.exec_request(req, "GET", &url).await?;
let parsed: CodeTaskDetailsResponse = self.decode_json(&url, &ct, &body)?;
Ok((parsed, body, ct))
}
pub async fn list_sibling_turns(
&self,
task_id: &str,
turn_id: &str,
) -> Result<TurnAttemptsSiblingTurnsResponse> {
let url = match self.path_style {
PathStyle::CodexApi => format!(
"{}/api/codex/tasks/{}/turns/{}/sibling_turns",
self.base_url, task_id, turn_id
),
PathStyle::ChatGptApi => format!(
"{}/wham/tasks/{}/turns/{}/sibling_turns",
self.base_url, task_id, turn_id
),
};
let req = self.http.get(&url).headers(self.headers());
let (body, ct) = self.exec_request(req, "GET", &url).await?;
self.decode_json::<TurnAttemptsSiblingTurnsResponse>(&url, &ct, &body)
}
/// Create a new task (user turn) by POSTing to the appropriate backend path
/// based on `path_style`. Returns the created task id.
pub async fn create_task(&self, request_body: serde_json::Value) -> Result<String> {
let url = match self.path_style {
PathStyle::CodexApi => format!("{}/api/codex/tasks", self.base_url),
PathStyle::ChatGptApi => format!("{}/wham/tasks", self.base_url),
};
let req = self
.http
.post(&url)
.headers(self.headers())
.header(CONTENT_TYPE, HeaderValue::from_static("application/json"))
.json(&request_body);
let (body, ct) = self.exec_request(req, "POST", &url).await?;
// Extract id from JSON: prefer `task.id`; fallback to top-level `id` when present.
match serde_json::from_str::<serde_json::Value>(&body) {
Ok(v) => {
if let Some(id) = v
.get("task")
.and_then(|t| t.get("id"))
.and_then(|s| s.as_str())
{
Ok(id.to_string())
} else if let Some(id) = v.get("id").and_then(|s| s.as_str()) {
Ok(id.to_string())
} else {
anyhow::bail!(
"POST {url} succeeded but no task id found; content-type={ct}; body={body}"
);
}
}
Err(e) => anyhow::bail!("Decode error for {url}: {e}; content-type={ct}; body={body}"),
}
}
}

View File

@@ -1,9 +0,0 @@
mod client;
pub mod types;
pub use client::Client;
pub use types::CodeTaskDetailsResponse;
pub use types::CodeTaskDetailsResponseExt;
pub use types::PaginatedListTaskListItem;
pub use types::TaskListItem;
pub use types::TurnAttemptsSiblingTurnsResponse;

View File

@@ -1,369 +0,0 @@
pub use codex_backend_openapi_models::models::PaginatedListTaskListItem;
pub use codex_backend_openapi_models::models::TaskListItem;
use serde::Deserialize;
use serde::de::Deserializer;
use serde_json::Value;
use std::collections::HashMap;
/// Hand-rolled models for the Cloud Tasks task-details response.
/// The generated OpenAPI models are pretty bad. This is a half-step
/// towards hand-rolling them.
#[derive(Clone, Debug, Deserialize)]
pub struct CodeTaskDetailsResponse {
#[serde(default)]
pub current_user_turn: Option<Turn>,
#[serde(default)]
pub current_assistant_turn: Option<Turn>,
#[serde(default)]
pub current_diff_task_turn: Option<Turn>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct Turn {
#[serde(default)]
pub id: Option<String>,
#[serde(default)]
pub attempt_placement: Option<i64>,
#[serde(default, rename = "turn_status")]
pub turn_status: Option<String>,
#[serde(default, deserialize_with = "deserialize_vec")]
pub sibling_turn_ids: Vec<String>,
#[serde(default, deserialize_with = "deserialize_vec")]
pub input_items: Vec<TurnItem>,
#[serde(default, deserialize_with = "deserialize_vec")]
pub output_items: Vec<TurnItem>,
#[serde(default)]
pub worklog: Option<Worklog>,
#[serde(default)]
pub error: Option<TurnError>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct TurnItem {
#[serde(rename = "type", default)]
pub kind: String,
#[serde(default)]
pub role: Option<String>,
#[serde(default, deserialize_with = "deserialize_vec")]
pub content: Vec<ContentFragment>,
#[serde(default)]
pub diff: Option<String>,
#[serde(default)]
pub output_diff: Option<DiffPayload>,
}
#[derive(Clone, Debug, Deserialize)]
#[serde(untagged)]
pub enum ContentFragment {
Structured(StructuredContent),
Text(String),
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct StructuredContent {
#[serde(rename = "content_type", default)]
pub content_type: Option<String>,
#[serde(default)]
pub text: Option<String>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct DiffPayload {
#[serde(default)]
pub diff: Option<String>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct Worklog {
#[serde(default, deserialize_with = "deserialize_vec")]
pub messages: Vec<WorklogMessage>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct WorklogMessage {
#[serde(default)]
pub author: Option<Author>,
#[serde(default)]
pub content: Option<WorklogContent>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct Author {
#[serde(default)]
pub role: Option<String>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct WorklogContent {
#[serde(default)]
pub parts: Vec<ContentFragment>,
}
#[derive(Clone, Debug, Default, Deserialize)]
pub struct TurnError {
#[serde(default)]
pub code: Option<String>,
#[serde(default)]
pub message: Option<String>,
}
impl ContentFragment {
fn text(&self) -> Option<&str> {
match self {
ContentFragment::Structured(inner) => {
if inner
.content_type
.as_deref()
.map(|ct| ct.eq_ignore_ascii_case("text"))
.unwrap_or(false)
{
inner.text.as_deref().filter(|s| !s.is_empty())
} else {
None
}
}
ContentFragment::Text(raw) => {
if raw.trim().is_empty() {
None
} else {
Some(raw.as_str())
}
}
}
}
}
impl TurnItem {
fn text_values(&self) -> Vec<String> {
self.content
.iter()
.filter_map(|fragment| fragment.text().map(str::to_string))
.collect()
}
fn diff_text(&self) -> Option<String> {
if self.kind == "output_diff" {
if let Some(diff) = &self.diff
&& !diff.is_empty()
{
return Some(diff.clone());
}
} else if self.kind == "pr"
&& let Some(payload) = &self.output_diff
&& let Some(diff) = &payload.diff
&& !diff.is_empty()
{
return Some(diff.clone());
}
None
}
}
impl Turn {
fn unified_diff(&self) -> Option<String> {
self.output_items.iter().find_map(TurnItem::diff_text)
}
fn message_texts(&self) -> Vec<String> {
let mut out: Vec<String> = self
.output_items
.iter()
.filter(|item| item.kind == "message")
.flat_map(TurnItem::text_values)
.collect();
if let Some(log) = &self.worklog {
for message in &log.messages {
if message.is_assistant() {
out.extend(message.text_values());
}
}
}
out
}
fn user_prompt(&self) -> Option<String> {
let parts: Vec<String> = self
.input_items
.iter()
.filter(|item| item.kind == "message")
.filter(|item| {
item.role
.as_deref()
.map(|r| r.eq_ignore_ascii_case("user"))
.unwrap_or(true)
})
.flat_map(TurnItem::text_values)
.collect();
if parts.is_empty() {
None
} else {
Some(parts.join(
"
",
))
}
}
fn error_summary(&self) -> Option<String> {
self.error.as_ref().and_then(TurnError::summary)
}
}
impl WorklogMessage {
fn is_assistant(&self) -> bool {
self.author
.as_ref()
.and_then(|a| a.role.as_deref())
.map(|role| role.eq_ignore_ascii_case("assistant"))
.unwrap_or(false)
}
fn text_values(&self) -> Vec<String> {
self.content
.as_ref()
.map(|content| {
content
.parts
.iter()
.filter_map(|fragment| fragment.text().map(str::to_string))
.collect()
})
.unwrap_or_default()
}
}
impl TurnError {
fn summary(&self) -> Option<String> {
let code = self.code.as_deref().unwrap_or("");
let message = self.message.as_deref().unwrap_or("");
match (code.is_empty(), message.is_empty()) {
(true, true) => None,
(false, true) => Some(code.to_string()),
(true, false) => Some(message.to_string()),
(false, false) => Some(format!("{code}: {message}")),
}
}
}
pub trait CodeTaskDetailsResponseExt {
/// Attempt to extract a unified diff string from the assistant or diff turn.
fn unified_diff(&self) -> Option<String>;
/// Extract assistant text output messages (no diff) from current turns.
fn assistant_text_messages(&self) -> Vec<String>;
/// Extract the user's prompt text from the current user turn, when present.
fn user_text_prompt(&self) -> Option<String>;
/// Extract an assistant error message (if the turn failed and provided one).
fn assistant_error_message(&self) -> Option<String>;
}
impl CodeTaskDetailsResponseExt for CodeTaskDetailsResponse {
fn unified_diff(&self) -> Option<String> {
[
self.current_diff_task_turn.as_ref(),
self.current_assistant_turn.as_ref(),
]
.into_iter()
.flatten()
.find_map(Turn::unified_diff)
}
fn assistant_text_messages(&self) -> Vec<String> {
let mut out = Vec::new();
for turn in [
self.current_diff_task_turn.as_ref(),
self.current_assistant_turn.as_ref(),
]
.into_iter()
.flatten()
{
out.extend(turn.message_texts());
}
out
}
fn user_text_prompt(&self) -> Option<String> {
self.current_user_turn.as_ref().and_then(Turn::user_prompt)
}
fn assistant_error_message(&self) -> Option<String> {
self.current_assistant_turn
.as_ref()
.and_then(Turn::error_summary)
}
}
fn deserialize_vec<'de, D, T>(deserializer: D) -> Result<Vec<T>, D::Error>
where
D: Deserializer<'de>,
T: Deserialize<'de>,
{
Option::<Vec<T>>::deserialize(deserializer).map(|opt| opt.unwrap_or_default())
}
#[derive(Clone, Debug, Deserialize)]
pub struct TurnAttemptsSiblingTurnsResponse {
#[serde(default)]
pub sibling_turns: Vec<HashMap<String, Value>>,
}
#[cfg(test)]
mod tests {
use super::*;
use pretty_assertions::assert_eq;
fn fixture(name: &str) -> CodeTaskDetailsResponse {
let json = match name {
"diff" => include_str!("../tests/fixtures/task_details_with_diff.json"),
"error" => include_str!("../tests/fixtures/task_details_with_error.json"),
other => panic!("unknown fixture {other}"),
};
serde_json::from_str(json).expect("fixture should deserialize")
}
#[test]
fn unified_diff_prefers_current_diff_task_turn() {
let details = fixture("diff");
let diff = details.unified_diff().expect("diff present");
assert!(diff.contains("diff --git"));
}
#[test]
fn unified_diff_falls_back_to_pr_output_diff() {
let details = fixture("error");
let diff = details.unified_diff().expect("diff from pr output");
assert!(diff.contains("lib.rs"));
}
#[test]
fn assistant_text_messages_extracts_text_content() {
let details = fixture("diff");
let messages = details.assistant_text_messages();
assert_eq!(messages, vec!["Assistant response".to_string()]);
}
#[test]
fn user_text_prompt_joins_parts_with_spacing() {
let details = fixture("diff");
let prompt = details.user_text_prompt().expect("prompt present");
assert_eq!(
prompt,
"First line
Second line"
);
}
#[test]
fn assistant_error_message_combines_code_and_message() {
let details = fixture("error");
let msg = details
.assistant_error_message()
.expect("error should be present");
assert_eq!(msg, "APPLY_FAILED: Patch could not be applied");
}
}

View File

@@ -1,38 +0,0 @@
{
"task": {
"id": "task_123",
"title": "Refactor cloud task client",
"archived": false,
"external_pull_requests": []
},
"current_user_turn": {
"input_items": [
{
"type": "message",
"role": "user",
"content": [
{ "content_type": "text", "text": "First line" },
{ "content_type": "text", "text": "Second line" }
]
}
]
},
"current_assistant_turn": {
"output_items": [
{
"type": "message",
"content": [
{ "content_type": "text", "text": "Assistant response" }
]
}
]
},
"current_diff_task_turn": {
"output_items": [
{
"type": "output_diff",
"diff": "diff --git a/src/main.rs b/src/main.rs\n+fn main() { println!(\"hi\"); }\n"
}
]
}
}

View File

@@ -1,22 +0,0 @@
{
"task": {
"id": "task_456",
"title": "Investigate failure",
"archived": false,
"external_pull_requests": []
},
"current_assistant_turn": {
"output_items": [
{
"type": "pr",
"output_diff": {
"diff": "diff --git a/lib.rs b/lib.rs\n+pub fn hello() {}\n"
}
}
],
"error": {
"code": "APPLY_FAILED",
"message": "Patch could not be applied"
}
}
}

View File

@@ -14,7 +14,6 @@ codex-core = { workspace = true }
serde = { workspace = true, features = ["derive"] }
serde_json = { workspace = true }
tokio = { workspace = true, features = ["full"] }
codex-git-apply = { path = "../git-apply" }
[dev-dependencies]
tempfile = { workspace = true }

View File

@@ -56,24 +56,46 @@ pub async fn apply_diff_from_task(
}
async fn apply_diff(diff: &str, cwd: Option<PathBuf>) -> anyhow::Result<()> {
let cwd = cwd.unwrap_or(std::env::current_dir().unwrap_or_else(|_| std::env::temp_dir()));
let req = codex_git_apply::ApplyGitRequest {
cwd,
diff: diff.to_string(),
revert: false,
preflight: false,
};
let res = codex_git_apply::apply_git_patch(&req)?;
if res.exit_code != 0 {
let mut cmd = tokio::process::Command::new("git");
if let Some(cwd) = cwd {
cmd.current_dir(cwd);
}
let toplevel_output = cmd
.args(vec!["rev-parse", "--show-toplevel"])
.output()
.await?;
if !toplevel_output.status.success() {
anyhow::bail!("apply must be run from a git repository.");
}
let repo_root = String::from_utf8(toplevel_output.stdout)?
.trim()
.to_string();
let mut git_apply_cmd = tokio::process::Command::new("git")
.args(vec!["apply", "--3way"])
.current_dir(&repo_root)
.stdin(std::process::Stdio::piped())
.stdout(std::process::Stdio::piped())
.stderr(std::process::Stdio::piped())
.spawn()?;
if let Some(mut stdin) = git_apply_cmd.stdin.take() {
tokio::io::AsyncWriteExt::write_all(&mut stdin, diff.as_bytes()).await?;
drop(stdin);
}
let output = git_apply_cmd.wait_with_output().await?;
if !output.status.success() {
anyhow::bail!(
"Git apply failed (applied={}, skipped={}, conflicts={})\nstdout:\n{}\nstderr:\n{}",
res.applied_paths.len(),
res.skipped_paths.len(),
res.conflicted_paths.len(),
res.stdout,
res.stderr
"Git apply failed with status {}: {}",
output.status,
String::from_utf8_lossy(&output.stderr)
);
}
println!("Successfully applied diff");
Ok(())
}

View File

@@ -44,6 +44,6 @@ pub(crate) async fn chatgpt_get_request<T: DeserializeOwned>(
} else {
let status = response.status();
let body = response.text().await.unwrap_or_default();
anyhow::bail!("Request failed with status {status}: {body}")
anyhow::bail!("Request failed with status {}: {}", status, body)
}
}

View File

@@ -18,7 +18,6 @@ workspace = true
anyhow = { workspace = true }
clap = { workspace = true, features = ["derive"] }
clap_complete = { workspace = true }
codex-app-server = { workspace = true }
codex-arg0 = { workspace = true }
codex-chatgpt = { workspace = true }
codex-common = { workspace = true, features = ["cli"] }
@@ -28,11 +27,8 @@ codex-login = { workspace = true }
codex-mcp-server = { workspace = true }
codex-process-hardening = { workspace = true }
codex-protocol = { workspace = true }
codex-app-server-protocol = { workspace = true }
codex-protocol-ts = { workspace = true }
codex-responses-api-proxy = { workspace = true }
codex-tui = { workspace = true }
codex-cloud-tasks = { path = "../cloud-tasks" }
ctor = { workspace = true }
owo-colors = { workspace = true }
serde_json = { workspace = true }
@@ -44,6 +40,8 @@ tokio = { workspace = true, features = [
"rt-multi-thread",
"signal",
] }
tracing = { workspace = true }
tracing-subscriber = { workspace = true }
[dev-dependencies]
assert_cmd = { workspace = true }

View File

@@ -1,6 +1,7 @@
pub mod debug_sandbox;
mod exit_status;
pub mod login;
pub mod proto;
use clap::Parser;
use codex_common::CliConfigOverrides;

View File

@@ -1,4 +1,3 @@
use codex_app_server_protocol::AuthMode;
use codex_common::CliConfigOverrides;
use codex_core::CodexAuth;
use codex_core::auth::CLIENT_ID;
@@ -7,8 +6,8 @@ use codex_core::auth::logout;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
use codex_login::ServerOptions;
use codex_login::run_device_code_login;
use codex_login::run_login_server;
use codex_protocol::mcp_protocol::AuthMode;
use std::path::PathBuf;
pub async fn login_with_chatgpt(codex_home: PathBuf) -> std::io::Result<()> {
@@ -56,32 +55,6 @@ pub async fn run_login_with_api_key(
}
}
/// Login using the OAuth device code flow.
pub async fn run_login_with_device_code(
cli_config_overrides: CliConfigOverrides,
issuer_base_url: Option<String>,
client_id: Option<String>,
) -> ! {
let config = load_config_or_exit(cli_config_overrides);
let mut opts = ServerOptions::new(
config.codex_home,
client_id.unwrap_or(CLIENT_ID.to_string()),
);
if let Some(iss) = issuer_base_url {
opts.issuer = iss;
}
match run_device_code_login(opts).await {
Ok(()) => {
eprintln!("Successfully logged in");
std::process::exit(0);
}
Err(e) => {
eprintln!("Error logging in with device code: {e}");
std::process::exit(1);
}
}
}
pub async fn run_login_status(cli_config_overrides: CliConfigOverrides) -> ! {
let config = load_config_or_exit(cli_config_overrides);

View File

@@ -10,12 +10,10 @@ use codex_cli::SeatbeltCommand;
use codex_cli::login::run_login_status;
use codex_cli::login::run_login_with_api_key;
use codex_cli::login::run_login_with_chatgpt;
use codex_cli::login::run_login_with_device_code;
use codex_cli::login::run_logout;
use codex_cloud_tasks::Cli as CloudTasksCli;
use codex_cli::proto;
use codex_common::CliConfigOverrides;
use codex_exec::Cli as ExecCli;
use codex_responses_api_proxy::Args as ResponsesApiProxyArgs;
use codex_tui::AppExitInfo;
use codex_tui::Cli as TuiCli;
use owo_colors::OwoColorize;
@@ -25,6 +23,7 @@ use supports_color::Stream;
mod mcp_cmd;
use crate::mcp_cmd::McpCli;
use crate::proto::ProtoCli;
/// Codex CLI
///
@@ -66,11 +65,9 @@ enum Subcommand {
/// [experimental] Run Codex as an MCP server and manage MCP servers.
Mcp(McpCli),
/// [experimental] Run the Codex MCP server (stdio transport).
McpServer,
/// [experimental] Run the app server.
AppServer,
/// Run the Protocol stream via stdin/stdout
#[clap(visible_alias = "p")]
Proto(ProtoCli),
/// Generate shell completion scripts.
Completion(CompletionCommand),
@@ -88,13 +85,6 @@ enum Subcommand {
/// Internal: generate TypeScript protocol bindings.
#[clap(hide = true)]
GenerateTs(GenerateTsCommand),
/// [EXPERIMENTAL] Browse tasks from Codex Cloud and apply changes locally.
#[clap(name = "cloud", alias = "cloud-tasks")]
Cloud(CloudTasksCli),
/// Internal: run the responses API proxy.
#[clap(hide = true)]
ResponsesApiProxy(ResponsesApiProxyArgs),
}
#[derive(Debug, Parser)]
@@ -142,20 +132,6 @@ struct LoginCommand {
#[arg(long = "api-key", value_name = "API_KEY")]
api_key: Option<String>,
/// EXPERIMENTAL: Use device code flow (not yet supported)
/// This feature is experimental and may changed in future releases.
#[arg(long = "experimental_use-device-code", hide = true)]
use_device_code: bool,
/// EXPERIMENTAL: Use custom OAuth issuer base URL (advanced)
/// Override the OAuth issuer base URL (advanced)
#[arg(long = "experimental_issuer", value_name = "URL", hide = true)]
issuer_base_url: Option<String>,
/// EXPERIMENTAL: Use custom OAuth client ID (advanced)
#[arg(long = "experimental_client-id", value_name = "CLIENT_ID", hide = true)]
client_id: Option<String>,
#[command(subcommand)]
action: Option<LoginSubcommand>,
}
@@ -218,12 +194,25 @@ fn print_exit_messages(exit_info: AppExitInfo) {
}
}
/// As early as possible in the process lifecycle, apply hardening measures. We
/// skip this in debug builds to avoid interfering with debugging.
pub(crate) const CODEX_SECURE_MODE_ENV_VAR: &str = "CODEX_SECURE_MODE";
/// As early as possible in the process lifecycle, apply hardening measures
/// if the CODEX_SECURE_MODE environment variable is set to "1".
#[ctor::ctor]
#[cfg(not(debug_assertions))]
fn pre_main_hardening() {
codex_process_hardening::pre_main_hardening();
let secure_mode = match std::env::var(CODEX_SECURE_MODE_ENV_VAR) {
Ok(value) => value,
Err(_) => return,
};
if secure_mode == "1" {
codex_process_hardening::pre_main_hardening();
}
// Always clear this env var so child processes don't inherit it.
unsafe {
std::env::remove_var(CODEX_SECURE_MODE_ENV_VAR);
}
}
fn main() -> anyhow::Result<()> {
@@ -256,16 +245,10 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
);
codex_exec::run_main(exec_cli, codex_linux_sandbox_exe).await?;
}
Some(Subcommand::McpServer) => {
codex_mcp_server::run_main(codex_linux_sandbox_exe, root_config_overrides).await?;
}
Some(Subcommand::Mcp(mut mcp_cli)) => {
// Propagate any root-level config overrides (e.g. `-c key=value`).
prepend_config_flags(&mut mcp_cli.config_overrides, root_config_overrides.clone());
mcp_cli.run().await?;
}
Some(Subcommand::AppServer) => {
codex_app_server::run_main(codex_linux_sandbox_exe, root_config_overrides).await?;
mcp_cli.run(codex_linux_sandbox_exe).await?;
}
Some(Subcommand::Resume(ResumeCommand {
session_id,
@@ -291,14 +274,7 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
run_login_status(login_cli.config_overrides).await;
}
None => {
if login_cli.use_device_code {
run_login_with_device_code(
login_cli.config_overrides,
login_cli.issuer_base_url,
login_cli.client_id,
)
.await;
} else if let Some(api_key) = login_cli.api_key {
if let Some(api_key) = login_cli.api_key {
run_login_with_api_key(login_cli.config_overrides, api_key).await;
} else {
run_login_with_chatgpt(login_cli.config_overrides).await;
@@ -313,15 +289,15 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
);
run_logout(logout_cli.config_overrides).await;
}
Some(Subcommand::Completion(completion_cli)) => {
print_completion(completion_cli);
}
Some(Subcommand::Cloud(mut cloud_cli)) => {
Some(Subcommand::Proto(mut proto_cli)) => {
prepend_config_flags(
&mut cloud_cli.config_overrides,
&mut proto_cli.config_overrides,
root_config_overrides.clone(),
);
codex_cloud_tasks::run_main(cloud_cli, codex_linux_sandbox_exe).await?;
proto::run_main(proto_cli).await?;
}
Some(Subcommand::Completion(completion_cli)) => {
print_completion(completion_cli);
}
Some(Subcommand::Debug(debug_args)) => match debug_args.cmd {
DebugCommand::Seatbelt(mut seatbelt_cli) => {
@@ -354,10 +330,6 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
);
run_apply_command(apply_cli, None).await?;
}
Some(Subcommand::ResponsesApiProxy(args)) => {
tokio::task::spawn_blocking(move || codex_responses_api_proxy::run_main(args))
.await??;
}
Some(Subcommand::GenerateTs(gen_cli)) => {
codex_protocol_ts::generate_ts(&gen_cli.out_dir, gen_cli.prettier.as_deref())?;
}
@@ -455,7 +427,7 @@ fn print_completion(cmd: CompletionCommand) {
mod tests {
use super::*;
use codex_core::protocol::TokenUsage;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::ConversationId;
fn finalize_from_args(args: &[&str]) -> TuiCli {
let cli = MultitoolCli::try_parse_from(args).expect("parse");

View File

@@ -1,4 +1,5 @@
use std::collections::HashMap;
use std::path::PathBuf;
use anyhow::Context;
use anyhow::Result;
@@ -27,11 +28,14 @@ pub struct McpCli {
pub config_overrides: CliConfigOverrides,
#[command(subcommand)]
pub subcommand: McpSubcommand,
pub cmd: Option<McpSubcommand>,
}
#[derive(Debug, clap::Subcommand)]
pub enum McpSubcommand {
/// [experimental] Run the Codex MCP server (stdio transport).
Serve,
/// [experimental] List configured MCP servers.
List(ListArgs),
@@ -83,13 +87,17 @@ pub struct RemoveArgs {
}
impl McpCli {
pub async fn run(self) -> Result<()> {
pub async fn run(self, codex_linux_sandbox_exe: Option<PathBuf>) -> Result<()> {
let McpCli {
config_overrides,
subcommand,
cmd,
} = self;
let subcommand = cmd.unwrap_or(McpSubcommand::Serve);
match subcommand {
McpSubcommand::Serve => {
codex_mcp_server::run_main(codex_linux_sandbox_exe, config_overrides).await?;
}
McpSubcommand::List(args) => {
run_list(&config_overrides, args)?;
}

133
codex-rs/cli/src/proto.rs Normal file
View File

@@ -0,0 +1,133 @@
use std::io::IsTerminal;
use clap::Parser;
use codex_common::CliConfigOverrides;
use codex_core::AuthManager;
use codex_core::ConversationManager;
use codex_core::NewConversation;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
use codex_core::protocol::Event;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Submission;
use tokio::io::AsyncBufReadExt;
use tokio::io::BufReader;
use tracing::error;
use tracing::info;
#[derive(Debug, Parser)]
pub struct ProtoCli {
#[clap(skip)]
pub config_overrides: CliConfigOverrides,
}
pub async fn run_main(opts: ProtoCli) -> anyhow::Result<()> {
if std::io::stdin().is_terminal() {
anyhow::bail!("Protocol mode expects stdin to be a pipe, not a terminal");
}
tracing_subscriber::fmt()
.with_writer(std::io::stderr)
.init();
let ProtoCli { config_overrides } = opts;
let overrides_vec = config_overrides
.parse_overrides()
.map_err(anyhow::Error::msg)?;
let config = Config::load_with_cli_overrides(overrides_vec, ConfigOverrides::default())?;
// Use conversation_manager API to start a conversation
let conversation_manager =
ConversationManager::new(AuthManager::shared(config.codex_home.clone()));
let NewConversation {
conversation_id: _,
conversation,
session_configured,
} = conversation_manager.new_conversation(config).await?;
// Simulate streaming the session_configured event.
let synthetic_event = Event {
// Fake id value.
id: "".to_string(),
msg: EventMsg::SessionConfigured(session_configured),
};
let session_configured_event = match serde_json::to_string(&synthetic_event) {
Ok(s) => s,
Err(e) => {
error!("Failed to serialize session_configured: {e}");
return Err(anyhow::Error::from(e));
}
};
println!("{session_configured_event}");
// Task that reads JSON lines from stdin and forwards to Submission Queue
let sq_fut = {
let conversation = conversation.clone();
async move {
let stdin = BufReader::new(tokio::io::stdin());
let mut lines = stdin.lines();
loop {
let result = tokio::select! {
_ = tokio::signal::ctrl_c() => {
break
},
res = lines.next_line() => res,
};
match result {
Ok(Some(line)) => {
let line = line.trim();
if line.is_empty() {
continue;
}
match serde_json::from_str::<Submission>(line) {
Ok(sub) => {
if let Err(e) = conversation.submit_with_id(sub).await {
error!("{e:#}");
break;
}
}
Err(e) => {
error!("invalid submission: {e}");
}
}
}
_ => {
info!("Submission queue closed");
break;
}
}
}
}
};
// Task that reads events from the agent and prints them as JSON lines to stdout
let eq_fut = async move {
loop {
let event = tokio::select! {
_ = tokio::signal::ctrl_c() => break,
event = conversation.next_event() => event,
};
match event {
Ok(event) => {
let event_str = match serde_json::to_string(&event) {
Ok(s) => s,
Err(e) => {
error!("Failed to serialize event: {e}");
continue;
}
};
println!("{event_str}");
}
Err(e) => {
error!("{e:#}");
break;
}
}
}
info!("Event queue closed");
};
tokio::join!(sq_fut, eq_fut);
Ok(())
}

View File

@@ -1,27 +0,0 @@
[package]
name = "codex-cloud-tasks-client"
version = { workspace = true }
edition = "2024"
[lib]
name = "codex_cloud_tasks_client"
path = "src/lib.rs"
[lints]
workspace = true
[features]
default = ["online"]
online = ["dep:codex-backend-client"]
mock = []
[dependencies]
anyhow = "1"
async-trait = "0.1"
chrono = { version = "0.4", features = ["serde"] }
diffy = "0.4.2"
serde = { version = "1", features = ["derive"] }
serde_json = "1"
thiserror = "2.0.12"
codex-backend-client = { path = "../backend-client", optional = true }
codex-git-apply = { path = "../git-apply" }

View File

@@ -1,158 +0,0 @@
use chrono::DateTime;
use chrono::Utc;
use serde::Deserialize;
use serde::Serialize;
pub type Result<T> = std::result::Result<T, CloudTaskError>;
#[derive(Debug, thiserror::Error)]
pub enum CloudTaskError {
#[error("unimplemented: {0}")]
Unimplemented(&'static str),
#[error("http error: {0}")]
Http(String),
#[error("io error: {0}")]
Io(String),
#[error("{0}")]
Msg(String),
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
#[serde(transparent)]
pub struct TaskId(pub String);
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "kebab-case")]
pub enum TaskStatus {
Pending,
Ready,
Applied,
Error,
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
pub struct TaskSummary {
pub id: TaskId,
pub title: String,
pub status: TaskStatus,
pub updated_at: DateTime<Utc>,
/// Backend environment identifier (when available)
pub environment_id: Option<String>,
/// Human-friendly environment label (when available)
pub environment_label: Option<String>,
pub summary: DiffSummary,
/// True when the backend reports this task as a code review.
#[serde(default)]
pub is_review: bool,
/// Number of assistant attempts (best-of-N), when reported by the backend.
#[serde(default)]
pub attempt_total: Option<usize>,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq, Default)]
pub enum AttemptStatus {
Pending,
InProgress,
Completed,
Failed,
Cancelled,
#[default]
Unknown,
}
#[derive(Clone, Debug, PartialEq, Eq)]
pub struct TurnAttempt {
pub turn_id: String,
pub attempt_placement: Option<i64>,
pub created_at: Option<DateTime<Utc>>,
pub status: AttemptStatus,
pub diff: Option<String>,
pub messages: Vec<String>,
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "lowercase")]
pub enum ApplyStatus {
Success,
Partial,
Error,
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
pub struct ApplyOutcome {
pub applied: bool,
pub status: ApplyStatus,
pub message: String,
#[serde(default)]
pub skipped_paths: Vec<String>,
#[serde(default)]
pub conflict_paths: Vec<String>,
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
pub struct CreatedTask {
pub id: TaskId,
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize, Default)]
pub struct DiffSummary {
pub files_changed: usize,
pub lines_added: usize,
pub lines_removed: usize,
}
#[derive(Clone, Debug, PartialEq, Eq)]
pub struct TaskText {
pub prompt: Option<String>,
pub messages: Vec<String>,
pub turn_id: Option<String>,
pub sibling_turn_ids: Vec<String>,
pub attempt_placement: Option<i64>,
pub attempt_status: AttemptStatus,
}
impl Default for TaskText {
fn default() -> Self {
Self {
prompt: None,
messages: Vec::new(),
turn_id: None,
sibling_turn_ids: Vec::new(),
attempt_placement: None,
attempt_status: AttemptStatus::Unknown,
}
}
}
#[async_trait::async_trait]
pub trait CloudBackend: Send + Sync {
async fn list_tasks(&self, env: Option<&str>) -> Result<Vec<TaskSummary>>;
async fn get_task_diff(&self, id: TaskId) -> Result<Option<String>>;
/// Return assistant output messages (no diff) when available.
async fn get_task_messages(&self, id: TaskId) -> Result<Vec<String>>;
/// Return the creating prompt and assistant messages (when available).
async fn get_task_text(&self, id: TaskId) -> Result<TaskText>;
/// Return any sibling attempts (best-of-N) for the given assistant turn.
async fn list_sibling_attempts(
&self,
task: TaskId,
turn_id: String,
) -> Result<Vec<TurnAttempt>>;
/// Dry-run apply (preflight) that validates whether the patch would apply cleanly.
/// Never modifies the working tree. When `diff_override` is supplied, the provided diff is
/// used instead of re-fetching the task details so callers can apply alternate attempts.
async fn apply_task_preflight(
&self,
id: TaskId,
diff_override: Option<String>,
) -> Result<ApplyOutcome>;
async fn apply_task(&self, id: TaskId, diff_override: Option<String>) -> Result<ApplyOutcome>;
async fn create_task(
&self,
env_id: &str,
prompt: &str,
git_ref: &str,
qa_mode: bool,
best_of_n: usize,
) -> Result<CreatedTask>;
}

View File

@@ -1,769 +0,0 @@
use crate::ApplyOutcome;
use crate::ApplyStatus;
use crate::AttemptStatus;
use crate::CloudBackend;
use crate::CloudTaskError;
use crate::DiffSummary;
use crate::Result;
use crate::TaskId;
use crate::TaskStatus;
use crate::TaskSummary;
use crate::TurnAttempt;
use crate::api::TaskText;
use chrono::DateTime;
use chrono::Utc;
use codex_backend_client as backend;
use codex_backend_client::CodeTaskDetailsResponseExt;
#[derive(Clone)]
pub struct HttpClient {
pub base_url: String,
backend: backend::Client,
}
impl HttpClient {
pub fn new(base_url: impl Into<String>) -> anyhow::Result<Self> {
let base_url = base_url.into();
let backend = backend::Client::new(base_url.clone())?;
Ok(Self { base_url, backend })
}
pub fn with_bearer_token(mut self, token: impl Into<String>) -> Self {
self.backend = self.backend.clone().with_bearer_token(token);
self
}
pub fn with_user_agent(mut self, ua: impl Into<String>) -> Self {
self.backend = self.backend.clone().with_user_agent(ua);
self
}
pub fn with_chatgpt_account_id(mut self, account_id: impl Into<String>) -> Self {
self.backend = self.backend.clone().with_chatgpt_account_id(account_id);
self
}
fn tasks_api(&self) -> api::Tasks<'_> {
api::Tasks::new(self)
}
fn attempts_api(&self) -> api::Attempts<'_> {
api::Attempts::new(self)
}
fn apply_api(&self) -> api::Apply<'_> {
api::Apply::new(self)
}
}
#[async_trait::async_trait]
impl CloudBackend for HttpClient {
async fn list_tasks(&self, env: Option<&str>) -> Result<Vec<TaskSummary>> {
self.tasks_api().list(env).await
}
async fn get_task_diff(&self, id: TaskId) -> Result<Option<String>> {
self.tasks_api().diff(id).await
}
async fn get_task_messages(&self, id: TaskId) -> Result<Vec<String>> {
self.tasks_api().messages(id).await
}
async fn get_task_text(&self, id: TaskId) -> Result<TaskText> {
self.tasks_api().task_text(id).await
}
async fn list_sibling_attempts(
&self,
task: TaskId,
turn_id: String,
) -> Result<Vec<TurnAttempt>> {
self.attempts_api().list(task, turn_id).await
}
async fn apply_task(&self, id: TaskId, diff_override: Option<String>) -> Result<ApplyOutcome> {
self.apply_api().run(id, diff_override, false).await
}
async fn apply_task_preflight(
&self,
id: TaskId,
diff_override: Option<String>,
) -> Result<ApplyOutcome> {
self.apply_api().run(id, diff_override, true).await
}
async fn create_task(
&self,
env_id: &str,
prompt: &str,
git_ref: &str,
qa_mode: bool,
best_of_n: usize,
) -> Result<crate::CreatedTask> {
self.tasks_api()
.create(env_id, prompt, git_ref, qa_mode, best_of_n)
.await
}
}
mod api {
use super::*;
use serde_json::Value;
use std::cmp::Ordering;
use std::collections::HashMap;
pub(crate) struct Tasks<'a> {
base_url: &'a str,
backend: &'a backend::Client,
}
impl<'a> Tasks<'a> {
pub(crate) fn new(client: &'a HttpClient) -> Self {
Self {
base_url: &client.base_url,
backend: &client.backend,
}
}
pub(crate) async fn list(&self, env: Option<&str>) -> Result<Vec<TaskSummary>> {
let resp = self
.backend
.list_tasks(Some(20), Some("current"), env)
.await
.map_err(|e| CloudTaskError::Http(format!("list_tasks failed: {e}")))?;
let tasks: Vec<TaskSummary> = resp
.items
.into_iter()
.map(map_task_list_item_to_summary)
.collect();
append_error_log(&format!(
"http.list_tasks: env={} items={}",
env.unwrap_or("<all>"),
tasks.len()
));
Ok(tasks)
}
pub(crate) async fn diff(&self, id: TaskId) -> Result<Option<String>> {
let (details, body, ct) = self
.details_with_body(&id.0)
.await
.map_err(|e| CloudTaskError::Http(format!("get_task_details failed: {e}")))?;
if let Some(diff) = details.unified_diff() {
return Ok(Some(diff));
}
let _ = (body, ct);
Ok(None)
}
pub(crate) async fn messages(&self, id: TaskId) -> Result<Vec<String>> {
let (details, body, ct) = self
.details_with_body(&id.0)
.await
.map_err(|e| CloudTaskError::Http(format!("get_task_details failed: {e}")))?;
let mut msgs = details.assistant_text_messages();
if msgs.is_empty() {
msgs.extend(extract_assistant_messages_from_body(&body));
}
if !msgs.is_empty() {
return Ok(msgs);
}
if let Some(err) = details.assistant_error_message() {
return Ok(vec![format!("Task failed: {err}")]);
}
let url = match details_path(self.base_url, &id.0) {
Some(url) => url,
None => format!("{}/api/codex/tasks/{}", self.base_url, id.0),
};
Err(CloudTaskError::Http(format!(
"No assistant text messages in response. GET {url}; content-type={ct}; body={body}"
)))
}
pub(crate) async fn task_text(&self, id: TaskId) -> Result<TaskText> {
let (details, body, _ct) = self
.details_with_body(&id.0)
.await
.map_err(|e| CloudTaskError::Http(format!("get_task_details failed: {e}")))?;
let prompt = details.user_text_prompt();
let mut messages = details.assistant_text_messages();
if messages.is_empty() {
messages.extend(extract_assistant_messages_from_body(&body));
}
let assistant_turn = details.current_assistant_turn.as_ref();
let turn_id = assistant_turn.and_then(|turn| turn.id.clone());
let sibling_turn_ids = assistant_turn
.map(|turn| turn.sibling_turn_ids.clone())
.unwrap_or_default();
let attempt_placement = assistant_turn.and_then(|turn| turn.attempt_placement);
let attempt_status = attempt_status_from_str(
assistant_turn.and_then(|turn| turn.turn_status.as_deref()),
);
Ok(TaskText {
prompt,
messages,
turn_id,
sibling_turn_ids,
attempt_placement,
attempt_status,
})
}
pub(crate) async fn create(
&self,
env_id: &str,
prompt: &str,
git_ref: &str,
qa_mode: bool,
best_of_n: usize,
) -> Result<crate::CreatedTask> {
let mut input_items: Vec<serde_json::Value> = Vec::new();
input_items.push(serde_json::json!({
"type": "message",
"role": "user",
"content": [{ "content_type": "text", "text": prompt }]
}));
if let Ok(diff) = std::env::var("CODEX_STARTING_DIFF")
&& !diff.is_empty()
{
input_items.push(serde_json::json!({
"type": "pre_apply_patch",
"output_diff": { "diff": diff }
}));
}
let mut request_body = serde_json::json!({
"new_task": {
"environment_id": env_id,
"branch": git_ref,
"run_environment_in_qa_mode": qa_mode,
},
"input_items": input_items,
});
if best_of_n > 1
&& let Some(obj) = request_body.as_object_mut()
{
obj.insert(
"metadata".to_string(),
serde_json::json!({ "best_of_n": best_of_n }),
);
}
match self.backend.create_task(request_body).await {
Ok(id) => {
append_error_log(&format!(
"new_task: created id={id} env={} prompt_chars={}",
env_id,
prompt.chars().count()
));
Ok(crate::CreatedTask { id: TaskId(id) })
}
Err(e) => {
append_error_log(&format!(
"new_task: create failed env={} prompt_chars={}: {}",
env_id,
prompt.chars().count(),
e
));
Err(CloudTaskError::Http(format!("create_task failed: {e}")))
}
}
}
async fn details_with_body(
&self,
id: &str,
) -> anyhow::Result<(backend::CodeTaskDetailsResponse, String, String)> {
let (parsed, body, ct) = self.backend.get_task_details_with_body(id).await?;
Ok((parsed, body, ct))
}
}
pub(crate) struct Attempts<'a> {
backend: &'a backend::Client,
}
impl<'a> Attempts<'a> {
pub(crate) fn new(client: &'a HttpClient) -> Self {
Self {
backend: &client.backend,
}
}
pub(crate) async fn list(&self, task: TaskId, turn_id: String) -> Result<Vec<TurnAttempt>> {
let resp = self
.backend
.list_sibling_turns(&task.0, &turn_id)
.await
.map_err(|e| CloudTaskError::Http(format!("list_sibling_turns failed: {e}")))?;
let mut attempts: Vec<TurnAttempt> = resp
.sibling_turns
.iter()
.filter_map(turn_attempt_from_map)
.collect();
attempts.sort_by(compare_attempts);
Ok(attempts)
}
}
pub(crate) struct Apply<'a> {
backend: &'a backend::Client,
}
impl<'a> Apply<'a> {
pub(crate) fn new(client: &'a HttpClient) -> Self {
Self {
backend: &client.backend,
}
}
pub(crate) async fn run(
&self,
task_id: TaskId,
diff_override: Option<String>,
preflight: bool,
) -> Result<ApplyOutcome> {
let id = task_id.0.clone();
let diff = match diff_override {
Some(diff) => diff,
None => {
let details = self.backend.get_task_details(&id).await.map_err(|e| {
CloudTaskError::Http(format!("get_task_details failed: {e}"))
})?;
details.unified_diff().ok_or_else(|| {
CloudTaskError::Msg(format!("No diff available for task {id}"))
})?
}
};
if !is_unified_diff(&diff) {
let summary = summarize_patch_for_logging(&diff);
let mode = if preflight { "preflight" } else { "apply" };
append_error_log(&format!(
"apply_error: id={id} mode={mode} format=non-unified; {summary}"
));
return Ok(ApplyOutcome {
applied: false,
status: ApplyStatus::Error,
message: "Expected unified git diff; backend returned an incompatible format."
.to_string(),
skipped_paths: Vec::new(),
conflict_paths: Vec::new(),
});
}
let req = codex_git_apply::ApplyGitRequest {
cwd: std::env::current_dir().unwrap_or_else(|_| std::env::temp_dir()),
diff: diff.clone(),
revert: false,
preflight,
};
let r = codex_git_apply::apply_git_patch(&req)
.map_err(|e| CloudTaskError::Io(format!("git apply failed to run: {e}")))?;
let status = if r.exit_code == 0 {
ApplyStatus::Success
} else if !r.applied_paths.is_empty() || !r.conflicted_paths.is_empty() {
ApplyStatus::Partial
} else {
ApplyStatus::Error
};
let applied = matches!(status, ApplyStatus::Success) && !preflight;
let message = if preflight {
match status {
ApplyStatus::Success => {
format!("Preflight passed for task {id} (applies cleanly)")
}
ApplyStatus::Partial => format!(
"Preflight: patch does not fully apply for task {id} (applied={}, skipped={}, conflicts={})",
r.applied_paths.len(),
r.skipped_paths.len(),
r.conflicted_paths.len()
),
ApplyStatus::Error => format!(
"Preflight failed for task {id} (applied={}, skipped={}, conflicts={})",
r.applied_paths.len(),
r.skipped_paths.len(),
r.conflicted_paths.len()
),
}
} else {
match status {
ApplyStatus::Success => format!(
"Applied task {id} locally ({} files)",
r.applied_paths.len()
),
ApplyStatus::Partial => format!(
"Apply partially succeeded for task {id} (applied={}, skipped={}, conflicts={})",
r.applied_paths.len(),
r.skipped_paths.len(),
r.conflicted_paths.len()
),
ApplyStatus::Error => format!(
"Apply failed for task {id} (applied={}, skipped={}, conflicts={})",
r.applied_paths.len(),
r.skipped_paths.len(),
r.conflicted_paths.len()
),
}
};
if matches!(status, ApplyStatus::Partial | ApplyStatus::Error)
|| (preflight && !matches!(status, ApplyStatus::Success))
{
let mut log = String::new();
let summary = summarize_patch_for_logging(&diff);
let mode = if preflight { "preflight" } else { "apply" };
use std::fmt::Write as _;
let _ = writeln!(
&mut log,
"apply_result: mode={} id={} status={:?} applied={} skipped={} conflicts={} cmd={}",
mode,
id,
status,
r.applied_paths.len(),
r.skipped_paths.len(),
r.conflicted_paths.len(),
r.cmd_for_log
);
let _ = writeln!(
&mut log,
"stdout_tail=\n{}\nstderr_tail=\n{}",
tail(&r.stdout, 2000),
tail(&r.stderr, 2000)
);
let _ = writeln!(&mut log, "{summary}");
let _ = writeln!(
&mut log,
"----- PATCH BEGIN -----\n{diff}\n----- PATCH END -----"
);
append_error_log(&log);
}
Ok(ApplyOutcome {
applied,
status,
message,
skipped_paths: r.skipped_paths,
conflict_paths: r.conflicted_paths,
})
}
}
fn details_path(base_url: &str, id: &str) -> Option<String> {
if base_url.contains("/backend-api") {
Some(format!("{base_url}/wham/tasks/{id}"))
} else if base_url.contains("/api/codex") {
Some(format!("{base_url}/tasks/{id}"))
} else {
None
}
}
fn extract_assistant_messages_from_body(body: &str) -> Vec<String> {
let mut msgs = Vec::new();
if let Ok(full) = serde_json::from_str::<serde_json::Value>(body)
&& let Some(arr) = full
.get("current_assistant_turn")
.and_then(|v| v.get("worklog"))
.and_then(|v| v.get("messages"))
.and_then(|v| v.as_array())
{
for m in arr {
let is_assistant = m
.get("author")
.and_then(|a| a.get("role"))
.and_then(|r| r.as_str())
== Some("assistant");
if !is_assistant {
continue;
}
if let Some(parts) = m
.get("content")
.and_then(|c| c.get("parts"))
.and_then(|p| p.as_array())
{
for p in parts {
if let Some(s) = p.as_str() {
if !s.is_empty() {
msgs.push(s.to_string());
}
continue;
}
if let Some(obj) = p.as_object()
&& obj.get("content_type").and_then(|t| t.as_str()) == Some("text")
&& let Some(txt) = obj.get("text").and_then(|t| t.as_str())
{
msgs.push(txt.to_string());
}
}
}
}
}
msgs
}
fn turn_attempt_from_map(turn: &HashMap<String, Value>) -> Option<TurnAttempt> {
let turn_id = turn.get("id").and_then(Value::as_str)?.to_string();
let attempt_placement = turn.get("attempt_placement").and_then(Value::as_i64);
let created_at = parse_timestamp_value(turn.get("created_at"));
let status = attempt_status_from_str(turn.get("turn_status").and_then(Value::as_str));
let diff = extract_diff_from_turn(turn);
let messages = extract_assistant_messages_from_turn(turn);
Some(TurnAttempt {
turn_id,
attempt_placement,
created_at,
status,
diff,
messages,
})
}
fn compare_attempts(a: &TurnAttempt, b: &TurnAttempt) -> Ordering {
match (a.attempt_placement, b.attempt_placement) {
(Some(lhs), Some(rhs)) => lhs.cmp(&rhs),
(Some(_), None) => Ordering::Less,
(None, Some(_)) => Ordering::Greater,
(None, None) => match (a.created_at, b.created_at) {
(Some(lhs), Some(rhs)) => lhs.cmp(&rhs),
(Some(_), None) => Ordering::Less,
(None, Some(_)) => Ordering::Greater,
(None, None) => a.turn_id.cmp(&b.turn_id),
},
}
}
fn extract_diff_from_turn(turn: &HashMap<String, Value>) -> Option<String> {
let items = turn.get("output_items").and_then(Value::as_array)?;
for item in items {
match item.get("type").and_then(Value::as_str) {
Some("output_diff") => {
if let Some(diff) = item.get("diff").and_then(Value::as_str)
&& !diff.is_empty()
{
return Some(diff.to_string());
}
}
Some("pr") => {
if let Some(diff) = item
.get("output_diff")
.and_then(Value::as_object)
.and_then(|od| od.get("diff"))
.and_then(Value::as_str)
&& !diff.is_empty()
{
return Some(diff.to_string());
}
}
_ => {}
}
}
None
}
fn extract_assistant_messages_from_turn(turn: &HashMap<String, Value>) -> Vec<String> {
let mut msgs = Vec::new();
if let Some(items) = turn.get("output_items").and_then(Value::as_array) {
for item in items {
if item.get("type").and_then(Value::as_str) != Some("message") {
continue;
}
if let Some(content) = item.get("content").and_then(Value::as_array) {
for part in content {
if part.get("content_type").and_then(Value::as_str) == Some("text")
&& let Some(txt) = part.get("text").and_then(Value::as_str)
&& !txt.is_empty()
{
msgs.push(txt.to_string());
}
}
}
}
}
msgs
}
fn attempt_status_from_str(raw: Option<&str>) -> AttemptStatus {
match raw.unwrap_or_default() {
"failed" => AttemptStatus::Failed,
"completed" => AttemptStatus::Completed,
"in_progress" => AttemptStatus::InProgress,
"pending" => AttemptStatus::Pending,
_ => AttemptStatus::Pending,
}
}
fn parse_timestamp_value(v: Option<&Value>) -> Option<DateTime<Utc>> {
let ts = v?.as_f64()?;
let secs = ts as i64;
let nanos = ((ts - secs as f64) * 1_000_000_000.0) as u32;
Some(DateTime::<Utc>::from(
std::time::UNIX_EPOCH + std::time::Duration::new(secs.max(0) as u64, nanos),
))
}
fn map_task_list_item_to_summary(src: backend::TaskListItem) -> TaskSummary {
let status_display = src.task_status_display.as_ref();
TaskSummary {
id: TaskId(src.id),
title: src.title,
status: map_status(status_display),
updated_at: parse_updated_at(src.updated_at.as_ref()),
environment_id: None,
environment_label: env_label_from_status_display(status_display),
summary: diff_summary_from_status_display(status_display),
is_review: src
.pull_requests
.as_ref()
.is_some_and(|prs| !prs.is_empty()),
attempt_total: attempt_total_from_status_display(status_display),
}
}
fn map_status(v: Option<&HashMap<String, Value>>) -> TaskStatus {
if let Some(val) = v {
if let Some(turn) = val
.get("latest_turn_status_display")
.and_then(Value::as_object)
&& let Some(s) = turn.get("turn_status").and_then(Value::as_str)
{
return match s {
"failed" => TaskStatus::Error,
"completed" => TaskStatus::Ready,
"in_progress" => TaskStatus::Pending,
"pending" => TaskStatus::Pending,
"cancelled" => TaskStatus::Error,
_ => TaskStatus::Pending,
};
}
if let Some(state) = val.get("state").and_then(Value::as_str) {
return match state {
"pending" => TaskStatus::Pending,
"ready" => TaskStatus::Ready,
"applied" => TaskStatus::Applied,
"error" => TaskStatus::Error,
_ => TaskStatus::Pending,
};
}
}
TaskStatus::Pending
}
fn parse_updated_at(ts: Option<&f64>) -> DateTime<Utc> {
if let Some(v) = ts {
let secs = *v as i64;
let nanos = ((*v - secs as f64) * 1_000_000_000.0) as u32;
return DateTime::<Utc>::from(
std::time::UNIX_EPOCH + std::time::Duration::new(secs.max(0) as u64, nanos),
);
}
Utc::now()
}
fn env_label_from_status_display(v: Option<&HashMap<String, Value>>) -> Option<String> {
let map = v?;
map.get("environment_label")
.and_then(Value::as_str)
.map(str::to_string)
}
fn diff_summary_from_status_display(v: Option<&HashMap<String, Value>>) -> DiffSummary {
let mut out = DiffSummary::default();
let Some(map) = v else { return out };
let latest = map
.get("latest_turn_status_display")
.and_then(Value::as_object);
let Some(latest) = latest else { return out };
if let Some(ds) = latest.get("diff_stats").and_then(Value::as_object) {
if let Some(n) = ds.get("files_modified").and_then(Value::as_i64) {
out.files_changed = n.max(0) as usize;
}
if let Some(n) = ds.get("lines_added").and_then(Value::as_i64) {
out.lines_added = n.max(0) as usize;
}
if let Some(n) = ds.get("lines_removed").and_then(Value::as_i64) {
out.lines_removed = n.max(0) as usize;
}
}
out
}
fn attempt_total_from_status_display(v: Option<&HashMap<String, Value>>) -> Option<usize> {
let map = v?;
let latest = map
.get("latest_turn_status_display")
.and_then(Value::as_object)?;
let siblings = latest.get("sibling_turn_ids").and_then(Value::as_array)?;
Some(siblings.len().saturating_add(1))
}
fn is_unified_diff(diff: &str) -> bool {
let t = diff.trim_start();
if t.starts_with("diff --git ") {
return true;
}
let has_dash_headers = diff.contains("\n--- ") && diff.contains("\n+++ ");
let has_hunk = diff.contains("\n@@ ") || diff.starts_with("@@ ");
has_dash_headers && has_hunk
}
fn tail(s: &str, max: usize) -> String {
if s.len() <= max {
s.to_string()
} else {
s[s.len() - max..].to_string()
}
}
fn summarize_patch_for_logging(patch: &str) -> String {
let trimmed = patch.trim_start();
let kind = if trimmed.starts_with("*** Begin Patch") {
"codex-patch"
} else if trimmed.starts_with("diff --git ") || trimmed.contains("\n*** End Patch\n") {
"git-diff"
} else if trimmed.starts_with("@@ ") || trimmed.contains("\n@@ ") {
"unified-diff"
} else {
"unknown"
};
let lines = patch.lines().count();
let chars = patch.len();
let cwd = std::env::current_dir()
.ok()
.map(|p| p.display().to_string())
.unwrap_or_else(|| "<unknown>".to_string());
let head: String = patch.lines().take(20).collect::<Vec<&str>>().join("\n");
let head_trunc = if head.len() > 800 {
format!("{}", &head[..800])
} else {
head
};
format!(
"patch_summary: kind={kind} lines={lines} chars={chars} cwd={cwd} ; head=\n{head_trunc}"
)
}
}
fn append_error_log(message: &str) {
let ts = Utc::now().to_rfc3339();
if let Ok(mut f) = std::fs::OpenOptions::new()
.create(true)
.append(true)
.open("error.log")
{
use std::io::Write as _;
let _ = writeln!(f, "[{ts}] {message}");
}
}

View File

@@ -1,29 +0,0 @@
mod api;
pub use api::ApplyOutcome;
pub use api::ApplyStatus;
pub use api::AttemptStatus;
pub use api::CloudBackend;
pub use api::CloudTaskError;
pub use api::CreatedTask;
pub use api::DiffSummary;
pub use api::Result;
pub use api::TaskId;
pub use api::TaskStatus;
pub use api::TaskSummary;
pub use api::TaskText;
pub use api::TurnAttempt;
#[cfg(feature = "mock")]
mod mock;
#[cfg(feature = "online")]
mod http;
#[cfg(feature = "mock")]
pub use mock::MockClient;
#[cfg(feature = "online")]
pub use http::HttpClient;
// Reusable apply engine now lives in the shared crate `codex-git-apply`.

View File

@@ -1,180 +0,0 @@
use crate::ApplyOutcome;
use crate::AttemptStatus;
use crate::CloudBackend;
use crate::DiffSummary;
use crate::Result;
use crate::TaskId;
use crate::TaskStatus;
use crate::TaskSummary;
use crate::TurnAttempt;
use crate::api::TaskText;
use chrono::Utc;
#[derive(Clone, Default)]
pub struct MockClient;
#[async_trait::async_trait]
impl CloudBackend for MockClient {
async fn list_tasks(&self, _env: Option<&str>) -> Result<Vec<TaskSummary>> {
// Slightly vary content by env to aid tests that rely on the mock
let rows = match _env {
Some("env-A") => vec![("T-2000", "A: First", TaskStatus::Ready)],
Some("env-B") => vec![
("T-3000", "B: One", TaskStatus::Ready),
("T-3001", "B: Two", TaskStatus::Pending),
],
_ => vec![
("T-1000", "Update README formatting", TaskStatus::Ready),
("T-1001", "Fix clippy warnings in core", TaskStatus::Pending),
("T-1002", "Add contributing guide", TaskStatus::Ready),
],
};
let environment_id = _env.map(str::to_string);
let environment_label = match _env {
Some("env-A") => Some("Env A".to_string()),
Some("env-B") => Some("Env B".to_string()),
Some(other) => Some(other.to_string()),
None => Some("Global".to_string()),
};
let mut out = Vec::new();
for (id_str, title, status) in rows {
let id = TaskId(id_str.to_string());
let diff = mock_diff_for(&id);
let (a, d) = count_from_unified(&diff);
out.push(TaskSummary {
id,
title: title.to_string(),
status,
updated_at: Utc::now(),
environment_id: environment_id.clone(),
environment_label: environment_label.clone(),
summary: DiffSummary {
files_changed: 1,
lines_added: a,
lines_removed: d,
},
is_review: false,
attempt_total: Some(if id_str == "T-1000" { 2 } else { 1 }),
});
}
Ok(out)
}
async fn get_task_diff(&self, id: TaskId) -> Result<Option<String>> {
Ok(Some(mock_diff_for(&id)))
}
async fn get_task_messages(&self, _id: TaskId) -> Result<Vec<String>> {
Ok(vec![
"Mock assistant output: this task contains no diff.".to_string(),
])
}
async fn get_task_text(&self, _id: TaskId) -> Result<TaskText> {
Ok(TaskText {
prompt: Some("Why is there no diff?".to_string()),
messages: vec!["Mock assistant output: this task contains no diff.".to_string()],
turn_id: Some("mock-turn".to_string()),
sibling_turn_ids: Vec::new(),
attempt_placement: Some(0),
attempt_status: AttemptStatus::Completed,
})
}
async fn apply_task(&self, id: TaskId, _diff_override: Option<String>) -> Result<ApplyOutcome> {
Ok(ApplyOutcome {
applied: true,
status: crate::ApplyStatus::Success,
message: format!("Applied task {} locally (mock)", id.0),
skipped_paths: Vec::new(),
conflict_paths: Vec::new(),
})
}
async fn apply_task_preflight(
&self,
id: TaskId,
_diff_override: Option<String>,
) -> Result<ApplyOutcome> {
Ok(ApplyOutcome {
applied: false,
status: crate::ApplyStatus::Success,
message: format!("Preflight passed for task {} (mock)", id.0),
skipped_paths: Vec::new(),
conflict_paths: Vec::new(),
})
}
async fn list_sibling_attempts(
&self,
task: TaskId,
_turn_id: String,
) -> Result<Vec<TurnAttempt>> {
if task.0 == "T-1000" {
return Ok(vec![TurnAttempt {
turn_id: "T-1000-attempt-2".to_string(),
attempt_placement: Some(1),
created_at: Some(Utc::now()),
status: AttemptStatus::Completed,
diff: Some(mock_diff_for(&task)),
messages: vec!["Mock alternate attempt".to_string()],
}]);
}
Ok(Vec::new())
}
async fn create_task(
&self,
env_id: &str,
prompt: &str,
git_ref: &str,
qa_mode: bool,
best_of_n: usize,
) -> Result<crate::CreatedTask> {
let _ = (env_id, prompt, git_ref, qa_mode, best_of_n);
let id = format!("task_local_{}", chrono::Utc::now().timestamp_millis());
Ok(crate::CreatedTask { id: TaskId(id) })
}
}
fn mock_diff_for(id: &TaskId) -> String {
match id.0.as_str() {
"T-1000" => {
"diff --git a/README.md b/README.md\nindex 000000..111111 100644\n--- a/README.md\n+++ b/README.md\n@@ -1,2 +1,3 @@\n Intro\n-Hello\n+Hello, world!\n+Task: T-1000\n".to_string()
}
"T-1001" => {
"diff --git a/core/src/lib.rs b/core/src/lib.rs\nindex 000000..111111 100644\n--- a/core/src/lib.rs\n+++ b/core/src/lib.rs\n@@ -1,2 +1,1 @@\n-use foo;\n use bar;\n".to_string()
}
_ => {
"diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md\nindex 000000..111111 100644\n--- /dev/null\n+++ b/CONTRIBUTING.md\n@@ -0,0 +1,3 @@\n+## Contributing\n+Please open PRs.\n+Thanks!\n".to_string()
}
}
}
fn count_from_unified(diff: &str) -> (usize, usize) {
if let Ok(patch) = diffy::Patch::from_str(diff) {
patch
.hunks()
.iter()
.flat_map(diffy::Hunk::lines)
.fold((0, 0), |(a, d), l| match l {
diffy::Line::Insert(_) => (a + 1, d),
diffy::Line::Delete(_) => (a, d + 1),
_ => (a, d),
})
} else {
let mut a = 0;
let mut d = 0;
for l in diff.lines() {
if l.starts_with("+++") || l.starts_with("---") || l.starts_with("@@") {
continue;
}
match l.as_bytes().first() {
Some(b'+') => a += 1,
Some(b'-') => d += 1,
_ => {}
}
}
(a, d)
}
}

View File

@@ -1,36 +0,0 @@
[package]
name = "codex-cloud-tasks"
version = { workspace = true }
edition = "2024"
[lib]
name = "codex_cloud_tasks"
path = "src/lib.rs"
[lints]
workspace = true
[dependencies]
anyhow = "1"
clap = { version = "4", features = ["derive"] }
codex-common = { path = "../common", features = ["cli"] }
tokio = { version = "1", features = ["macros", "rt-multi-thread"] }
tracing = { version = "0.1.41", features = ["log"] }
tracing-subscriber = { version = "0.3.19", features = ["env-filter"] }
codex-cloud-tasks-client = { path = "../cloud-tasks-client", features = ["mock", "online"] }
ratatui = { version = "0.29.0" }
crossterm = { version = "0.28.1", features = ["event-stream"] }
tokio-stream = "0.1.17"
chrono = { version = "0.4", features = ["serde"] }
codex-login = { path = "../login" }
codex-core = { path = "../core" }
throbber-widgets-tui = "0.8.0"
base64 = "0.22"
serde_json = "1"
reqwest = { version = "0.12", features = ["json"] }
serde = { version = "1", features = ["derive"] }
unicode-width = "0.1"
codex-tui = { path = "../tui" }
[dev-dependencies]
async-trait = "0.1"

View File

@@ -1,482 +0,0 @@
use std::time::Duration;
// Environment filter data models for the TUI
#[derive(Clone, Debug, Default)]
pub struct EnvironmentRow {
pub id: String,
pub label: Option<String>,
pub is_pinned: bool,
pub repo_hints: Option<String>, // e.g., "openai/codex"
}
#[derive(Clone, Debug, Default)]
pub struct EnvModalState {
pub query: String,
pub selected: usize,
}
#[derive(Clone, Debug, Default)]
pub struct BestOfModalState {
pub selected: usize,
}
#[derive(Clone, Debug, Copy, PartialEq, Eq)]
pub enum ApplyResultLevel {
Success,
Partial,
Error,
}
#[derive(Clone, Debug)]
pub struct ApplyModalState {
pub task_id: TaskId,
pub title: String,
pub result_message: Option<String>,
pub result_level: Option<ApplyResultLevel>,
pub skipped_paths: Vec<String>,
pub conflict_paths: Vec<String>,
pub diff_override: Option<String>,
}
use crate::scrollable_diff::ScrollableDiff;
use codex_cloud_tasks_client::CloudBackend;
use codex_cloud_tasks_client::TaskId;
use codex_cloud_tasks_client::TaskSummary;
use throbber_widgets_tui::ThrobberState;
#[derive(Default)]
pub struct App {
pub tasks: Vec<TaskSummary>,
pub selected: usize,
pub status: String,
pub diff_overlay: Option<DiffOverlay>,
pub throbber: ThrobberState,
pub refresh_inflight: bool,
pub details_inflight: bool,
// Environment filter state
pub env_filter: Option<String>,
pub env_modal: Option<EnvModalState>,
pub apply_modal: Option<ApplyModalState>,
pub best_of_modal: Option<BestOfModalState>,
pub environments: Vec<EnvironmentRow>,
pub env_last_loaded: Option<std::time::Instant>,
pub env_loading: bool,
pub env_error: Option<String>,
// New Task page
pub new_task: Option<crate::new_task::NewTaskPage>,
pub best_of_n: usize,
// Apply preflight spinner state
pub apply_preflight_inflight: bool,
// Apply action spinner state
pub apply_inflight: bool,
// Background enrichment coordination
pub list_generation: u64,
pub in_flight: std::collections::HashSet<String>,
// Background enrichment caches were planned; currently unused.
}
impl App {
pub fn new() -> Self {
Self {
tasks: Vec::new(),
selected: 0,
status: "Press r to refresh".to_string(),
diff_overlay: None,
throbber: ThrobberState::default(),
refresh_inflight: false,
details_inflight: false,
env_filter: None,
env_modal: None,
apply_modal: None,
best_of_modal: None,
environments: Vec::new(),
env_last_loaded: None,
env_loading: false,
env_error: None,
new_task: None,
best_of_n: 1,
apply_preflight_inflight: false,
apply_inflight: false,
list_generation: 0,
in_flight: std::collections::HashSet::new(),
}
}
pub fn next(&mut self) {
if self.tasks.is_empty() {
return;
}
self.selected = (self.selected + 1).min(self.tasks.len().saturating_sub(1));
}
pub fn prev(&mut self) {
if self.tasks.is_empty() {
return;
}
if self.selected > 0 {
self.selected -= 1;
}
}
}
pub async fn load_tasks(
backend: &dyn CloudBackend,
env: Option<&str>,
) -> anyhow::Result<Vec<TaskSummary>> {
// In later milestones, add a small debounce, spinner, and error display.
let tasks = tokio::time::timeout(Duration::from_secs(5), backend.list_tasks(env)).await??;
// Hide review-only tasks from the main list.
let filtered: Vec<TaskSummary> = tasks.into_iter().filter(|t| !t.is_review).collect();
Ok(filtered)
}
pub struct DiffOverlay {
pub title: String,
pub task_id: TaskId,
pub sd: ScrollableDiff,
pub base_can_apply: bool,
pub diff_lines: Vec<String>,
pub text_lines: Vec<String>,
pub prompt: Option<String>,
pub attempts: Vec<AttemptView>,
pub selected_attempt: usize,
pub current_view: DetailView,
pub base_turn_id: Option<String>,
pub sibling_turn_ids: Vec<String>,
pub attempt_total_hint: Option<usize>,
}
#[derive(Clone, Debug, Default)]
pub struct AttemptView {
pub turn_id: Option<String>,
pub status: codex_cloud_tasks_client::AttemptStatus,
pub attempt_placement: Option<i64>,
pub diff_lines: Vec<String>,
pub text_lines: Vec<String>,
pub prompt: Option<String>,
pub diff_raw: Option<String>,
}
impl AttemptView {
pub fn has_diff(&self) -> bool {
!self.diff_lines.is_empty()
}
pub fn has_text(&self) -> bool {
!self.text_lines.is_empty() || self.prompt.is_some()
}
}
impl DiffOverlay {
pub fn new(task_id: TaskId, title: String, attempt_total_hint: Option<usize>) -> Self {
let mut sd = ScrollableDiff::new();
sd.set_content(Vec::new());
Self {
title,
task_id,
sd,
base_can_apply: false,
diff_lines: Vec::new(),
text_lines: Vec::new(),
prompt: None,
attempts: vec![AttemptView::default()],
selected_attempt: 0,
current_view: DetailView::Prompt,
base_turn_id: None,
sibling_turn_ids: Vec::new(),
attempt_total_hint,
}
}
pub fn current_attempt(&self) -> Option<&AttemptView> {
self.attempts.get(self.selected_attempt)
}
pub fn base_attempt_mut(&mut self) -> &mut AttemptView {
if self.attempts.is_empty() {
self.attempts.push(AttemptView::default());
}
&mut self.attempts[0]
}
pub fn set_view(&mut self, view: DetailView) {
self.current_view = view;
self.apply_selection_to_fields();
}
pub fn expected_attempts(&self) -> Option<usize> {
self.attempt_total_hint.or({
if self.attempts.is_empty() {
None
} else {
Some(self.attempts.len())
}
})
}
pub fn attempt_count(&self) -> usize {
self.attempts.len()
}
pub fn attempt_display_total(&self) -> usize {
self.expected_attempts()
.unwrap_or_else(|| self.attempts.len().max(1))
}
pub fn step_attempt(&mut self, delta: isize) -> bool {
let total = self.attempts.len();
if total <= 1 {
return false;
}
let total_isize = total as isize;
let current = self.selected_attempt as isize;
let mut next = current + delta;
next = ((next % total_isize) + total_isize) % total_isize;
let next = next as usize;
self.selected_attempt = next;
self.apply_selection_to_fields();
true
}
pub fn current_can_apply(&self) -> bool {
matches!(self.current_view, DetailView::Diff)
&& self
.current_attempt()
.and_then(|attempt| attempt.diff_raw.as_ref())
.map(|diff| !diff.is_empty())
.unwrap_or(false)
}
pub fn apply_selection_to_fields(&mut self) {
let (diff_lines, text_lines, prompt) = if let Some(attempt) = self.current_attempt() {
(
attempt.diff_lines.clone(),
attempt.text_lines.clone(),
attempt.prompt.clone(),
)
} else {
self.diff_lines.clear();
self.text_lines.clear();
self.prompt = None;
self.sd.set_content(vec!["<loading attempt>".to_string()]);
return;
};
self.diff_lines = diff_lines.clone();
self.text_lines = text_lines.clone();
self.prompt = prompt;
match self.current_view {
DetailView::Diff => {
if diff_lines.is_empty() {
self.sd.set_content(vec!["<no diff available>".to_string()]);
} else {
self.sd.set_content(diff_lines);
}
}
DetailView::Prompt => {
if text_lines.is_empty() {
self.sd.set_content(vec!["<no output>".to_string()]);
} else {
self.sd.set_content(text_lines);
}
}
}
}
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum DetailView {
Diff,
Prompt,
}
/// Internal app events delivered from background tasks.
/// These let the UI event loop remain responsive and keep the spinner animating.
#[derive(Debug)]
pub enum AppEvent {
TasksLoaded {
env: Option<String>,
result: anyhow::Result<Vec<TaskSummary>>,
},
// Background diff summary events were planned; removed for now to keep code minimal.
/// Autodetection of a likely environment id finished
EnvironmentAutodetected(anyhow::Result<crate::env_detect::AutodetectSelection>),
/// Background completion of environment list fetch
EnvironmentsLoaded(anyhow::Result<Vec<EnvironmentRow>>),
DetailsDiffLoaded {
id: TaskId,
title: String,
diff: String,
},
DetailsMessagesLoaded {
id: TaskId,
title: String,
messages: Vec<String>,
prompt: Option<String>,
turn_id: Option<String>,
sibling_turn_ids: Vec<String>,
attempt_placement: Option<i64>,
attempt_status: codex_cloud_tasks_client::AttemptStatus,
},
DetailsFailed {
id: TaskId,
title: String,
error: String,
},
AttemptsLoaded {
id: TaskId,
attempts: Vec<codex_cloud_tasks_client::TurnAttempt>,
},
/// Background completion of new task submission
NewTaskSubmitted(Result<codex_cloud_tasks_client::CreatedTask, String>),
/// Background completion of apply preflight when opening modal or on demand
ApplyPreflightFinished {
id: TaskId,
title: String,
message: String,
level: ApplyResultLevel,
skipped: Vec<String>,
conflicts: Vec<String>,
},
/// Background completion of apply action (actual patch application)
ApplyFinished {
id: TaskId,
result: std::result::Result<codex_cloud_tasks_client::ApplyOutcome, String>,
},
}
// Convenience aliases; currently unused.
#[cfg(test)]
mod tests {
use super::*;
use chrono::Utc;
struct FakeBackend {
// maps env key to titles
by_env: std::collections::HashMap<Option<String>, Vec<&'static str>>,
}
#[async_trait::async_trait]
impl codex_cloud_tasks_client::CloudBackend for FakeBackend {
async fn list_tasks(
&self,
env: Option<&str>,
) -> codex_cloud_tasks_client::Result<Vec<TaskSummary>> {
let key = env.map(str::to_string);
let titles = self
.by_env
.get(&key)
.cloned()
.unwrap_or_else(|| vec!["default-a", "default-b"]);
let mut out = Vec::new();
for (i, t) in titles.into_iter().enumerate() {
out.push(TaskSummary {
id: TaskId(format!("T-{i}")),
title: t.to_string(),
status: codex_cloud_tasks_client::TaskStatus::Ready,
updated_at: Utc::now(),
environment_id: env.map(str::to_string),
environment_label: None,
summary: codex_cloud_tasks_client::DiffSummary::default(),
is_review: false,
attempt_total: Some(1),
});
}
Ok(out)
}
async fn get_task_diff(
&self,
_id: TaskId,
) -> codex_cloud_tasks_client::Result<Option<String>> {
Err(codex_cloud_tasks_client::CloudTaskError::Unimplemented(
"not used in test",
))
}
async fn get_task_messages(
&self,
_id: TaskId,
) -> codex_cloud_tasks_client::Result<Vec<String>> {
Ok(vec![])
}
async fn get_task_text(
&self,
_id: TaskId,
) -> codex_cloud_tasks_client::Result<codex_cloud_tasks_client::TaskText> {
Ok(codex_cloud_tasks_client::TaskText {
prompt: Some("Example prompt".to_string()),
messages: Vec::new(),
turn_id: Some("fake-turn".to_string()),
sibling_turn_ids: Vec::new(),
attempt_placement: Some(0),
attempt_status: codex_cloud_tasks_client::AttemptStatus::Completed,
})
}
async fn list_sibling_attempts(
&self,
_task: TaskId,
_turn_id: String,
) -> codex_cloud_tasks_client::Result<Vec<codex_cloud_tasks_client::TurnAttempt>> {
Ok(Vec::new())
}
async fn apply_task(
&self,
_id: TaskId,
_diff_override: Option<String>,
) -> codex_cloud_tasks_client::Result<codex_cloud_tasks_client::ApplyOutcome> {
Err(codex_cloud_tasks_client::CloudTaskError::Unimplemented(
"not used in test",
))
}
async fn apply_task_preflight(
&self,
_id: TaskId,
_diff_override: Option<String>,
) -> codex_cloud_tasks_client::Result<codex_cloud_tasks_client::ApplyOutcome> {
Err(codex_cloud_tasks_client::CloudTaskError::Unimplemented(
"not used in test",
))
}
async fn create_task(
&self,
_env_id: &str,
_prompt: &str,
_git_ref: &str,
_qa_mode: bool,
_best_of_n: usize,
) -> codex_cloud_tasks_client::Result<codex_cloud_tasks_client::CreatedTask> {
Err(codex_cloud_tasks_client::CloudTaskError::Unimplemented(
"not used in test",
))
}
}
#[tokio::test]
async fn load_tasks_uses_env_parameter() {
// Arrange: env-specific task titles
let mut by_env = std::collections::HashMap::new();
by_env.insert(None, vec!["root-1", "root-2"]);
by_env.insert(Some("env-A".to_string()), vec!["A-1"]);
by_env.insert(Some("env-B".to_string()), vec!["B-1", "B-2", "B-3"]);
let backend = FakeBackend { by_env };
// Act + Assert
let root = load_tasks(&backend, None).await.unwrap();
assert_eq!(root.len(), 2);
assert_eq!(root[0].title, "root-1");
let a = load_tasks(&backend, Some("env-A")).await.unwrap();
assert_eq!(a.len(), 1);
assert_eq!(a[0].title, "A-1");
let b = load_tasks(&backend, Some("env-B")).await.unwrap();
assert_eq!(b.len(), 3);
assert_eq!(b[2].title, "B-3");
}
}

View File

@@ -1,9 +0,0 @@
use clap::Parser;
use codex_common::CliConfigOverrides;
#[derive(Parser, Debug, Default)]
#[command(version)]
pub struct Cli {
#[clap(skip)]
pub config_overrides: CliConfigOverrides,
}

View File

@@ -1,361 +0,0 @@
use reqwest::header::CONTENT_TYPE;
use reqwest::header::HeaderMap;
use std::collections::HashMap;
use tracing::info;
use tracing::warn;
#[derive(Debug, Clone, serde::Deserialize)]
struct CodeEnvironment {
id: String,
#[serde(default)]
label: Option<String>,
#[serde(default)]
is_pinned: Option<bool>,
#[serde(default)]
task_count: Option<i64>,
}
#[derive(Debug, Clone)]
pub struct AutodetectSelection {
pub id: String,
pub label: Option<String>,
}
pub async fn autodetect_environment_id(
base_url: &str,
headers: &HeaderMap,
desired_label: Option<String>,
) -> anyhow::Result<AutodetectSelection> {
// 1) Try repo-specific environments based on local git origins (GitHub only, like VSCode)
let origins = get_git_origins();
crate::append_error_log(format!("env: git origins: {origins:?}"));
let mut by_repo_envs: Vec<CodeEnvironment> = Vec::new();
for origin in &origins {
if let Some((owner, repo)) = parse_owner_repo(origin) {
let url = if base_url.contains("/backend-api") {
format!(
"{}/wham/environments/by-repo/{}/{}/{}",
base_url, "github", owner, repo
)
} else {
format!(
"{}/api/codex/environments/by-repo/{}/{}/{}",
base_url, "github", owner, repo
)
};
crate::append_error_log(format!("env: GET {url}"));
match get_json::<Vec<CodeEnvironment>>(&url, headers).await {
Ok(mut list) => {
crate::append_error_log(format!(
"env: by-repo returned {} env(s) for {owner}/{repo}",
list.len(),
));
by_repo_envs.append(&mut list);
}
Err(e) => crate::append_error_log(format!(
"env: by-repo fetch failed for {owner}/{repo}: {e}"
)),
}
}
}
if let Some(env) = pick_environment_row(&by_repo_envs, desired_label.as_deref()) {
return Ok(AutodetectSelection {
id: env.id.clone(),
label: env.label.as_deref().map(str::to_owned),
});
}
// 2) Fallback to the full list
let list_url = if base_url.contains("/backend-api") {
format!("{base_url}/wham/environments")
} else {
format!("{base_url}/api/codex/environments")
};
crate::append_error_log(format!("env: GET {list_url}"));
// Fetch and log the full environments JSON for debugging
let http = reqwest::Client::builder().build()?;
let res = http.get(&list_url).headers(headers.clone()).send().await?;
let status = res.status();
let ct = res
.headers()
.get(CONTENT_TYPE)
.and_then(|v| v.to_str().ok())
.unwrap_or("")
.to_string();
let body = res.text().await.unwrap_or_default();
crate::append_error_log(format!("env: status={status} content-type={ct}"));
match serde_json::from_str::<serde_json::Value>(&body) {
Ok(v) => {
let pretty = serde_json::to_string_pretty(&v).unwrap_or(body.clone());
crate::append_error_log(format!("env: /environments JSON (pretty):\n{pretty}"));
}
Err(_) => crate::append_error_log(format!("env: /environments (raw):\n{body}")),
}
if !status.is_success() {
anyhow::bail!("GET {list_url} failed: {status}; content-type={ct}; body={body}");
}
let all_envs: Vec<CodeEnvironment> = serde_json::from_str(&body).map_err(|e| {
anyhow::anyhow!("Decode error for {list_url}: {e}; content-type={ct}; body={body}")
})?;
if let Some(env) = pick_environment_row(&all_envs, desired_label.as_deref()) {
return Ok(AutodetectSelection {
id: env.id.clone(),
label: env.label.as_deref().map(str::to_owned),
});
}
anyhow::bail!("no environments available")
}
fn pick_environment_row(
envs: &[CodeEnvironment],
desired_label: Option<&str>,
) -> Option<CodeEnvironment> {
if envs.is_empty() {
return None;
}
if let Some(label) = desired_label {
let lc = label.to_lowercase();
if let Some(e) = envs
.iter()
.find(|e| e.label.as_deref().unwrap_or("").to_lowercase() == lc)
{
crate::append_error_log(format!("env: matched by label: {label} -> {}", e.id));
return Some(e.clone());
}
}
if envs.len() == 1 {
crate::append_error_log("env: single environment available; selecting it");
return Some(envs[0].clone());
}
if let Some(e) = envs.iter().find(|e| e.is_pinned.unwrap_or(false)) {
crate::append_error_log(format!("env: selecting pinned environment: {}", e.id));
return Some(e.clone());
}
// Highest task_count as heuristic
if let Some(e) = envs
.iter()
.max_by_key(|e| e.task_count.unwrap_or(0))
.or_else(|| envs.first())
{
crate::append_error_log(format!("env: selecting by task_count/first: {}", e.id));
return Some(e.clone());
}
None
}
async fn get_json<T: serde::de::DeserializeOwned>(
url: &str,
headers: &HeaderMap,
) -> anyhow::Result<T> {
let http = reqwest::Client::builder().build()?;
let res = http.get(url).headers(headers.clone()).send().await?;
let status = res.status();
let ct = res
.headers()
.get(CONTENT_TYPE)
.and_then(|v| v.to_str().ok())
.unwrap_or("")
.to_string();
let body = res.text().await.unwrap_or_default();
crate::append_error_log(format!("env: status={status} content-type={ct}"));
if !status.is_success() {
anyhow::bail!("GET {url} failed: {status}; content-type={ct}; body={body}");
}
let parsed = serde_json::from_str::<T>(&body).map_err(|e| {
anyhow::anyhow!("Decode error for {url}: {e}; content-type={ct}; body={body}")
})?;
Ok(parsed)
}
fn get_git_origins() -> Vec<String> {
// Prefer: git config --get-regexp remote\..*\.url
let out = std::process::Command::new("git")
.args(["config", "--get-regexp", "remote\\..*\\.url"])
.output();
if let Ok(ok) = out
&& ok.status.success()
{
let s = String::from_utf8_lossy(&ok.stdout);
let mut urls = Vec::new();
for line in s.lines() {
if let Some((_, url)) = line.split_once(' ') {
urls.push(url.trim().to_string());
}
}
if !urls.is_empty() {
return uniq(urls);
}
}
// Fallback: git remote -v
let out = std::process::Command::new("git")
.args(["remote", "-v"])
.output();
if let Ok(ok) = out
&& ok.status.success()
{
let s = String::from_utf8_lossy(&ok.stdout);
let mut urls = Vec::new();
for line in s.lines() {
let parts: Vec<&str> = line.split_whitespace().collect();
if parts.len() >= 2 {
urls.push(parts[1].to_string());
}
}
if !urls.is_empty() {
return uniq(urls);
}
}
Vec::new()
}
fn uniq(mut v: Vec<String>) -> Vec<String> {
v.sort();
v.dedup();
v
}
fn parse_owner_repo(url: &str) -> Option<(String, String)> {
// Normalize common prefixes and handle multiple SSH/HTTPS variants.
let mut s = url.trim().to_string();
// Drop protocol scheme for ssh URLs
if let Some(rest) = s.strip_prefix("ssh://") {
s = rest.to_string();
}
// Accept any user before @github.com (e.g., git@, org-123@)
if let Some(idx) = s.find("@github.com:") {
let rest = &s[idx + "@github.com:".len()..];
let rest = rest.trim_start_matches('/').trim_end_matches(".git");
let mut parts = rest.splitn(2, '/');
let owner = parts.next()?.to_string();
let repo = parts.next()?.to_string();
crate::append_error_log(format!("env: parsed SSH GitHub origin => {owner}/{repo}"));
return Some((owner, repo));
}
// HTTPS or git protocol
for prefix in [
"https://github.com/",
"http://github.com/",
"git://github.com/",
"github.com/",
] {
if let Some(rest) = s.strip_prefix(prefix) {
let rest = rest.trim_start_matches('/').trim_end_matches(".git");
let mut parts = rest.splitn(2, '/');
let owner = parts.next()?.to_string();
let repo = parts.next()?.to_string();
crate::append_error_log(format!("env: parsed HTTP GitHub origin => {owner}/{repo}"));
return Some((owner, repo));
}
}
None
}
/// List environments for the current repo(s) with a fallback to the global list.
/// Returns a de-duplicated, sorted set suitable for the TUI modal.
pub async fn list_environments(
base_url: &str,
headers: &HeaderMap,
) -> anyhow::Result<Vec<crate::app::EnvironmentRow>> {
let mut map: HashMap<String, crate::app::EnvironmentRow> = HashMap::new();
// 1) By-repo lookup for each parsed GitHub origin
let origins = get_git_origins();
for origin in &origins {
if let Some((owner, repo)) = parse_owner_repo(origin) {
let url = if base_url.contains("/backend-api") {
format!(
"{}/wham/environments/by-repo/{}/{}/{}",
base_url, "github", owner, repo
)
} else {
format!(
"{}/api/codex/environments/by-repo/{}/{}/{}",
base_url, "github", owner, repo
)
};
match get_json::<Vec<CodeEnvironment>>(&url, headers).await {
Ok(list) => {
info!("env_tui: by-repo {}:{} -> {} envs", owner, repo, list.len());
for e in list {
let entry =
map.entry(e.id.clone())
.or_insert_with(|| crate::app::EnvironmentRow {
id: e.id.clone(),
label: e.label.clone(),
is_pinned: e.is_pinned.unwrap_or(false),
repo_hints: Some(format!("{owner}/{repo}")),
});
// Merge: keep label if present, or use new; accumulate pinned flag
if entry.label.is_none() {
entry.label = e.label.clone();
}
entry.is_pinned = entry.is_pinned || e.is_pinned.unwrap_or(false);
if entry.repo_hints.is_none() {
entry.repo_hints = Some(format!("{owner}/{repo}"));
}
}
}
Err(e) => {
warn!(
"env_tui: by-repo fetch failed for {}/{}: {}",
owner, repo, e
);
}
}
}
}
// 2) Fallback to the full list; on error return what we have if any.
let list_url = if base_url.contains("/backend-api") {
format!("{base_url}/wham/environments")
} else {
format!("{base_url}/api/codex/environments")
};
match get_json::<Vec<CodeEnvironment>>(&list_url, headers).await {
Ok(list) => {
info!("env_tui: global list -> {} envs", list.len());
for e in list {
let entry = map
.entry(e.id.clone())
.or_insert_with(|| crate::app::EnvironmentRow {
id: e.id.clone(),
label: e.label.clone(),
is_pinned: e.is_pinned.unwrap_or(false),
repo_hints: None,
});
if entry.label.is_none() {
entry.label = e.label.clone();
}
entry.is_pinned = entry.is_pinned || e.is_pinned.unwrap_or(false);
}
}
Err(e) => {
if map.is_empty() {
return Err(e);
} else {
warn!(
"env_tui: global list failed; using by-repo results only: {}",
e
);
}
}
}
let mut rows: Vec<crate::app::EnvironmentRow> = map.into_values().collect();
rows.sort_by(|a, b| {
// pinned first
let p = b.is_pinned.cmp(&a.is_pinned);
if p != std::cmp::Ordering::Equal {
return p;
}
// then label (ci), then id
let al = a.label.as_deref().unwrap_or("").to_lowercase();
let bl = b.label.as_deref().unwrap_or("").to_lowercase();
let l = al.cmp(&bl);
if l != std::cmp::Ordering::Equal {
return l;
}
a.id.cmp(&b.id)
});
Ok(rows)
}

File diff suppressed because it is too large Load Diff

View File

@@ -1,35 +0,0 @@
use codex_tui::ComposerInput;
pub struct NewTaskPage {
pub composer: ComposerInput,
pub submitting: bool,
pub env_id: Option<String>,
pub best_of_n: usize,
}
impl NewTaskPage {
pub fn new(env_id: Option<String>, best_of_n: usize) -> Self {
let mut composer = ComposerInput::new();
composer.set_hint_items(vec![
("", "send"),
("Shift+⏎", "newline"),
("Ctrl+O", "env"),
("Ctrl+N", "attempts"),
("Ctrl+C", "quit"),
]);
Self {
composer,
submitting: false,
env_id,
best_of_n,
}
}
// Additional helpers can be added as usage evolves.
}
impl Default for NewTaskPage {
fn default() -> Self {
Self::new(None, 1)
}
}

View File

@@ -1,176 +0,0 @@
use unicode_width::UnicodeWidthChar;
use unicode_width::UnicodeWidthStr;
/// Scroll position and geometry for a vertical scroll view.
#[derive(Clone, Copy, Debug, Default)]
pub struct ScrollViewState {
pub scroll: u16,
pub viewport_h: u16,
pub content_h: u16,
}
impl ScrollViewState {
pub fn clamp(&mut self) {
let max_scroll = self.content_h.saturating_sub(self.viewport_h);
if self.scroll > max_scroll {
self.scroll = max_scroll;
}
}
}
/// A simple, local scrollable view for diffs or message text.
///
/// Owns raw lines, caches wrapped lines for a given width, and maintains
/// a small scroll state that is clamped whenever geometry shrinks.
#[derive(Clone, Debug, Default)]
pub struct ScrollableDiff {
raw: Vec<String>,
wrapped: Vec<String>,
wrapped_src_idx: Vec<usize>,
wrap_cols: Option<u16>,
pub state: ScrollViewState,
}
impl ScrollableDiff {
pub fn new() -> Self {
Self::default()
}
/// Replace the raw content lines. Does not rewrap immediately; call `set_width` next.
pub fn set_content(&mut self, lines: Vec<String>) {
self.raw = lines;
self.wrapped.clear();
self.wrapped_src_idx.clear();
self.state.content_h = 0;
// Force rewrap on next set_width even if width is unchanged
self.wrap_cols = None;
}
/// Set the wrap width. If changed, rebuild wrapped lines and clamp scroll.
pub fn set_width(&mut self, width: u16) {
if self.wrap_cols == Some(width) {
return;
}
self.wrap_cols = Some(width);
self.rewrap(width);
self.state.clamp();
}
/// Update viewport height and clamp scroll if needed.
pub fn set_viewport(&mut self, height: u16) {
self.state.viewport_h = height;
self.state.clamp();
}
/// Return the cached wrapped lines. Call `set_width` first when area changes.
pub fn wrapped_lines(&self) -> &[String] {
&self.wrapped
}
pub fn wrapped_src_indices(&self) -> &[usize] {
&self.wrapped_src_idx
}
pub fn raw_line_at(&self, idx: usize) -> &str {
self.raw.get(idx).map(String::as_str).unwrap_or("")
}
/// Scroll by a signed delta; clamps to content.
pub fn scroll_by(&mut self, delta: i16) {
let s = self.state.scroll as i32 + delta as i32;
self.state.scroll = s.clamp(0, self.max_scroll() as i32) as u16;
}
/// Page by a signed delta; typically viewport_h - 1.
pub fn page_by(&mut self, delta: i16) {
self.scroll_by(delta);
}
pub fn to_top(&mut self) {
self.state.scroll = 0;
}
pub fn to_bottom(&mut self) {
self.state.scroll = self.max_scroll();
}
/// Optional percent scrolled; None when not enough geometry is known.
pub fn percent_scrolled(&self) -> Option<u8> {
if self.state.content_h == 0 || self.state.viewport_h == 0 {
return None;
}
if self.state.content_h <= self.state.viewport_h {
return None;
}
let visible_bottom = self.state.scroll.saturating_add(self.state.viewport_h) as f32;
let pct = (visible_bottom / self.state.content_h as f32 * 100.0).round();
Some(pct.clamp(0.0, 100.0) as u8)
}
fn max_scroll(&self) -> u16 {
self.state.content_h.saturating_sub(self.state.viewport_h)
}
fn rewrap(&mut self, width: u16) {
if width == 0 {
self.wrapped = self.raw.clone();
self.state.content_h = self.wrapped.len() as u16;
return;
}
let max_cols = width as usize;
let mut out: Vec<String> = Vec::new();
let mut out_idx: Vec<usize> = Vec::new();
for (raw_idx, raw) in self.raw.iter().enumerate() {
// Normalize tabs for width accounting (MVP: 4 spaces).
let raw = raw.replace('\t', " ");
if raw.is_empty() {
out.push(String::new());
out_idx.push(raw_idx);
continue;
}
let mut line = String::new();
let mut line_cols = 0usize;
let mut last_soft_idx: Option<usize> = None; // last whitespace or punctuation break
for (_i, ch) in raw.char_indices() {
if ch == '\n' {
out.push(std::mem::take(&mut line));
out_idx.push(raw_idx);
line_cols = 0;
last_soft_idx = None;
continue;
}
let w = UnicodeWidthChar::width(ch).unwrap_or(0);
if line_cols.saturating_add(w) > max_cols {
if let Some(split) = last_soft_idx {
let (prefix, rest) = line.split_at(split);
out.push(prefix.trim_end().to_string());
out_idx.push(raw_idx);
line = rest.trim_start().to_string();
last_soft_idx = None;
// retry add current ch now that line may be shorter
} else if !line.is_empty() {
out.push(std::mem::take(&mut line));
out_idx.push(raw_idx);
}
}
if ch.is_whitespace()
|| matches!(
ch,
',' | ';' | '.' | ':' | ')' | ']' | '}' | '|' | '/' | '?' | '!' | '-' | '_'
)
{
last_soft_idx = Some(line.len());
}
line.push(ch);
line_cols = UnicodeWidthStr::width(line.as_str());
}
if !line.is_empty() {
out.push(line);
out_idx.push(raw_idx);
}
}
self.wrapped = out;
self.wrapped_src_idx = out_idx;
self.state.content_h = self.wrapped.len() as u16;
}
}

File diff suppressed because it is too large Load Diff

View File

@@ -1,93 +0,0 @@
use base64::Engine as _;
use chrono::Utc;
use reqwest::header::HeaderMap;
pub fn set_user_agent_suffix(suffix: &str) {
if let Ok(mut guard) = codex_core::default_client::USER_AGENT_SUFFIX.lock() {
guard.replace(suffix.to_string());
}
}
pub fn append_error_log(message: impl AsRef<str>) {
let ts = Utc::now().to_rfc3339();
if let Ok(mut f) = std::fs::OpenOptions::new()
.create(true)
.append(true)
.open("error.log")
{
use std::io::Write as _;
let _ = writeln!(f, "[{ts}] {}", message.as_ref());
}
}
/// Normalize the configured base URL to a canonical form used by the backend client.
/// - trims trailing '/'
/// - appends '/backend-api' for ChatGPT hosts when missing
pub fn normalize_base_url(input: &str) -> String {
let mut base_url = input.to_string();
while base_url.ends_with('/') {
base_url.pop();
}
if (base_url.starts_with("https://chatgpt.com")
|| base_url.starts_with("https://chat.openai.com"))
&& !base_url.contains("/backend-api")
{
base_url = format!("{base_url}/backend-api");
}
base_url
}
/// Extract the ChatGPT account id from a JWT token, when present.
pub fn extract_chatgpt_account_id(token: &str) -> Option<String> {
let mut parts = token.split('.');
let (_h, payload_b64, _s) = match (parts.next(), parts.next(), parts.next()) {
(Some(h), Some(p), Some(s)) if !h.is_empty() && !p.is_empty() && !s.is_empty() => (h, p, s),
_ => return None,
};
let payload_bytes = base64::engine::general_purpose::URL_SAFE_NO_PAD
.decode(payload_b64)
.ok()?;
let v: serde_json::Value = serde_json::from_slice(&payload_bytes).ok()?;
v.get("https://api.openai.com/auth")
.and_then(|auth| auth.get("chatgpt_account_id"))
.and_then(|id| id.as_str())
.map(str::to_string)
}
/// Build headers for ChatGPT-backed requests: `User-Agent`, optional `Authorization`,
/// and optional `ChatGPT-Account-Id`.
pub async fn build_chatgpt_headers() -> HeaderMap {
use reqwest::header::AUTHORIZATION;
use reqwest::header::HeaderName;
use reqwest::header::HeaderValue;
use reqwest::header::USER_AGENT;
set_user_agent_suffix("codex_cloud_tasks_tui");
let ua = codex_core::default_client::get_codex_user_agent();
let mut headers = HeaderMap::new();
headers.insert(
USER_AGENT,
HeaderValue::from_str(&ua).unwrap_or(HeaderValue::from_static("codex-cli")),
);
if let Ok(home) = codex_core::config::find_codex_home() {
let am = codex_login::AuthManager::new(home);
if let Some(auth) = am.auth()
&& let Ok(tok) = auth.get_token().await
&& !tok.is_empty()
{
let v = format!("Bearer {tok}");
if let Ok(hv) = HeaderValue::from_str(&v) {
headers.insert(AUTHORIZATION, hv);
}
if let Some(acc) = auth
.get_account_id()
.or_else(|| extract_chatgpt_account_id(&tok))
&& let Ok(name) = HeaderName::from_bytes(b"ChatGPT-Account-Id")
&& let Ok(hv) = HeaderValue::from_str(&acc)
{
headers.insert(name, hv);
}
}
}
headers
}

View File

@@ -1,22 +0,0 @@
use codex_cloud_tasks_client::CloudBackend;
use codex_cloud_tasks_client::MockClient;
#[tokio::test]
async fn mock_backend_varies_by_env() {
let client = MockClient;
let root = CloudBackend::list_tasks(&client, None).await.unwrap();
assert!(root.iter().any(|t| t.title.contains("Update README")));
let a = CloudBackend::list_tasks(&client, Some("env-A"))
.await
.unwrap();
assert_eq!(a.len(), 1);
assert_eq!(a[0].title, "A: First");
let b = CloudBackend::list_tasks(&client, Some("env-B"))
.await
.unwrap();
assert_eq!(b.len(), 2);
assert!(b[0].title.starts_with("B: "));
}

View File

@@ -1,17 +0,0 @@
[package]
name = "codex-backend-openapi-models"
version = { workspace = true }
edition = "2024"
[lib]
name = "codex_backend_openapi_models"
path = "src/lib.rs"
# Important: generated code often violates our workspace lints.
# Allow unwrap/expect in this crate so the workspace builds cleanly
# after models are regenerated.
# Lint overrides are applied in src/lib.rs via crate attributes
[dependencies]
serde = { version = "1", features = ["derive"] }
serde_json = "1"

View File

@@ -1,6 +0,0 @@
#![allow(clippy::unwrap_used, clippy::expect_used)]
// Re-export generated OpenAPI models.
// The regen script populates `src/models/*.rs` and writes `src/models/mod.rs`.
// This module intentionally contains no hand-written types.
pub mod models;

View File

@@ -1,42 +0,0 @@
/*
* codex-backend
*
* codex-backend
*
* The version of the OpenAPI document: 0.0.1
*
* Generated by: https://openapi-generator.tech
*/
use crate::models;
use serde::Deserialize;
use serde::Serialize;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct CodeTaskDetailsResponse {
#[serde(rename = "task")]
pub task: Box<models::TaskResponse>,
#[serde(rename = "current_user_turn", skip_serializing_if = "Option::is_none")]
pub current_user_turn: Option<std::collections::HashMap<String, serde_json::Value>>,
#[serde(
rename = "current_assistant_turn",
skip_serializing_if = "Option::is_none"
)]
pub current_assistant_turn: Option<std::collections::HashMap<String, serde_json::Value>>,
#[serde(
rename = "current_diff_task_turn",
skip_serializing_if = "Option::is_none"
)]
pub current_diff_task_turn: Option<std::collections::HashMap<String, serde_json::Value>>,
}
impl CodeTaskDetailsResponse {
pub fn new(task: models::TaskResponse) -> CodeTaskDetailsResponse {
CodeTaskDetailsResponse {
task: Box::new(task),
current_user_turn: None,
current_assistant_turn: None,
current_diff_task_turn: None,
}
}
}

View File

@@ -1,40 +0,0 @@
/*
* codex-backend
*
* codex-backend
*
* The version of the OpenAPI document: 0.0.1
*
* Generated by: https://openapi-generator.tech
*/
use crate::models;
use serde::Deserialize;
use serde::Serialize;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct ExternalPullRequestResponse {
#[serde(rename = "id")]
pub id: String,
#[serde(rename = "assistant_turn_id")]
pub assistant_turn_id: String,
#[serde(rename = "pull_request")]
pub pull_request: Box<models::GitPullRequest>,
#[serde(rename = "codex_updated_sha", skip_serializing_if = "Option::is_none")]
pub codex_updated_sha: Option<String>,
}
impl ExternalPullRequestResponse {
pub fn new(
id: String,
assistant_turn_id: String,
pull_request: models::GitPullRequest,
) -> ExternalPullRequestResponse {
ExternalPullRequestResponse {
id,
assistant_turn_id,
pull_request: Box::new(pull_request),
codex_updated_sha: None,
}
}
}

View File

@@ -1,77 +0,0 @@
/*
* codex-backend
*
* codex-backend
*
* The version of the OpenAPI document: 0.0.1
*
* Generated by: https://openapi-generator.tech
*/
use serde::Deserialize;
use serde::Serialize;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct GitPullRequest {
#[serde(rename = "number")]
pub number: i32,
#[serde(rename = "url")]
pub url: String,
#[serde(rename = "state")]
pub state: String,
#[serde(rename = "merged")]
pub merged: bool,
#[serde(rename = "mergeable")]
pub mergeable: bool,
#[serde(rename = "draft", skip_serializing_if = "Option::is_none")]
pub draft: Option<bool>,
#[serde(rename = "title", skip_serializing_if = "Option::is_none")]
pub title: Option<String>,
#[serde(rename = "body", skip_serializing_if = "Option::is_none")]
pub body: Option<String>,
#[serde(rename = "base", skip_serializing_if = "Option::is_none")]
pub base: Option<String>,
#[serde(rename = "head", skip_serializing_if = "Option::is_none")]
pub head: Option<String>,
#[serde(rename = "base_sha", skip_serializing_if = "Option::is_none")]
pub base_sha: Option<String>,
#[serde(rename = "head_sha", skip_serializing_if = "Option::is_none")]
pub head_sha: Option<String>,
#[serde(rename = "merge_commit_sha", skip_serializing_if = "Option::is_none")]
pub merge_commit_sha: Option<String>,
#[serde(rename = "comments", skip_serializing_if = "Option::is_none")]
pub comments: Option<serde_json::Value>,
#[serde(rename = "diff", skip_serializing_if = "Option::is_none")]
pub diff: Option<serde_json::Value>,
#[serde(rename = "user", skip_serializing_if = "Option::is_none")]
pub user: Option<serde_json::Value>,
}
impl GitPullRequest {
pub fn new(
number: i32,
url: String,
state: String,
merged: bool,
mergeable: bool,
) -> GitPullRequest {
GitPullRequest {
number,
url,
state,
merged,
mergeable,
draft: None,
title: None,
body: None,
base: None,
head: None,
base_sha: None,
head_sha: None,
merge_commit_sha: None,
comments: None,
diff: None,
user: None,
}
}
}

View File

@@ -1,22 +0,0 @@
// Curated minimal export list for current workspace usage.
// NOTE: This file was previously auto-generated by the OpenAPI generator.
// Currently export only the types referenced by the workspace
// The process for this will change
pub mod code_task_details_response;
pub use self::code_task_details_response::CodeTaskDetailsResponse;
pub mod task_response;
pub use self::task_response::TaskResponse;
pub mod external_pull_request_response;
pub use self::external_pull_request_response::ExternalPullRequestResponse;
pub mod git_pull_request;
pub use self::git_pull_request::GitPullRequest;
pub mod task_list_item;
pub use self::task_list_item::TaskListItem;
pub mod paginated_list_task_list_item_;
pub use self::paginated_list_task_list_item_::PaginatedListTaskListItem;

View File

@@ -1,30 +0,0 @@
/*
* codex-backend
*
* codex-backend
*
* The version of the OpenAPI document: 0.0.1
*
* Generated by: https://openapi-generator.tech
*/
use crate::models;
use serde::Deserialize;
use serde::Serialize;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct PaginatedListTaskListItem {
#[serde(rename = "items")]
pub items: Vec<models::TaskListItem>,
#[serde(rename = "cursor", skip_serializing_if = "Option::is_none")]
pub cursor: Option<String>,
}
impl PaginatedListTaskListItem {
pub fn new(items: Vec<models::TaskListItem>) -> PaginatedListTaskListItem {
PaginatedListTaskListItem {
items,
cursor: None,
}
}
}

View File

@@ -1,63 +0,0 @@
/*
* codex-backend
*
* codex-backend
*
* The version of the OpenAPI document: 0.0.1
*
* Generated by: https://openapi-generator.tech
*/
use crate::models;
use serde::Deserialize;
use serde::Serialize;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct TaskListItem {
#[serde(rename = "id")]
pub id: String,
#[serde(rename = "title")]
pub title: String,
#[serde(
rename = "has_generated_title",
skip_serializing_if = "Option::is_none"
)]
pub has_generated_title: Option<bool>,
#[serde(rename = "updated_at", skip_serializing_if = "Option::is_none")]
pub updated_at: Option<f64>,
#[serde(rename = "created_at", skip_serializing_if = "Option::is_none")]
pub created_at: Option<f64>,
#[serde(
rename = "task_status_display",
skip_serializing_if = "Option::is_none"
)]
pub task_status_display: Option<std::collections::HashMap<String, serde_json::Value>>,
#[serde(rename = "archived")]
pub archived: bool,
#[serde(rename = "has_unread_turn")]
pub has_unread_turn: bool,
#[serde(rename = "pull_requests", skip_serializing_if = "Option::is_none")]
pub pull_requests: Option<Vec<models::ExternalPullRequestResponse>>,
}
impl TaskListItem {
pub fn new(
id: String,
title: String,
has_generated_title: Option<bool>,
archived: bool,
has_unread_turn: bool,
) -> TaskListItem {
TaskListItem {
id,
title,
has_generated_title,
updated_at: None,
created_at: None,
task_status_display: None,
archived,
has_unread_turn,
pull_requests: None,
}
}
}

View File

@@ -1,62 +0,0 @@
/*
* codex-backend
*
* codex-backend
*
* The version of the OpenAPI document: 0.0.1
*
* Generated by: https://openapi-generator.tech
*/
use crate::models;
use serde::Deserialize;
use serde::Serialize;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct TaskResponse {
#[serde(rename = "id")]
pub id: String,
#[serde(rename = "created_at", skip_serializing_if = "Option::is_none")]
pub created_at: Option<f64>,
#[serde(rename = "title")]
pub title: String,
#[serde(
rename = "has_generated_title",
skip_serializing_if = "Option::is_none"
)]
pub has_generated_title: Option<bool>,
#[serde(rename = "current_turn_id", skip_serializing_if = "Option::is_none")]
pub current_turn_id: Option<String>,
#[serde(rename = "has_unread_turn", skip_serializing_if = "Option::is_none")]
pub has_unread_turn: Option<bool>,
#[serde(
rename = "denormalized_metadata",
skip_serializing_if = "Option::is_none"
)]
pub denormalized_metadata: Option<std::collections::HashMap<String, serde_json::Value>>,
#[serde(rename = "archived")]
pub archived: bool,
#[serde(rename = "external_pull_requests")]
pub external_pull_requests: Vec<models::ExternalPullRequestResponse>,
}
impl TaskResponse {
pub fn new(
id: String,
title: String,
archived: bool,
external_pull_requests: Vec<models::ExternalPullRequestResponse>,
) -> TaskResponse {
TaskResponse {
id,
created_at: None,
title,
has_generated_title: None,
current_turn_id: None,
has_unread_turn: None,
denormalized_metadata: None,
archived,
external_pull_requests,
}
}
}

View File

@@ -10,7 +10,6 @@ workspace = true
clap = { workspace = true, features = ["derive", "wrap_help"], optional = true }
codex-core = { workspace = true }
codex-protocol = { workspace = true }
codex-app-server-protocol = { workspace = true }
serde = { workspace = true, optional = true }
toml = { workspace = true, optional = true }

View File

@@ -1,5 +1,5 @@
use codex_app_server_protocol::AuthMode;
use codex_core::protocol_config_types::ReasoningEffort;
use codex_protocol::mcp_protocol::AuthMode;
/// A simple preset pairing a model slug with a reasoning effort.
#[derive(Debug, Clone, Copy)]

View File

@@ -24,10 +24,7 @@ codex-file-search = { workspace = true }
codex-mcp-client = { workspace = true }
codex-rmcp-client = { workspace = true }
codex-protocol = { workspace = true }
codex-app-server-protocol = { workspace = true }
codex-otel = { workspace = true, features = ["otel"] }
dirs = { workspace = true }
dunce = { workspace = true }
env-flags = { workspace = true }
eventsource-stream = { workspace = true }
futures = { workspace = true }
@@ -94,7 +91,6 @@ tempfile = { workspace = true }
tokio-test = { workspace = true }
walkdir = { workspace = true }
wiremock = { workspace = true }
tracing-test = { workspace = true, features = ["no-env-filter"] }
[package.metadata.cargo-shear]
ignored = ["openssl-sys"]

View File

@@ -27,7 +27,6 @@ pub(crate) enum InternalApplyPatchInvocation {
DelegateToExec(ApplyPatchExec),
}
#[derive(Debug)]
pub(crate) struct ApplyPatchExec {
pub(crate) action: ApplyPatchAction,
pub(crate) user_explicitly_approved_this_action: bool,
@@ -46,13 +45,12 @@ pub(crate) async fn apply_patch(
&turn_context.sandbox_policy,
&turn_context.cwd,
) {
SafetyCheck::AutoApprove {
user_explicitly_approved,
..
} => InternalApplyPatchInvocation::DelegateToExec(ApplyPatchExec {
action,
user_explicitly_approved_this_action: user_explicitly_approved,
}),
SafetyCheck::AutoApprove { .. } => {
InternalApplyPatchInvocation::DelegateToExec(ApplyPatchExec {
action,
user_explicitly_approved_this_action: false,
})
}
SafetyCheck::AskUser => {
// Compute a readable summary of path changes to include in the
// approval request so the user can make an informed decision.
@@ -110,28 +108,3 @@ pub(crate) fn convert_apply_patch_to_protocol(
}
result
}
#[cfg(test)]
mod tests {
use super::*;
use pretty_assertions::assert_eq;
use tempfile::tempdir;
#[test]
fn convert_apply_patch_maps_add_variant() {
let tmp = tempdir().expect("tmp");
let p = tmp.path().join("a.txt");
// Create an action with a single Add change
let action = ApplyPatchAction::new_add_for_test(&p, "hello".to_string());
let got = convert_apply_patch_to_protocol(&action);
assert_eq!(
got.get(&p),
Some(&FileChange::Add {
content: "hello".to_string()
})
);
}
}

View File

@@ -15,7 +15,7 @@ use std::sync::Arc;
use std::sync::Mutex;
use std::time::Duration;
use codex_app_server_protocol::AuthMode;
use codex_protocol::mcp_protocol::AuthMode;
use crate::token_data::PlanType;
use crate::token_data::TokenData;

View File

@@ -1,21 +1,6 @@
use std::time::Duration;
use crate::ModelProviderInfo;
use crate::client_common::Prompt;
use crate::client_common::ResponseEvent;
use crate::client_common::ResponseStream;
use crate::error::CodexErr;
use crate::error::Result;
use crate::error::RetryLimitReachedError;
use crate::error::UnexpectedResponseError;
use crate::model_family::ModelFamily;
use crate::openai_tools::create_tools_json_for_chat_completions_api;
use crate::util::backoff;
use bytes::Bytes;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ReasoningItemContent;
use codex_protocol::models::ResponseItem;
use eventsource_stream::Eventsource;
use futures::Stream;
use futures::StreamExt;
@@ -30,13 +15,25 @@ use tokio::time::timeout;
use tracing::debug;
use tracing::trace;
use crate::ModelProviderInfo;
use crate::client_common::Prompt;
use crate::client_common::ResponseEvent;
use crate::client_common::ResponseStream;
use crate::error::CodexErr;
use crate::error::Result;
use crate::model_family::ModelFamily;
use crate::openai_tools::create_tools_json_for_chat_completions_api;
use crate::util::backoff;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ReasoningItemContent;
use codex_protocol::models::ResponseItem;
/// Implementation for the classic Chat Completions API.
pub(crate) async fn stream_chat_completions(
prompt: &Prompt,
model_family: &ModelFamily,
client: &reqwest::Client,
provider: &ModelProviderInfo,
otel_event_manager: &OtelEventManager,
) -> Result<ResponseStream> {
if prompt.output_schema.is_some() {
return Err(CodexErr::UnsupportedOperation(
@@ -297,13 +294,10 @@ pub(crate) async fn stream_chat_completions(
let req_builder = provider.create_request_builder(client, &None).await?;
let res = otel_event_manager
.log_request(attempt, || {
req_builder
.header(reqwest::header::ACCEPT, "text/event-stream")
.json(&payload)
.send()
})
let res = req_builder
.header(reqwest::header::ACCEPT, "text/event-stream")
.json(&payload)
.send()
.await;
match res {
@@ -314,7 +308,6 @@ pub(crate) async fn stream_chat_completions(
stream,
tx_event,
provider.stream_idle_timeout(),
otel_event_manager.clone(),
));
return Ok(ResponseStream { rx_event });
}
@@ -322,18 +315,11 @@ pub(crate) async fn stream_chat_completions(
let status = res.status();
if !(status == StatusCode::TOO_MANY_REQUESTS || status.is_server_error()) {
let body = (res.text().await).unwrap_or_default();
return Err(CodexErr::UnexpectedStatus(UnexpectedResponseError {
status,
body,
request_id: None,
}));
return Err(CodexErr::UnexpectedStatus(status, body));
}
if attempt > max_retries {
return Err(CodexErr::RetryLimit(RetryLimitReachedError {
status,
request_id: None,
}));
return Err(CodexErr::RetryLimit(status));
}
let retry_after_secs = res
@@ -365,7 +351,6 @@ async fn process_chat_sse<S>(
stream: S,
tx_event: mpsc::Sender<Result<ResponseEvent>>,
idle_timeout: Duration,
otel_event_manager: OtelEventManager,
) where
S: Stream<Item = Result<Bytes>> + Unpin,
{
@@ -389,10 +374,7 @@ async fn process_chat_sse<S>(
let mut reasoning_text = String::new();
loop {
let sse = match otel_event_manager
.log_sse_event(|| timeout(idle_timeout, stream.next()))
.await
{
let sse = match timeout(idle_timeout, stream.next()).await {
Ok(Some(Ok(ev))) => ev,
Ok(Some(Err(e))) => {
let _ = tx_event

View File

@@ -5,11 +5,9 @@ use std::time::Duration;
use crate::AuthManager;
use crate::auth::CodexAuth;
use crate::error::RetryLimitReachedError;
use crate::error::UnexpectedResponseError;
use bytes::Bytes;
use codex_app_server_protocol::AuthMode;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::AuthMode;
use codex_protocol::mcp_protocol::ConversationId;
use eventsource_stream::Eventsource;
use futures::prelude::*;
use regex_lite::Regex;
@@ -49,7 +47,6 @@ use crate::protocol::RateLimitWindow;
use crate::protocol::TokenUsage;
use crate::token_data::PlanType;
use crate::util::backoff;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_protocol::config_types::ReasoningEffort as ReasoningEffortConfig;
use codex_protocol::config_types::ReasoningSummary as ReasoningSummaryConfig;
use codex_protocol::models::ResponseItem;
@@ -76,7 +73,6 @@ struct Error {
pub struct ModelClient {
config: Arc<Config>,
auth_manager: Option<Arc<AuthManager>>,
otel_event_manager: OtelEventManager,
client: reqwest::Client,
provider: ModelProviderInfo,
conversation_id: ConversationId,
@@ -88,7 +84,6 @@ impl ModelClient {
pub fn new(
config: Arc<Config>,
auth_manager: Option<Arc<AuthManager>>,
otel_event_manager: OtelEventManager,
provider: ModelProviderInfo,
effort: Option<ReasoningEffortConfig>,
summary: ReasoningSummaryConfig,
@@ -99,7 +94,6 @@ impl ModelClient {
Self {
config,
auth_manager,
otel_event_manager,
client,
provider,
conversation_id,
@@ -133,7 +127,6 @@ impl ModelClient {
&self.config.model_family,
&self.client,
&self.provider,
&self.otel_event_manager,
)
.await?;
@@ -170,12 +163,7 @@ impl ModelClient {
if let Some(path) = &*CODEX_RS_SSE_FIXTURE {
// short circuit for tests
warn!(path, "Streaming from fixture");
return stream_from_fixture(
path,
self.provider.clone(),
self.otel_event_manager.clone(),
)
.await;
return stream_from_fixture(path, self.provider.clone()).await;
}
let auth_manager = self.auth_manager.clone();
@@ -245,7 +233,7 @@ impl ModelClient {
let max_attempts = self.provider.request_max_retries();
for attempt in 0..=max_attempts {
match self
.attempt_stream_responses(attempt, &payload_json, &auth_manager)
.attempt_stream_responses(&payload_json, &auth_manager)
.await
{
Ok(stream) => {
@@ -270,7 +258,6 @@ impl ModelClient {
/// Single attempt to start a streaming Responses API call.
async fn attempt_stream_responses(
&self,
attempt: u64,
payload_json: &Value,
auth_manager: &Option<Arc<AuthManager>>,
) -> std::result::Result<ResponseStream, StreamAttemptError> {
@@ -304,22 +291,15 @@ impl ModelClient {
req_builder = req_builder.header("chatgpt-account-id", account_id);
}
let res = self
.otel_event_manager
.log_request(attempt, || req_builder.send())
.await;
let mut request_id = None;
let res = req_builder.send().await;
if let Ok(resp) = &res {
request_id = resp
.headers()
.get("cf-ray")
.map(|v| v.to_str().unwrap_or_default().to_string());
trace!(
"Response status: {}, cf-ray: {:?}",
"Response status: {}, cf-ray: {}",
resp.status(),
request_id
resp.headers()
.get("cf-ray")
.map(|v| v.to_str().unwrap_or_default())
.unwrap_or_default()
);
}
@@ -342,7 +322,6 @@ impl ModelClient {
stream,
tx_event,
self.provider.stream_idle_timeout(),
self.otel_event_manager.clone(),
));
Ok(ResponseStream { rx_event })
@@ -379,11 +358,7 @@ impl ModelClient {
// Surface the error body to callers. Use `unwrap_or_default` per Clippy.
let body = res.text().await.unwrap_or_default();
return Err(StreamAttemptError::Fatal(CodexErr::UnexpectedStatus(
UnexpectedResponseError {
status,
body,
request_id: None,
},
status, body,
)));
}
@@ -414,7 +389,6 @@ impl ModelClient {
Err(StreamAttemptError::RetryableHttpError {
status,
retry_after,
request_id,
})
}
Err(e) => Err(StreamAttemptError::RetryableTransportError(e.into())),
@@ -425,10 +399,6 @@ impl ModelClient {
self.provider.clone()
}
pub fn get_otel_event_manager(&self) -> OtelEventManager {
self.otel_event_manager.clone()
}
/// Returns the currently configured model slug.
pub fn get_model(&self) -> String {
self.config.model.clone()
@@ -458,7 +428,6 @@ enum StreamAttemptError {
RetryableHttpError {
status: StatusCode,
retry_after: Option<Duration>,
request_id: Option<String>,
},
RetryableTransportError(CodexErr),
Fatal(CodexErr),
@@ -483,13 +452,11 @@ impl StreamAttemptError {
fn into_error(self) -> CodexErr {
match self {
Self::RetryableHttpError {
status, request_id, ..
} => {
Self::RetryableHttpError { status, .. } => {
if status == StatusCode::INTERNAL_SERVER_ERROR {
CodexErr::InternalServerError
} else {
CodexErr::RetryLimit(RetryLimitReachedError { status, request_id })
CodexErr::RetryLimit(status)
}
}
Self::RetryableTransportError(error) => error,
@@ -638,7 +605,6 @@ async fn process_sse<S>(
stream: S,
tx_event: mpsc::Sender<Result<ResponseEvent>>,
idle_timeout: Duration,
otel_event_manager: OtelEventManager,
) where
S: Stream<Item = Result<Bytes>> + Unpin,
{
@@ -650,10 +616,7 @@ async fn process_sse<S>(
let mut response_error: Option<CodexErr> = None;
loop {
let sse = match otel_event_manager
.log_sse_event(|| timeout(idle_timeout, stream.next()))
.await
{
let sse = match timeout(idle_timeout, stream.next()).await {
Ok(Some(Ok(sse))) => sse,
Ok(Some(Err(e))) => {
debug!("SSE Error: {e:#}");
@@ -667,21 +630,6 @@ async fn process_sse<S>(
id: response_id,
usage,
}) => {
if let Some(token_usage) = &usage {
otel_event_manager.sse_event_completed(
token_usage.input_tokens,
token_usage.output_tokens,
token_usage
.input_tokens_details
.as_ref()
.map(|d| d.cached_tokens),
token_usage
.output_tokens_details
.as_ref()
.map(|d| d.reasoning_tokens),
token_usage.total_tokens,
);
}
let event = ResponseEvent::Completed {
response_id,
token_usage: usage.map(Into::into),
@@ -689,13 +637,12 @@ async fn process_sse<S>(
let _ = tx_event.send(Ok(event)).await;
}
None => {
let error = response_error.unwrap_or(CodexErr::Stream(
"stream closed before response.completed".into(),
None,
));
otel_event_manager.see_event_completed_failed(&error);
let _ = tx_event.send(Err(error)).await;
let _ = tx_event
.send(Err(response_error.unwrap_or(CodexErr::Stream(
"stream closed before response.completed".into(),
None,
))))
.await;
}
}
return;
@@ -799,9 +746,7 @@ async fn process_sse<S>(
response_error = Some(CodexErr::Stream(message, delay));
}
Err(e) => {
let error = format!("failed to parse ErrorResponse: {e}");
debug!(error);
response_error = Some(CodexErr::Stream(error, None))
debug!("failed to parse ErrorResponse: {e}");
}
}
}
@@ -815,9 +760,7 @@ async fn process_sse<S>(
response_completed = Some(r);
}
Err(e) => {
let error = format!("failed to parse ResponseCompleted: {e}");
debug!(error);
response_error = Some(CodexErr::Stream(error, None));
debug!("failed to parse ResponseCompleted: {e}");
continue;
}
};
@@ -864,7 +807,6 @@ async fn process_sse<S>(
async fn stream_from_fixture(
path: impl AsRef<Path>,
provider: ModelProviderInfo,
otel_event_manager: OtelEventManager,
) -> Result<ResponseStream> {
let (tx_event, rx_event) = mpsc::channel::<Result<ResponseEvent>>(1600);
let f = std::fs::File::open(path.as_ref())?;
@@ -883,7 +825,6 @@ async fn stream_from_fixture(
stream,
tx_event,
provider.stream_idle_timeout(),
otel_event_manager,
));
Ok(ResponseStream { rx_event })
}
@@ -939,7 +880,6 @@ mod tests {
async fn collect_events(
chunks: &[&[u8]],
provider: ModelProviderInfo,
otel_event_manager: OtelEventManager,
) -> Vec<Result<ResponseEvent>> {
let mut builder = IoBuilder::new();
for chunk in chunks {
@@ -949,12 +889,7 @@ mod tests {
let reader = builder.build();
let stream = ReaderStream::new(reader).map_err(CodexErr::Io);
let (tx, mut rx) = mpsc::channel::<Result<ResponseEvent>>(16);
tokio::spawn(process_sse(
stream,
tx,
provider.stream_idle_timeout(),
otel_event_manager,
));
tokio::spawn(process_sse(stream, tx, provider.stream_idle_timeout()));
let mut events = Vec::new();
while let Some(ev) = rx.recv().await {
@@ -968,7 +903,6 @@ mod tests {
async fn run_sse(
events: Vec<serde_json::Value>,
provider: ModelProviderInfo,
otel_event_manager: OtelEventManager,
) -> Vec<ResponseEvent> {
let mut body = String::new();
for e in events {
@@ -985,12 +919,7 @@ mod tests {
let (tx, mut rx) = mpsc::channel::<Result<ResponseEvent>>(8);
let stream = ReaderStream::new(std::io::Cursor::new(body)).map_err(CodexErr::Io);
tokio::spawn(process_sse(
stream,
tx,
provider.stream_idle_timeout(),
otel_event_manager,
));
tokio::spawn(process_sse(stream, tx, provider.stream_idle_timeout()));
let mut out = Vec::new();
while let Some(ev) = rx.recv().await {
@@ -999,18 +928,6 @@ mod tests {
out
}
fn otel_event_manager() -> OtelEventManager {
OtelEventManager::new(
ConversationId::new(),
"test",
"test",
None,
Some(AuthMode::ChatGPT),
false,
"test".to_string(),
)
}
// ────────────────────────────
// Tests from `implement-test-for-responses-api-sse-parser`
// ────────────────────────────
@@ -1062,12 +979,9 @@ mod tests {
requires_openai_auth: false,
};
let otel_event_manager = otel_event_manager();
let events = collect_events(
&[sse1.as_bytes(), sse2.as_bytes(), sse3.as_bytes()],
provider,
otel_event_manager,
)
.await;
@@ -1125,9 +1039,7 @@ mod tests {
requires_openai_auth: false,
};
let otel_event_manager = otel_event_manager();
let events = collect_events(&[sse1.as_bytes()], provider, otel_event_manager).await;
let events = collect_events(&[sse1.as_bytes()], provider).await;
assert_eq!(events.len(), 2);
@@ -1161,9 +1073,7 @@ mod tests {
requires_openai_auth: false,
};
let otel_event_manager = otel_event_manager();
let events = collect_events(&[sse1.as_bytes()], provider, otel_event_manager).await;
let events = collect_events(&[sse1.as_bytes()], provider).await;
assert_eq!(events.len(), 1);
@@ -1268,9 +1178,7 @@ mod tests {
requires_openai_auth: false,
};
let otel_event_manager = otel_event_manager();
let out = run_sse(evs, provider, otel_event_manager).await;
let out = run_sse(evs, provider).await;
assert_eq!(out.len(), case.expected_len, "case {}", case.name);
assert!(
(case.expect_first)(&out[0]),

View File

@@ -1,23 +1,23 @@
use std::borrow::Cow;
use std::collections::HashMap;
use std::fmt::Debug;
use std::path::Path;
use std::path::PathBuf;
use std::sync::Arc;
use std::sync::atomic::AtomicU64;
use std::time::Duration;
use crate::AuthManager;
use crate::client_common::REVIEW_PROMPT;
use crate::event_mapping::map_response_item_to_event_messages;
use crate::function_tool::FunctionCallError;
use crate::review_format::format_review_findings_block;
use crate::terminal;
use crate::user_notification::UserNotifier;
use async_channel::Receiver;
use async_channel::Sender;
use codex_apply_patch::ApplyPatchAction;
use codex_apply_patch::MaybeApplyPatchVerified;
use codex_apply_patch::maybe_parse_apply_patch_verified;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::ConversationId;
use codex_protocol::protocol::ConversationPathResponseEvent;
use codex_protocol::protocol::ExitedReviewModeEvent;
use codex_protocol::protocol::ReviewRequest;
@@ -42,6 +42,7 @@ use tracing::warn;
use crate::ModelProviderInfo;
use crate::apply_patch;
use crate::apply_patch::ApplyPatchExec;
use crate::apply_patch::CODEX_APPLY_PATCH_ARG1;
use crate::apply_patch::InternalApplyPatchInvocation;
use crate::apply_patch::convert_apply_patch_to_protocol;
use crate::client::ModelClient;
@@ -54,21 +55,19 @@ use crate::environment_context::EnvironmentContext;
use crate::error::CodexErr;
use crate::error::Result as CodexResult;
use crate::error::SandboxErr;
use crate::error::get_error_message_ui;
use crate::exec::ExecParams;
use crate::exec::ExecToolCallOutput;
use crate::exec::SandboxType;
use crate::exec::StdoutStream;
#[cfg(test)]
use crate::exec::StreamOutput;
use crate::exec::process_exec_tool_call;
use crate::exec_command::EXEC_COMMAND_TOOL_NAME;
use crate::exec_command::ExecCommandParams;
use crate::exec_command::ExecSessionManager;
use crate::exec_command::WRITE_STDIN_TOOL_NAME;
use crate::exec_command::WriteStdinParams;
use crate::exec_env::create_env;
use crate::executor::ExecutionMode;
use crate::executor::Executor;
use crate::executor::ExecutorConfig;
use crate::executor::normalize_exec_result;
use crate::mcp_connection_manager::McpConnectionManager;
use crate::mcp_tool_call::handle_mcp_tool_call;
use crate::model_family::find_family_for_model;
@@ -112,6 +111,9 @@ use crate::protocol::TurnDiffEvent;
use crate::protocol::WebSearchBeginEvent;
use crate::rollout::RolloutRecorder;
use crate::rollout::RolloutRecorderParams;
use crate::safety::SafetyCheck;
use crate::safety::assess_command_safety;
use crate::safety::assess_safety_for_untrusted_command;
use crate::shell;
use crate::state::ActiveTurn;
use crate::state::SessionServices;
@@ -123,7 +125,6 @@ use crate::unified_exec::UnifiedExecSessionManager;
use crate::user_instructions::UserInstructions;
use crate::user_notification::UserNotification;
use crate::util::backoff;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_protocol::config_types::ReasoningEffort as ReasoningEffortConfig;
use codex_protocol::config_types::ReasoningSummary as ReasoningSummaryConfig;
use codex_protocol::custom_prompts::CustomPrompt;
@@ -421,35 +422,11 @@ impl Session {
}
}
let otel_event_manager = OtelEventManager::new(
conversation_id,
config.model.as_str(),
config.model_family.slug.as_str(),
auth_manager.auth().and_then(|a| a.get_account_id()),
auth_manager.auth().map(|a| a.mode),
config.otel.log_user_prompt,
terminal::user_agent(),
);
otel_event_manager.conversation_starts(
config.model_provider.name.as_str(),
config.model_reasoning_effort,
config.model_reasoning_summary,
config.model_context_window,
config.model_max_output_tokens,
config.model_auto_compact_token_limit,
config.approval_policy,
config.sandbox_policy.clone(),
config.mcp_servers.keys().map(String::as_str).collect(),
config.active_profile.clone(),
);
// Now that the conversation id is final (may have been updated by resume),
// construct the model client.
let client = ModelClient::new(
config.clone(),
Some(auth_manager.clone()),
otel_event_manager,
provider.clone(),
model_reasoning_effort,
model_reasoning_summary,
@@ -481,13 +458,9 @@ impl Session {
unified_exec_manager: UnifiedExecSessionManager::default(),
notifier: notify,
rollout: Mutex::new(Some(rollout_recorder)),
codex_linux_sandbox_exe: config.codex_linux_sandbox_exe.clone(),
user_shell: default_shell,
show_raw_agent_reasoning: config.show_raw_agent_reasoning,
executor: Executor::new(ExecutorConfig::new(
turn_context.sandbox_policy.clone(),
turn_context.cwd.clone(),
config.codex_linux_sandbox_exe.clone(),
)),
};
let sess = Arc::new(Session {
@@ -572,11 +545,6 @@ impl Session {
}
}
/// Emit an exec approval request event and await the user's decision.
///
/// The request is keyed by `sub_id`/`call_id` so matching responses are delivered
/// to the correct in-flight turn. If the task is aborted, this returns the
/// default `ReviewDecision` (`Denied`).
pub async fn request_command_approval(
&self,
sub_id: String,
@@ -674,6 +642,11 @@ impl Session {
}
}
pub async fn add_approved_command(&self, cmd: Vec<String>) {
let mut state = self.state.lock().await;
state.add_approved_command(cmd);
}
/// Records input items: always append to conversation history and
/// persist these response items to rollout.
async fn record_conversation_items(&self, items: &[ResponseItem]) {
@@ -831,7 +804,6 @@ impl Session {
command_for_display,
cwd,
apply_patch,
..
} = exec_command_context;
let msg = match apply_patch {
Some(ApplyPatchCommandContext {
@@ -928,29 +900,45 @@ impl Session {
/// command even on error.
///
/// Returns the output of the exec tool call.
async fn run_exec_with_events(
async fn run_exec_with_events<'a>(
&self,
turn_diff_tracker: &mut TurnDiffTracker,
prepared: PreparedExec,
approval_policy: AskForApproval,
) -> Result<ExecToolCallOutput, ExecError> {
let PreparedExec { context, request } = prepared;
let is_apply_patch = context.apply_patch.is_some();
let sub_id = context.sub_id.clone();
let call_id = context.call_id.clone();
begin_ctx: ExecCommandContext,
exec_args: ExecInvokeArgs<'a>,
) -> crate::error::Result<ExecToolCallOutput> {
let is_apply_patch = begin_ctx.apply_patch.is_some();
let sub_id = begin_ctx.sub_id.clone();
let call_id = begin_ctx.call_id.clone();
self.on_exec_command_begin(turn_diff_tracker, context.clone())
self.on_exec_command_begin(turn_diff_tracker, begin_ctx.clone())
.await;
let result = self
.services
.executor
.run(request, self, approval_policy, &context)
.await;
let normalized = normalize_exec_result(&result);
let borrowed = normalized.event_output();
let result = process_exec_tool_call(
exec_args.params,
exec_args.sandbox_type,
exec_args.sandbox_policy,
exec_args.sandbox_cwd,
exec_args.codex_linux_sandbox_exe,
exec_args.stdout_stream,
)
.await;
let output_stderr;
let borrowed: &ExecToolCallOutput = match &result {
Ok(output) => output,
Err(CodexErr::Sandbox(SandboxErr::Timeout { output })) => output,
Err(e) => {
output_stderr = ExecToolCallOutput {
exit_code: -1,
stdout: StreamOutput::new(String::new()),
stderr: StreamOutput::new(get_error_message_ui(e)),
aggregated_output: StreamOutput::new(get_error_message_ui(e)),
duration: Duration::default(),
timed_out: false,
};
&output_stderr
}
};
self.on_exec_command_end(
turn_diff_tracker,
&sub_id,
@@ -960,15 +948,13 @@ impl Session {
)
.await;
drop(normalized);
result
}
/// Helper that emits a BackgroundEvent with the given message. This keeps
/// the callsites terse so adding more diagnostics does not clutter the
/// core agent logic.
pub(crate) async fn notify_background_event(&self, sub_id: &str, message: impl Into<String>) {
async fn notify_background_event(&self, sub_id: &str, message: impl Into<String>) {
let event = Event {
id: sub_id.to_string(),
msg: EventMsg::BackgroundEvent(BackgroundEventEvent {
@@ -1056,7 +1042,7 @@ impl Session {
&self.services.notifier
}
pub(crate) fn user_shell(&self) -> &shell::Shell {
fn user_shell(&self) -> &shell::Shell {
&self.services.user_shell
}
@@ -1078,8 +1064,6 @@ pub(crate) struct ExecCommandContext {
pub(crate) command_for_display: Vec<String>,
pub(crate) cwd: PathBuf,
pub(crate) apply_patch: Option<ApplyPatchCommandContext>,
pub(crate) tool_name: String,
pub(crate) otel_event_manager: OtelEventManager,
}
#[derive(Clone, Debug)]
@@ -1138,15 +1122,9 @@ async fn submission_loop(
updated_config.model_context_window = Some(model_info.context_window);
}
let otel_event_manager = prev.client.get_otel_event_manager().with_model(
updated_config.model.as_str(),
updated_config.model_family.slug.as_str(),
);
let client = ModelClient::new(
Arc::new(updated_config),
auth_manager,
otel_event_manager,
provider,
effective_effort,
effective_summary,
@@ -1198,10 +1176,6 @@ async fn submission_loop(
}
}
Op::UserInput { items } => {
turn_context
.client
.get_otel_event_manager()
.user_prompt(&items);
// attempt to inject input into current task
if let Err(items) = sess.inject_input(items).await {
// no current task, spawn a new one
@@ -1219,10 +1193,6 @@ async fn submission_loop(
summary,
final_output_json_schema,
} => {
turn_context
.client
.get_otel_event_manager()
.user_prompt(&items);
// attempt to inject input into current task
if let Err(items) = sess.inject_input(items).await {
// Derive a fresh TurnContext for this turn using the provided overrides.
@@ -1241,18 +1211,11 @@ async fn submission_loop(
per_turn_config.model_context_window = Some(model_info.context_window);
}
let otel_event_manager =
turn_context.client.get_otel_event_manager().with_model(
per_turn_config.model.as_str(),
per_turn_config.model_family.slug.as_str(),
);
// Build a new client with perturn reasoning settings.
// Reuse the same provider and session id; auth defaults to env/API key.
let client = ModelClient::new(
Arc::new(per_turn_config),
auth_manager,
otel_event_manager,
provider,
effort,
summary,
@@ -1286,19 +1249,8 @@ async fn submission_loop(
let previous_env_context = EnvironmentContext::from(turn_context.as_ref());
let new_env_context = EnvironmentContext::from(&fresh_turn_context);
if !new_env_context.equals_except_shell(&previous_env_context) {
let env_response_item = ResponseItem::from(new_env_context);
sess.record_conversation_items(std::slice::from_ref(&env_response_item))
sess.record_conversation_items(&[ResponseItem::from(new_env_context)])
.await;
for msg in map_response_item_to_event_messages(
&env_response_item,
sess.show_raw_agent_reasoning(),
) {
let event = Event {
id: sub.id.clone(),
msg,
};
sess.send_event(event).await;
}
}
// Install the new persistent context for subsequent tasks/turns.
@@ -1520,19 +1472,10 @@ async fn spawn_review_thread(
per_turn_config.model_context_window = Some(model_info.context_window);
}
let otel_event_manager = parent_turn_context
.client
.get_otel_event_manager()
.with_model(
per_turn_config.model.as_str(),
per_turn_config.model_family.slug.as_str(),
);
let per_turn_config = Arc::new(per_turn_config);
let client = ModelClient::new(
per_turn_config.clone(),
auth_manager,
otel_event_manager,
provider,
per_turn_config.model_reasoning_effort,
per_turn_config.model_reasoning_summary,
@@ -2197,21 +2140,16 @@ async fn handle_response_item(
.await;
Some(resp)
} else {
let result = turn_context
.client
.get_otel_event_manager()
.log_tool_result(name.as_str(), call_id.as_str(), arguments.as_str(), || {
handle_function_call(
sess,
turn_context,
turn_diff_tracker,
sub_id.to_string(),
name.to_owned(),
arguments.to_owned(),
call_id.clone(),
)
})
.await;
let result = handle_function_call(
sess,
turn_context,
turn_diff_tracker,
sub_id.to_string(),
name,
arguments,
call_id.clone(),
)
.await;
let output = match result {
Ok(content) => FunctionCallOutputPayload {
@@ -2232,7 +2170,6 @@ async fn handle_response_item(
status: _,
action,
} => {
let name = "local_shell";
let LocalShellAction::Exec(action) = action;
tracing::info!("LocalShellCall: {action:?}");
let params = ShellToolCallParams {
@@ -2246,18 +2183,11 @@ async fn handle_response_item(
(Some(call_id), _) => call_id,
(None, Some(id)) => id,
(None, None) => {
let error_message = "LocalShellCall without call_id or id";
turn_context
.client
.get_otel_event_manager()
.log_tool_failed(name, error_message);
error!(error_message);
error!("LocalShellCall without call_id or id");
return Ok(Some(ResponseInputItem::FunctionCallOutput {
call_id: "".to_string(),
output: FunctionCallOutputPayload {
content: error_message.to_string(),
content: "LocalShellCall without call_id or id".to_string(),
success: None,
},
}));
@@ -2266,26 +2196,15 @@ async fn handle_response_item(
let exec_params = to_exec_params(params, turn_context);
{
let result = turn_context
.client
.get_otel_event_manager()
.log_tool_result(
name,
effective_call_id.as_str(),
exec_params.command.join(" ").as_str(),
|| {
handle_container_exec_with_params(
name,
exec_params,
sess,
turn_context,
turn_diff_tracker,
sub_id.to_string(),
effective_call_id.clone(),
)
},
)
.await;
let result = handle_container_exec_with_params(
exec_params,
sess,
turn_context,
turn_diff_tracker,
sub_id.to_string(),
effective_call_id.clone(),
)
.await;
let output = match result {
Ok(content) => FunctionCallOutputPayload {
@@ -2310,21 +2229,16 @@ async fn handle_response_item(
input,
status: _,
} => {
let result = turn_context
.client
.get_otel_event_manager()
.log_tool_result(name.as_str(), call_id.as_str(), input.as_str(), || {
handle_custom_tool_call(
sess,
turn_context,
turn_diff_tracker,
sub_id.to_string(),
name.to_owned(),
input.to_owned(),
call_id.clone(),
)
})
.await;
let result = handle_custom_tool_call(
sess,
turn_context,
turn_diff_tracker,
sub_id.to_string(),
name,
input,
call_id.clone(),
)
.await;
let output = match result {
Ok(content) => content,
@@ -2430,7 +2344,6 @@ async fn handle_function_call(
"container.exec" | "shell" => {
let params = parse_container_exec_arguments(arguments, turn_context, &call_id)?;
handle_container_exec_with_params(
name.as_str(),
params,
sess,
turn_context,
@@ -2494,7 +2407,6 @@ async fn handle_function_call(
justification: None,
};
handle_container_exec_with_params(
name.as_str(),
exec_params,
sess,
turn_context,
@@ -2567,7 +2479,6 @@ async fn handle_custom_tool_call(
};
handle_container_exec_with_params(
name.as_str(),
exec_params,
sess,
turn_context,
@@ -2609,8 +2520,34 @@ fn parse_container_exec_arguments(
})
}
pub struct ExecInvokeArgs<'a> {
pub params: ExecParams,
pub sandbox_type: SandboxType,
pub sandbox_policy: &'a SandboxPolicy,
pub sandbox_cwd: &'a Path,
pub codex_linux_sandbox_exe: &'a Option<PathBuf>,
pub stdout_stream: Option<StdoutStream>,
}
fn maybe_translate_shell_command(
params: ExecParams,
sess: &Session,
turn_context: &TurnContext,
) -> ExecParams {
let should_translate = matches!(sess.user_shell(), crate::shell::Shell::PowerShell(_))
|| turn_context.shell_environment_policy.use_profile;
if should_translate
&& let Some(command) = sess
.user_shell()
.format_default_shell_invocation(params.command.clone())
{
return ExecParams { command, ..params };
}
params
}
async fn handle_container_exec_with_params(
tool_name: &str,
params: ExecParams,
sess: &Session,
turn_context: &TurnContext,
@@ -2618,8 +2555,6 @@ async fn handle_container_exec_with_params(
sub_id: String,
call_id: String,
) -> Result<String, FunctionCallError> {
let otel_event_manager = turn_context.client.get_otel_event_manager();
if params.with_escalated_permissions.unwrap_or(false)
&& !matches!(turn_context.approval_policy, AskForApproval::OnRequest)
{
@@ -2654,10 +2589,99 @@ async fn handle_container_exec_with_params(
MaybeApplyPatchVerified::NotApplyPatch => None,
};
let command_for_display = if let Some(exec) = apply_patch_exec.as_ref() {
vec!["apply_patch".to_string(), exec.action.patch.clone()]
} else {
params.command.clone()
let (params, safety, command_for_display) = match &apply_patch_exec {
Some(ApplyPatchExec {
action: ApplyPatchAction { patch, cwd, .. },
user_explicitly_approved_this_action,
}) => {
let path_to_codex = std::env::current_exe()
.ok()
.map(|p| p.to_string_lossy().to_string());
let Some(path_to_codex) = path_to_codex else {
return Err(FunctionCallError::RespondToModel(
"failed to determine path to codex executable".to_string(),
));
};
let params = ExecParams {
command: vec![
path_to_codex,
CODEX_APPLY_PATCH_ARG1.to_string(),
patch.clone(),
],
cwd: cwd.clone(),
timeout_ms: params.timeout_ms,
env: HashMap::new(),
with_escalated_permissions: params.with_escalated_permissions,
justification: params.justification.clone(),
};
let safety = if *user_explicitly_approved_this_action {
SafetyCheck::AutoApprove {
sandbox_type: SandboxType::None,
}
} else {
assess_safety_for_untrusted_command(
turn_context.approval_policy,
&turn_context.sandbox_policy,
params.with_escalated_permissions.unwrap_or(false),
)
};
(
params,
safety,
vec!["apply_patch".to_string(), patch.clone()],
)
}
None => {
let safety = {
let state = sess.state.lock().await;
assess_command_safety(
&params.command,
turn_context.approval_policy,
&turn_context.sandbox_policy,
state.approved_commands_ref(),
params.with_escalated_permissions.unwrap_or(false),
)
};
let command_for_display = params.command.clone();
(params, safety, command_for_display)
}
};
let sandbox_type = match safety {
SafetyCheck::AutoApprove { sandbox_type } => sandbox_type,
SafetyCheck::AskUser => {
let decision = sess
.request_command_approval(
sub_id.clone(),
call_id.clone(),
params.command.clone(),
params.cwd.clone(),
params.justification.clone(),
)
.await;
match decision {
ReviewDecision::Approved => (),
ReviewDecision::ApprovedForSession => {
sess.add_approved_command(params.command.clone()).await;
}
ReviewDecision::Denied | ReviewDecision::Abort => {
return Err(FunctionCallError::RespondToModel(
"exec command rejected by user".to_string(),
));
}
}
// No sandboxing is applied because the user has given
// explicit approval. Often, we end up in this case because
// the command cannot be run in a sandbox, such as
// installing a new dependency that requires network access.
SandboxType::None
}
SafetyCheck::Reject { reason } => {
return Err(FunctionCallError::RespondToModel(format!(
"exec command rejected: {reason:?}"
)));
}
};
let exec_command_context = ExecCommandContext {
@@ -2665,47 +2689,38 @@ async fn handle_container_exec_with_params(
call_id: call_id.clone(),
command_for_display: command_for_display.clone(),
cwd: params.cwd.clone(),
apply_patch: apply_patch_exec.as_ref().map(
apply_patch: apply_patch_exec.map(
|ApplyPatchExec {
action,
user_explicitly_approved_this_action,
}| ApplyPatchCommandContext {
user_explicitly_approved_this_action: *user_explicitly_approved_this_action,
changes: convert_apply_patch_to_protocol(action),
user_explicitly_approved_this_action,
changes: convert_apply_patch_to_protocol(&action),
},
),
tool_name: tool_name.to_string(),
otel_event_manager,
};
let mode = match apply_patch_exec {
Some(exec) => ExecutionMode::ApplyPatch(exec),
None => ExecutionMode::Shell,
};
sess.services.executor.update_environment(
turn_context.sandbox_policy.clone(),
turn_context.cwd.clone(),
);
let prepared_exec = PreparedExec::new(
exec_command_context,
params,
command_for_display,
mode,
Some(StdoutStream {
sub_id: sub_id.clone(),
call_id: call_id.clone(),
tx_event: sess.tx_event.clone(),
}),
turn_context.shell_environment_policy.use_profile,
);
let params = maybe_translate_shell_command(params, sess, turn_context);
let output_result = sess
.run_exec_with_events(
turn_diff_tracker,
prepared_exec,
turn_context.approval_policy,
exec_command_context.clone(),
ExecInvokeArgs {
params: params.clone(),
sandbox_type,
sandbox_policy: &turn_context.sandbox_policy,
sandbox_cwd: &turn_context.cwd,
codex_linux_sandbox_exe: &sess.services.codex_linux_sandbox_exe,
stdout_stream: if exec_command_context.apply_patch.is_some() {
None
} else {
Some(StdoutStream {
sub_id: sub_id.clone(),
call_id: call_id.clone(),
tx_event: sess.tx_event.clone(),
})
},
},
)
.await;
@@ -2719,16 +2734,135 @@ async fn handle_container_exec_with_params(
Err(FunctionCallError::RespondToModel(content))
}
}
Err(ExecError::Function(err)) => Err(err),
Err(ExecError::Codex(CodexErr::Sandbox(SandboxErr::Timeout { output }))) => Err(
FunctionCallError::RespondToModel(format_exec_output(&output)),
),
Err(ExecError::Codex(err)) => Err(FunctionCallError::RespondToModel(format!(
"execution error: {err:?}"
Err(CodexErr::Sandbox(error)) => {
handle_sandbox_error(
turn_diff_tracker,
params,
exec_command_context,
error,
sandbox_type,
sess,
turn_context,
)
.await
}
Err(e) => Err(FunctionCallError::RespondToModel(format!(
"execution error: {e:?}"
))),
}
}
async fn handle_sandbox_error(
turn_diff_tracker: &mut TurnDiffTracker,
params: ExecParams,
exec_command_context: ExecCommandContext,
error: SandboxErr,
sandbox_type: SandboxType,
sess: &Session,
turn_context: &TurnContext,
) -> Result<String, FunctionCallError> {
let call_id = exec_command_context.call_id.clone();
let sub_id = exec_command_context.sub_id.clone();
let cwd = exec_command_context.cwd.clone();
if let SandboxErr::Timeout { output } = &error {
let content = format_exec_output(output);
return Err(FunctionCallError::RespondToModel(content));
}
// Early out if either the user never wants to be asked for approval, or
// we're letting the model manage escalation requests. Otherwise, continue
match turn_context.approval_policy {
AskForApproval::Never | AskForApproval::OnRequest => {
return Err(FunctionCallError::RespondToModel(format!(
"failed in sandbox {sandbox_type:?} with execution error: {error:?}"
)));
}
AskForApproval::UnlessTrusted | AskForApproval::OnFailure => (),
}
// Note that when `error` is `SandboxErr::Denied`, it could be a false
// positive. That is, it may have exited with a non-zero exit code, not
// because the sandbox denied it, but because that is its expected behavior,
// i.e., a grep command that did not match anything. Ideally we would
// include additional metadata on the command to indicate whether non-zero
// exit codes merit a retry.
// For now, we categorically ask the user to retry without sandbox and
// emit the raw error as a background event.
sess.notify_background_event(&sub_id, format!("Execution failed: {error}"))
.await;
let decision = sess
.request_command_approval(
sub_id.clone(),
call_id.clone(),
params.command.clone(),
cwd.clone(),
Some("command failed; retry without sandbox?".to_string()),
)
.await;
match decision {
ReviewDecision::Approved | ReviewDecision::ApprovedForSession => {
// Persist this command as preapproved for the
// remainder of the session so future
// executions skip the sandbox directly.
// TODO(ragona): Isn't this a bug? It always saves the command in an | fork?
sess.add_approved_command(params.command.clone()).await;
// Inform UI we are retrying without sandbox.
sess.notify_background_event(&sub_id, "retrying command without sandbox")
.await;
// This is an escalated retry; the policy will not be
// examined and the sandbox has been set to `None`.
let retry_output_result = sess
.run_exec_with_events(
turn_diff_tracker,
exec_command_context.clone(),
ExecInvokeArgs {
params,
sandbox_type: SandboxType::None,
sandbox_policy: &turn_context.sandbox_policy,
sandbox_cwd: &turn_context.cwd,
codex_linux_sandbox_exe: &sess.services.codex_linux_sandbox_exe,
stdout_stream: if exec_command_context.apply_patch.is_some() {
None
} else {
Some(StdoutStream {
sub_id: sub_id.clone(),
call_id: call_id.clone(),
tx_event: sess.tx_event.clone(),
})
},
},
)
.await;
match retry_output_result {
Ok(retry_output) => {
let ExecToolCallOutput { exit_code, .. } = &retry_output;
let content = format_exec_output(&retry_output);
if *exit_code == 0 {
Ok(content)
} else {
Err(FunctionCallError::RespondToModel(content))
}
}
Err(e) => Err(FunctionCallError::RespondToModel(format!(
"retry failed: {e}"
))),
}
}
ReviewDecision::Denied | ReviewDecision::Abort => {
// Fall through to original failure handling.
Err(FunctionCallError::RespondToModel(
"exec command rejected by user".to_string(),
))
}
}
}
fn format_exec_output_str(exec_output: &ExecToolCallOutput) -> String {
let ExecToolCallOutput {
aggregated_output, ..
@@ -2987,8 +3121,6 @@ pub(crate) async fn exit_review_mode(
.await;
}
use crate::executor::errors::ExecError;
use crate::executor::linkers::PreparedExec;
#[cfg(test)]
pub(crate) use tests::make_session_and_context;
@@ -2997,17 +3129,13 @@ mod tests {
use super::*;
use crate::config::ConfigOverrides;
use crate::config::ConfigToml;
use crate::protocol::CompactedItem;
use crate::protocol::InitialHistory;
use crate::protocol::ResumedHistory;
use crate::state::TaskKind;
use crate::tasks::SessionTask;
use crate::tasks::SessionTaskContext;
use codex_app_server_protocol::AuthMode;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ResponseItem;
use mcp_types::ContentBlock;
use mcp_types::TextContent;
use pretty_assertions::assert_eq;
@@ -3242,18 +3370,6 @@ mod tests {
})
}
fn otel_event_manager(conversation_id: ConversationId, config: &Config) -> OtelEventManager {
OtelEventManager::new(
conversation_id,
config.model.as_str(),
config.model_family.slug.as_str(),
None,
Some(AuthMode::ChatGPT),
false,
"test".to_string(),
)
}
pub(crate) fn make_session_and_context() -> (Session, TurnContext) {
let (tx_event, _rx_event) = async_channel::unbounded();
let codex_home = tempfile::tempdir().expect("create temp dir");
@@ -3265,11 +3381,9 @@ mod tests {
.expect("load default test config");
let config = Arc::new(config);
let conversation_id = ConversationId::default();
let otel_event_manager = otel_event_manager(conversation_id, config.as_ref());
let client = ModelClient::new(
config.clone(),
None,
otel_event_manager,
config.model_provider.clone(),
config.model_reasoning_effort,
config.model_reasoning_summary,
@@ -3302,13 +3416,9 @@ mod tests {
unified_exec_manager: UnifiedExecSessionManager::default(),
notifier: UserNotifier::default(),
rollout: Mutex::new(None),
codex_linux_sandbox_exe: None,
user_shell: shell::Shell::Unknown,
show_raw_agent_reasoning: config.show_raw_agent_reasoning,
executor: Executor::new(ExecutorConfig::new(
turn_context.sandbox_policy.clone(),
turn_context.cwd.clone(),
None,
)),
};
let session = Session {
conversation_id,
@@ -3338,11 +3448,9 @@ mod tests {
.expect("load default test config");
let config = Arc::new(config);
let conversation_id = ConversationId::default();
let otel_event_manager = otel_event_manager(conversation_id, config.as_ref());
let client = ModelClient::new(
config.clone(),
None,
otel_event_manager,
config.model_provider.clone(),
config.model_reasoning_effort,
config.model_reasoning_summary,
@@ -3375,13 +3483,9 @@ mod tests {
unified_exec_manager: UnifiedExecSessionManager::default(),
notifier: UserNotifier::default(),
rollout: Mutex::new(None),
codex_linux_sandbox_exe: None,
user_shell: shell::Shell::Unknown,
show_raw_agent_reasoning: config.show_raw_agent_reasoning,
executor: Executor::new(ExecutorConfig::new(
config.sandbox_policy.clone(),
config.cwd.clone(),
None,
)),
};
let session = Arc::new(Session {
conversation_id,
@@ -3637,12 +3741,10 @@ mod tests {
let mut turn_diff_tracker = TurnDiffTracker::new();
let tool_name = "shell";
let sub_id = "test-sub".to_string();
let call_id = "test-call".to_string();
let resp = handle_container_exec_with_params(
tool_name,
params,
&session,
&turn_context,
@@ -3668,7 +3770,6 @@ mod tests {
turn_context.sandbox_policy = SandboxPolicy::DangerFullAccess;
let resp2 = handle_container_exec_with_params(
tool_name,
params2,
&session,
&turn_context,

View File

@@ -1,431 +1,25 @@
use shlex::split as shlex_split;
/// On Windows, we conservatively allow only clearly read-only PowerShell invocations
/// that match a small safelist. Anything else (including direct CMD commands) is unsafe.
pub fn is_safe_command_windows(command: &[String]) -> bool {
if let Some(commands) = try_parse_powershell_command_sequence(command) {
return commands
.iter()
.all(|cmd| is_safe_powershell_command(cmd.as_slice()));
}
// Only PowerShell invocations are allowed on Windows for now; anything else is unsafe.
false
}
/// Returns each command sequence if the invocation starts with a PowerShell binary.
/// For example, the tokens from `pwsh Get-ChildItem | Measure-Object` become two sequences.
fn try_parse_powershell_command_sequence(command: &[String]) -> Option<Vec<Vec<String>>> {
let (exe, rest) = command.split_first()?;
if !is_powershell_executable(exe) {
return None;
}
parse_powershell_invocation(rest)
}
/// Parses a PowerShell invocation into discrete command vectors, rejecting unsafe patterns.
fn parse_powershell_invocation(args: &[String]) -> Option<Vec<Vec<String>>> {
if args.is_empty() {
// Examples rejected here: "pwsh" and "powershell.exe" with no additional arguments.
return None;
}
let mut idx = 0;
while idx < args.len() {
let arg = &args[idx];
let lower = arg.to_ascii_lowercase();
match lower.as_str() {
"-command" | "/command" | "-c" => {
let script = args.get(idx + 1)?;
if idx + 2 != args.len() {
// Reject if there is more than one token representing the actual command.
// Examples rejected here: "pwsh -Command foo bar" and "powershell -c ls extra".
return None;
}
return parse_powershell_script(script);
}
_ if lower.starts_with("-command:") || lower.starts_with("/command:") => {
if idx + 1 != args.len() {
// Reject if there are more tokens after the command itself.
// Examples rejected here: "pwsh -Command:dir C:\\" and "powershell /Command:dir C:\\" with trailing args.
return None;
}
let script = arg.split_once(':')?.1;
return parse_powershell_script(script);
}
// Benign, no-arg flags we tolerate.
"-nologo" | "-noprofile" | "-noninteractive" | "-mta" | "-sta" => {
idx += 1;
continue;
}
// Explicitly forbidden/opaque or unnecessary for read-only operations.
"-encodedcommand" | "-ec" | "-file" | "/file" | "-windowstyle" | "-executionpolicy"
| "-workingdirectory" => {
// Examples rejected here: "pwsh -EncodedCommand ..." and "powershell -File script.ps1".
return None;
}
// Unknown switch → bail conservatively.
_ if lower.starts_with('-') => {
// Examples rejected here: "pwsh -UnknownFlag" and "powershell -foo bar".
return None;
}
// If we hit non-flag tokens, treat the remainder as a command sequence.
// This happens if powershell is invoked without -Command, e.g.
// ["pwsh", "-NoLogo", "git", "-c", "core.pager=cat", "status"]
_ => {
return split_into_commands(args[idx..].to_vec());
}
}
}
// Examples rejected here: "pwsh" and "powershell.exe -NoLogo" without a script.
None
}
/// Tokenizes an inline PowerShell script and delegates to the command splitter.
/// Examples of when this is called: pwsh.exe -Command '<script>' or pwsh.exe -Command:<script>
fn parse_powershell_script(script: &str) -> Option<Vec<Vec<String>>> {
let tokens = shlex_split(script)?;
split_into_commands(tokens)
}
/// Splits tokens into pipeline segments while ensuring no unsafe separators slip through.
/// e.g. Get-ChildItem | Measure-Object -> [['Get-ChildItem'], ['Measure-Object']]
fn split_into_commands(tokens: Vec<String>) -> Option<Vec<Vec<String>>> {
if tokens.is_empty() {
// Examples rejected here: "pwsh -Command ''" and "powershell -Command \"\"".
return None;
}
let mut commands = Vec::new();
let mut current = Vec::new();
for token in tokens.into_iter() {
match token.as_str() {
"|" | "||" | "&&" | ";" => {
if current.is_empty() {
// Examples rejected here: "pwsh -Command '| Get-ChildItem'" and "pwsh -Command '; dir'".
return None;
}
commands.push(current);
current = Vec::new();
}
// Reject if any token embeds separators, redirection, or call operator characters.
_ if token.contains(['|', ';', '>', '<', '&']) || token.contains("$(") => {
// Examples rejected here: "pwsh -Command 'dir|select'" and "pwsh -Command 'echo hi > out.txt'".
return None;
}
_ => current.push(token),
}
}
if current.is_empty() {
// Examples rejected here: "pwsh -Command 'dir |'" and "pwsh -Command 'Get-ChildItem ;'".
return None;
}
commands.push(current);
Some(commands)
}
/// Returns true when the executable name is one of the supported PowerShell binaries.
fn is_powershell_executable(exe: &str) -> bool {
matches!(
exe.to_ascii_lowercase().as_str(),
"powershell" | "powershell.exe" | "pwsh" | "pwsh.exe"
)
}
/// Validates that a parsed PowerShell command stays within our read-only safelist.
/// Everything before this is parsing, and rejecting things that make us feel uncomfortable.
fn is_safe_powershell_command(words: &[String]) -> bool {
if words.is_empty() {
// Examples rejected here: "pwsh -Command ''" and "pwsh -Command \"\"".
return false;
}
// Reject nested unsafe cmdlets inside parentheses or arguments
for w in words.iter() {
let inner = w
.trim_matches(|c| c == '(' || c == ')')
.trim_start_matches('-')
.to_ascii_lowercase();
if matches!(
inner.as_str(),
"set-content"
| "add-content"
| "out-file"
| "new-item"
| "remove-item"
| "move-item"
| "copy-item"
| "rename-item"
| "start-process"
| "stop-process"
) {
// Examples rejected here: "Write-Output (Set-Content foo6.txt 'abc')" and "Get-Content (New-Item bar.txt)".
return false;
}
}
// Block PowerShell call operator or any redirection explicitly.
if words.iter().any(|w| {
matches!(
w.as_str(),
"&" | ">" | ">>" | "1>" | "2>" | "2>&1" | "*>" | "<" | "<<"
)
}) {
// Examples rejected here: "pwsh -Command '& Remove-Item foo'" and "pwsh -Command 'Get-Content foo > bar'".
return false;
}
let command = words[0]
.trim_matches(|c| c == '(' || c == ')')
.trim_start_matches('-')
.to_ascii_lowercase();
match command.as_str() {
"echo" | "write-output" | "write-host" => true, // (no redirection allowed)
"dir" | "ls" | "get-childitem" | "gci" => true,
"cat" | "type" | "gc" | "get-content" => true,
"select-string" | "sls" | "findstr" => true,
"measure-object" | "measure" => true,
"get-location" | "gl" | "pwd" => true,
"test-path" | "tp" => true,
"resolve-path" | "rvpa" => true,
"select-object" | "select" => true,
"get-item" => true,
"git" => is_safe_git_command(words),
"rg" => is_safe_ripgrep(words),
// Extra safety: explicitly prohibit common side-effecting cmdlets regardless of args.
"set-content" | "add-content" | "out-file" | "new-item" | "remove-item" | "move-item"
| "copy-item" | "rename-item" | "start-process" | "stop-process" => {
// Examples rejected here: "pwsh -Command 'Set-Content notes.txt data'" and "pwsh -Command 'Remove-Item temp.log'".
false
}
_ => {
// Examples rejected here: "pwsh -Command 'Invoke-WebRequest https://example.com'" and "pwsh -Command 'Start-Service Spooler'".
false
}
}
}
/// Checks that an `rg` invocation avoids options that can spawn arbitrary executables.
fn is_safe_ripgrep(words: &[String]) -> bool {
const UNSAFE_RIPGREP_OPTIONS_WITH_ARGS: &[&str] = &["--pre", "--hostname-bin"];
const UNSAFE_RIPGREP_OPTIONS_WITHOUT_ARGS: &[&str] = &["--search-zip", "-z"];
!words.iter().skip(1).any(|arg| {
let arg_lc = arg.to_ascii_lowercase();
// Examples rejected here: "pwsh -Command 'rg --pre cat pattern'" and "pwsh -Command 'rg --search-zip pattern'".
UNSAFE_RIPGREP_OPTIONS_WITHOUT_ARGS.contains(&arg_lc.as_str())
|| UNSAFE_RIPGREP_OPTIONS_WITH_ARGS
.iter()
.any(|opt| arg_lc == *opt || arg_lc.starts_with(&format!("{opt}=")))
})
}
/// Ensures a Git command sticks to whitelisted read-only subcommands and flags.
fn is_safe_git_command(words: &[String]) -> bool {
const SAFE_SUBCOMMANDS: &[&str] = &["status", "log", "show", "diff", "cat-file"];
let mut iter = words.iter().skip(1);
while let Some(arg) = iter.next() {
let arg_lc = arg.to_ascii_lowercase();
if arg.starts_with('-') {
if arg.eq_ignore_ascii_case("-c") || arg.eq_ignore_ascii_case("--config") {
if iter.next().is_none() {
// Examples rejected here: "pwsh -Command 'git -c'" and "pwsh -Command 'git --config'".
return false;
}
continue;
}
if arg_lc.starts_with("-c=")
|| arg_lc.starts_with("--config=")
|| arg_lc.starts_with("--git-dir=")
|| arg_lc.starts_with("--work-tree=")
{
continue;
}
if arg.eq_ignore_ascii_case("--git-dir") || arg.eq_ignore_ascii_case("--work-tree") {
if iter.next().is_none() {
// Examples rejected here: "pwsh -Command 'git --git-dir'" and "pwsh -Command 'git --work-tree'".
return false;
}
continue;
}
continue;
}
return SAFE_SUBCOMMANDS.contains(&arg_lc.as_str());
}
// Examples rejected here: "pwsh -Command 'git'" and "pwsh -Command 'git status --short | Remove-Item foo'".
// This is a WIP. This will eventually contain a real list of common safe Windows commands.
pub fn is_safe_command_windows(_command: &[String]) -> bool {
false
}
#[cfg(test)]
mod tests {
use super::is_safe_command_windows;
use std::string::ToString;
/// Converts a slice of string literals into owned `String`s for the tests.
fn vec_str(args: &[&str]) -> Vec<String> {
args.iter().map(ToString::to_string).collect()
}
#[test]
fn recognizes_safe_powershell_wrappers() {
assert!(is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-NoLogo",
"-Command",
"Get-ChildItem -Path .",
])));
assert!(is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-NoProfile",
"-Command",
"git status",
])));
assert!(is_safe_command_windows(&vec_str(&[
"powershell.exe",
"Get-Content",
"Cargo.toml",
])));
// pwsh parity
assert!(is_safe_command_windows(&vec_str(&[
"pwsh.exe",
"-NoProfile",
"-Command",
"Get-ChildItem",
])));
}
#[test]
fn allows_read_only_pipelines_and_git_usage() {
assert!(is_safe_command_windows(&vec_str(&[
"pwsh",
"-NoLogo",
"-NoProfile",
"-Command",
"rg --files-with-matches foo | Measure-Object | Select-Object -ExpandProperty Count",
])));
assert!(is_safe_command_windows(&vec_str(&[
"pwsh",
"-NoLogo",
"-NoProfile",
"-Command",
"Get-Content foo.rs | Select-Object -Skip 200",
])));
assert!(is_safe_command_windows(&vec_str(&[
"pwsh",
"-NoLogo",
"-NoProfile",
"-Command",
"git -c core.pager=cat show HEAD:foo.rs",
])));
assert!(is_safe_command_windows(&vec_str(&[
"pwsh",
"-Command",
"-git cat-file -p HEAD:foo.rs",
])));
assert!(is_safe_command_windows(&vec_str(&[
"pwsh",
"-Command",
"(Get-Content foo.rs -Raw)",
])));
assert!(is_safe_command_windows(&vec_str(&[
"pwsh",
"-Command",
"Get-Item foo.rs | Select-Object Length",
])));
}
#[test]
fn rejects_powershell_commands_with_side_effects() {
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-NoLogo",
"-Command",
"Remove-Item foo.txt",
])));
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-NoProfile",
"-Command",
"rg --pre cat",
])));
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Set-Content foo.txt 'hello'",
])));
// Redirections are blocked
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"echo hi > out.txt",
])));
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Get-Content x | Out-File y",
])));
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Write-Output foo 2> err.txt",
])));
// Call operator is blocked
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"& Remove-Item foo",
])));
// Chained safe + unsafe must fail
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Get-ChildItem; Remove-Item foo",
])));
// Nested unsafe cmdlet inside safe command must fail
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Write-Output (Set-Content foo6.txt 'abc')",
])));
// Additional nested unsafe cmdlet examples must fail
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Write-Host (Remove-Item foo.txt)",
])));
assert!(!is_safe_command_windows(&vec_str(&[
"powershell.exe",
"-Command",
"Get-Content (New-Item bar.txt)",
])));
fn everything_is_unsafe() {
for cmd in [
vec_str(&["powershell.exe", "-NoLogo", "-Command", "echo hello"]),
vec_str(&["copy", "foo", "bar"]),
vec_str(&["del", "file.txt"]),
vec_str(&["powershell.exe", "Get-ChildItem"]),
] {
assert!(!is_safe_command_windows(&cmd));
}
}
}

View File

@@ -1,12 +1,8 @@
use crate::config_profile::ConfigProfile;
use crate::config_types::DEFAULT_OTEL_ENVIRONMENT;
use crate::config_types::History;
use crate::config_types::McpServerConfig;
use crate::config_types::McpServerTransportConfig;
use crate::config_types::Notifications;
use crate::config_types::OtelConfig;
use crate::config_types::OtelConfigToml;
use crate::config_types::OtelExporterKind;
use crate::config_types::ReasoningSummaryFormat;
use crate::config_types::SandboxWorkspaceWrite;
use crate::config_types::ShellEnvironmentPolicy;
@@ -23,12 +19,12 @@ use crate::openai_model_info::get_model_info;
use crate::protocol::AskForApproval;
use crate::protocol::SandboxPolicy;
use anyhow::Context;
use codex_app_server_protocol::Tools;
use codex_app_server_protocol::UserSavedConfig;
use codex_protocol::config_types::ReasoningEffort;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::config_types::SandboxMode;
use codex_protocol::config_types::Verbosity;
use codex_protocol::mcp_protocol::Tools;
use codex_protocol::mcp_protocol::UserSavedConfig;
use dirs::home_dir;
use serde::Deserialize;
use std::collections::BTreeMap;
@@ -141,9 +137,6 @@ pub struct Config {
/// Maximum number of bytes to include from an AGENTS.md project doc file.
pub project_doc_max_bytes: usize,
/// Additional filenames to try when looking for project-level docs.
pub project_doc_fallback_filenames: Vec<String>,
/// Directory containing all Codex state (defaults to `~/.codex` but can be
/// overridden by the `CODEX_HOME` environment variable).
pub codex_home: PathBuf,
@@ -206,9 +199,6 @@ pub struct Config {
/// All characters are inserted as they are received, and no buffering
/// or placeholder replacement will occur for fast keypress bursts.
pub disable_paste_burst: bool,
/// OTEL configuration (exporter type, endpoint, headers, etc.).
pub otel: crate::config_types::OtelConfig,
}
impl Config {
@@ -436,7 +426,7 @@ fn set_project_trusted_inner(doc: &mut DocumentMut, project_path: &Path) -> anyh
.get_mut(project_key.as_str())
.and_then(|i| i.as_table_mut())
else {
return Err(anyhow::anyhow!("project table missing for {project_key}"));
return Err(anyhow::anyhow!("project table missing for {}", project_key));
};
proj_tbl.set_implicit(false);
proj_tbl["trust_level"] = toml_edit::value("trusted");
@@ -673,9 +663,6 @@ pub struct ConfigToml {
/// Maximum number of bytes to include from an AGENTS.md project doc file.
pub project_doc_max_bytes: Option<usize>,
/// Ordered list of fallback filenames to look for when AGENTS.md is missing.
pub project_doc_fallback_filenames: Option<Vec<String>>,
/// Profile to use from the `profiles` map.
pub profile: Option<String>,
@@ -732,9 +719,6 @@ pub struct ConfigToml {
/// All characters are inserted as they are received, and no buffering
/// or placeholder replacement will occur for fast keypress bursts.
pub disable_paste_burst: Option<bool>,
/// OTEL configuration.
pub otel: Option<crate::config_types::OtelConfigToml>,
}
impl From<ConfigToml> for UserSavedConfig {
@@ -1044,19 +1028,6 @@ impl Config {
mcp_servers: cfg.mcp_servers,
model_providers,
project_doc_max_bytes: cfg.project_doc_max_bytes.unwrap_or(PROJECT_DOC_MAX_BYTES),
project_doc_fallback_filenames: cfg
.project_doc_fallback_filenames
.unwrap_or_default()
.into_iter()
.filter_map(|name| {
let trimmed = name.trim();
if trimmed.is_empty() {
None
} else {
Some(trimmed.to_string())
}
})
.collect(),
codex_home,
history,
file_opener: cfg.file_opener.unwrap_or(UriBasedFileOpener::VsCode),
@@ -1079,7 +1050,7 @@ impl Config {
.chatgpt_base_url
.or(cfg.chatgpt_base_url)
.unwrap_or("https://chatgpt.com/backend-api/".to_string()),
include_plan_tool: include_plan_tool.unwrap_or(true),
include_plan_tool: include_plan_tool.unwrap_or(false),
include_apply_patch_tool: include_apply_patch_tool.unwrap_or(false),
tools_web_search_request,
use_experimental_streamable_shell_tool: cfg
@@ -1097,19 +1068,6 @@ impl Config {
.as_ref()
.map(|t| t.notifications.clone())
.unwrap_or_default(),
otel: {
let t: OtelConfigToml = cfg.otel.unwrap_or_default();
let log_user_prompt = t.log_user_prompt.unwrap_or(false);
let environment = t
.environment
.unwrap_or(DEFAULT_OTEL_ENVIRONMENT.to_string());
let exporter = t.exporter.unwrap_or(OtelExporterKind::None);
OtelConfig {
log_user_prompt,
environment,
exporter,
}
},
};
Ok(config)
}
@@ -1830,7 +1788,6 @@ model_verbosity = "high"
mcp_servers: HashMap::new(),
model_providers: fixture.model_provider_map.clone(),
project_doc_max_bytes: PROJECT_DOC_MAX_BYTES,
project_doc_fallback_filenames: Vec::new(),
codex_home: fixture.codex_home(),
history: History::default(),
file_opener: UriBasedFileOpener::VsCode,
@@ -1842,7 +1799,7 @@ model_verbosity = "high"
model_verbosity: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
include_plan_tool: true,
include_plan_tool: false,
include_apply_patch_tool: false,
tools_web_search_request: false,
use_experimental_streamable_shell_tool: false,
@@ -1852,7 +1809,6 @@ model_verbosity = "high"
active_profile: Some("o3".to_string()),
disable_paste_burst: false,
tui_notifications: Default::default(),
otel: OtelConfig::default(),
},
o3_profile_config
);
@@ -1891,7 +1847,6 @@ model_verbosity = "high"
mcp_servers: HashMap::new(),
model_providers: fixture.model_provider_map.clone(),
project_doc_max_bytes: PROJECT_DOC_MAX_BYTES,
project_doc_fallback_filenames: Vec::new(),
codex_home: fixture.codex_home(),
history: History::default(),
file_opener: UriBasedFileOpener::VsCode,
@@ -1903,7 +1858,7 @@ model_verbosity = "high"
model_verbosity: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
include_plan_tool: true,
include_plan_tool: false,
include_apply_patch_tool: false,
tools_web_search_request: false,
use_experimental_streamable_shell_tool: false,
@@ -1913,7 +1868,6 @@ model_verbosity = "high"
active_profile: Some("gpt3".to_string()),
disable_paste_burst: false,
tui_notifications: Default::default(),
otel: OtelConfig::default(),
};
assert_eq!(expected_gpt3_profile_config, gpt3_profile_config);
@@ -1967,7 +1921,6 @@ model_verbosity = "high"
mcp_servers: HashMap::new(),
model_providers: fixture.model_provider_map.clone(),
project_doc_max_bytes: PROJECT_DOC_MAX_BYTES,
project_doc_fallback_filenames: Vec::new(),
codex_home: fixture.codex_home(),
history: History::default(),
file_opener: UriBasedFileOpener::VsCode,
@@ -1979,7 +1932,7 @@ model_verbosity = "high"
model_verbosity: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
include_plan_tool: true,
include_plan_tool: false,
include_apply_patch_tool: false,
tools_web_search_request: false,
use_experimental_streamable_shell_tool: false,
@@ -1989,7 +1942,6 @@ model_verbosity = "high"
active_profile: Some("zdr".to_string()),
disable_paste_burst: false,
tui_notifications: Default::default(),
otel: OtelConfig::default(),
};
assert_eq!(expected_zdr_profile_config, zdr_profile_config);
@@ -2029,7 +1981,6 @@ model_verbosity = "high"
mcp_servers: HashMap::new(),
model_providers: fixture.model_provider_map.clone(),
project_doc_max_bytes: PROJECT_DOC_MAX_BYTES,
project_doc_fallback_filenames: Vec::new(),
codex_home: fixture.codex_home(),
history: History::default(),
file_opener: UriBasedFileOpener::VsCode,
@@ -2041,7 +1992,7 @@ model_verbosity = "high"
model_verbosity: Some(Verbosity::High),
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
include_plan_tool: true,
include_plan_tool: false,
include_apply_patch_tool: false,
tools_web_search_request: false,
use_experimental_streamable_shell_tool: false,
@@ -2051,7 +2002,6 @@ model_verbosity = "high"
active_profile: Some("gpt5".to_string()),
disable_paste_burst: false,
tui_notifications: Default::default(),
otel: OtelConfig::default(),
};
assert_eq!(expected_gpt5_profile_config, gpt5_profile_config);

View File

@@ -22,7 +22,7 @@ pub struct ConfigProfile {
pub experimental_instructions_file: Option<PathBuf>,
}
impl From<ConfigProfile> for codex_app_server_protocol::Profile {
impl From<ConfigProfile> for codex_protocol::mcp_protocol::Profile {
fn from(config_profile: ConfigProfile) -> Self {
Self {
model: config_profile.model,

View File

@@ -13,8 +13,6 @@ use serde::Deserialize;
use serde::Serialize;
use serde::de::Error as SerdeError;
pub const DEFAULT_OTEL_ENVIRONMENT: &str = "dev";
#[derive(Serialize, Debug, Clone, PartialEq)]
pub struct McpServerConfig {
#[serde(flatten)]
@@ -221,64 +219,6 @@ pub enum HistoryPersistence {
None,
}
// ===== OTEL configuration =====
#[derive(Deserialize, Debug, Clone, PartialEq)]
#[serde(rename_all = "kebab-case")]
pub enum OtelHttpProtocol {
/// Binary payload
Binary,
/// JSON payload
Json,
}
/// Which OTEL exporter to use.
#[derive(Deserialize, Debug, Clone, PartialEq)]
#[serde(rename_all = "kebab-case")]
pub enum OtelExporterKind {
None,
OtlpHttp {
endpoint: String,
headers: HashMap<String, String>,
protocol: OtelHttpProtocol,
},
OtlpGrpc {
endpoint: String,
headers: HashMap<String, String>,
},
}
/// OTEL settings loaded from config.toml. Fields are optional so we can apply defaults.
#[derive(Deserialize, Debug, Clone, PartialEq, Default)]
pub struct OtelConfigToml {
/// Log user prompt in traces
pub log_user_prompt: Option<bool>,
/// Mark traces with environment (dev, staging, prod, test). Defaults to dev.
pub environment: Option<String>,
/// Exporter to use. Defaults to `otlp-file`.
pub exporter: Option<OtelExporterKind>,
}
/// Effective OTEL settings after defaults are applied.
#[derive(Debug, Clone, PartialEq)]
pub struct OtelConfig {
pub log_user_prompt: bool,
pub environment: String,
pub exporter: OtelExporterKind,
}
impl Default for OtelConfig {
fn default() -> Self {
OtelConfig {
log_user_prompt: false,
environment: DEFAULT_OTEL_ENVIRONMENT.to_owned(),
exporter: OtelExporterKind::None,
}
}
}
#[derive(Debug, Clone, PartialEq, Eq, Deserialize)]
#[serde(untagged)]
pub enum Notifications {
@@ -313,7 +253,7 @@ pub struct SandboxWorkspaceWrite {
pub exclude_slash_tmp: bool,
}
impl From<SandboxWorkspaceWrite> for codex_app_server_protocol::SandboxSettings {
impl From<SandboxWorkspaceWrite> for codex_protocol::mcp_protocol::SandboxSettings {
fn from(sandbox_workspace_write: SandboxWorkspaceWrite) -> Self {
Self {
writable_roots: sandbox_workspace_write.writable_roots,

View File

@@ -13,7 +13,7 @@ use crate::protocol::Event;
use crate::protocol::EventMsg;
use crate::protocol::SessionConfiguredEvent;
use crate::rollout::RolloutRecorder;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::ConversationId;
use codex_protocol::models::ResponseItem;
use codex_protocol::protocol::InitialHistory;
use codex_protocol::protocol::RolloutItem;

View File

@@ -63,88 +63,16 @@ pub async fn discover_prompts_in_excluding(
Ok(s) => s,
Err(_) => continue,
};
let (description, argument_hint, body) = parse_frontmatter(&content);
out.push(CustomPrompt {
name,
path,
content: body,
description,
argument_hint,
content,
});
}
out.sort_by(|a, b| a.name.cmp(&b.name));
out
}
/// Parse optional YAML-like frontmatter at the beginning of `content`.
/// Supported keys:
/// - `description`: short description shown in the slash popup
/// - `argument-hint` or `argument_hint`: brief hint string shown after the description
/// Returns (description, argument_hint, body_without_frontmatter).
fn parse_frontmatter(content: &str) -> (Option<String>, Option<String>, String) {
let mut segments = content.split_inclusive('\n');
let Some(first_segment) = segments.next() else {
return (None, None, String::new());
};
let first_line = first_segment.trim_end_matches(['\r', '\n']);
if first_line.trim() != "---" {
return (None, None, content.to_string());
}
let mut desc: Option<String> = None;
let mut hint: Option<String> = None;
let mut frontmatter_closed = false;
let mut consumed = first_segment.len();
for segment in segments {
let line = segment.trim_end_matches(['\r', '\n']);
let trimmed = line.trim();
if trimmed == "---" {
frontmatter_closed = true;
consumed += segment.len();
break;
}
if trimmed.is_empty() || trimmed.starts_with('#') {
consumed += segment.len();
continue;
}
if let Some((k, v)) = trimmed.split_once(':') {
let key = k.trim().to_ascii_lowercase();
let mut val = v.trim().to_string();
if val.len() >= 2 {
let bytes = val.as_bytes();
let first = bytes[0];
let last = bytes[bytes.len() - 1];
if (first == b'\"' && last == b'\"') || (first == b'\'' && last == b'\'') {
val = val[1..val.len().saturating_sub(1)].to_string();
}
}
match key.as_str() {
"description" => desc = Some(val),
"argument-hint" | "argument_hint" => hint = Some(val),
_ => {}
}
}
consumed += segment.len();
}
if !frontmatter_closed {
// Unterminated frontmatter: treat input as-is.
return (None, None, content.to_string());
}
let body = if consumed >= content.len() {
String::new()
} else {
content[consumed..].to_string()
};
(desc, hint, body)
}
#[cfg(test)]
mod tests {
use super::*;
@@ -196,31 +124,4 @@ mod tests {
let names: Vec<String> = found.into_iter().map(|e| e.name).collect();
assert_eq!(names, vec!["good"]);
}
#[tokio::test]
async fn parses_frontmatter_and_strips_from_body() {
let tmp = tempdir().expect("create TempDir");
let dir = tmp.path();
let file = dir.join("withmeta.md");
let text = "---\nname: ignored\ndescription: \"Quick review command\"\nargument-hint: \"[file] [priority]\"\n---\nActual body with $1 and $ARGUMENTS";
fs::write(&file, text).unwrap();
let found = discover_prompts_in(dir).await;
assert_eq!(found.len(), 1);
let p = &found[0];
assert_eq!(p.name, "withmeta");
assert_eq!(p.description.as_deref(), Some("Quick review command"));
assert_eq!(p.argument_hint.as_deref(), Some("[file] [priority]"));
// Body should not include the frontmatter delimiters.
assert_eq!(p.content, "Actual body with $1 and $ARGUMENTS");
}
#[test]
fn parse_frontmatter_preserves_body_newlines() {
let content = "---\r\ndescription: \"Line endings\"\r\nargument_hint: \"[arg]\"\r\n---\r\nFirst line\r\nSecond line\r\n";
let (desc, hint, body) = parse_frontmatter(content);
assert_eq!(desc.as_deref(), Some("Line endings"));
assert_eq!(hint.as_deref(), Some("[arg]"));
assert_eq!(body, "First line\r\nSecond line\r\n");
}
}

View File

@@ -2,7 +2,6 @@ use crate::spawn::CODEX_SANDBOX_ENV_VAR;
use reqwest::header::HeaderValue;
use std::sync::LazyLock;
use std::sync::Mutex;
use std::sync::OnceLock;
/// Set this to add a suffix to the User-Agent string.
///
@@ -27,15 +26,8 @@ pub struct Originator {
pub value: String,
pub header_value: HeaderValue,
}
static ORIGINATOR: OnceLock<Originator> = OnceLock::new();
#[derive(Debug)]
pub enum SetOriginatorError {
InvalidHeaderValue,
AlreadyInitialized,
}
fn init_originator_from_env() -> Originator {
pub static ORIGINATOR: LazyLock<Originator> = LazyLock::new(|| {
let default = "codex_cli_rs";
let value = std::env::var(CODEX_INTERNAL_ORIGINATOR_OVERRIDE_ENV_VAR)
.unwrap_or_else(|_| default.to_string());
@@ -53,34 +45,14 @@ fn init_originator_from_env() -> Originator {
}
}
}
}
fn build_originator(value: String) -> Result<Originator, SetOriginatorError> {
let header_value =
HeaderValue::from_str(&value).map_err(|_| SetOriginatorError::InvalidHeaderValue)?;
Ok(Originator {
value,
header_value,
})
}
pub fn set_default_originator(value: &str) -> Result<(), SetOriginatorError> {
let originator = build_originator(value.to_string())?;
ORIGINATOR
.set(originator)
.map_err(|_| SetOriginatorError::AlreadyInitialized)
}
pub fn originator() -> &'static Originator {
ORIGINATOR.get_or_init(init_originator_from_env)
}
});
pub fn get_codex_user_agent() -> String {
let build_version = env!("CARGO_PKG_VERSION");
let os_info = os_info::get();
let prefix = format!(
"{}/{build_version} ({} {}; {}) {}",
originator().value.as_str(),
ORIGINATOR.value.as_str(),
os_info.os_type(),
os_info.version(),
os_info.architecture().unwrap_or("unknown"),
@@ -128,7 +100,7 @@ fn sanitize_user_agent(candidate: String, fallback: &str) -> String {
tracing::warn!(
"Falling back to default Codex originator because base user agent string is invalid"
);
originator().value.clone()
ORIGINATOR.value.clone()
}
}
@@ -137,7 +109,7 @@ pub fn create_client() -> reqwest::Client {
use reqwest::header::HeaderMap;
let mut headers = HeaderMap::new();
headers.insert("originator", originator().header_value.clone());
headers.insert("originator", ORIGINATOR.header_value.clone());
let ua = get_codex_user_agent();
let mut builder = reqwest::Client::builder()

View File

@@ -1,7 +1,7 @@
use crate::exec::ExecToolCallOutput;
use crate::token_data::KnownPlan;
use crate::token_data::PlanType;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::ConversationId;
use codex_protocol::protocol::RateLimitSnapshot;
use reqwest::StatusCode;
use serde_json;
@@ -76,8 +76,8 @@ pub enum CodexErr {
Interrupted,
/// Unexpected HTTP status code.
#[error("{0}")]
UnexpectedStatus(UnexpectedResponseError),
#[error("unexpected status {0}: {1}")]
UnexpectedStatus(StatusCode, String),
#[error("{0}")]
UsageLimitReached(UsageLimitReachedError),
@@ -91,8 +91,8 @@ pub enum CodexErr {
InternalServerError,
/// Retry limit exceeded.
#[error("{0}")]
RetryLimit(RetryLimitReachedError),
#[error("exceeded retry limit, last status: {0}")]
RetryLimit(StatusCode),
/// Agent loop died unexpectedly
#[error("internal error; agent loop died unexpectedly")]
@@ -135,49 +135,6 @@ pub enum CodexErr {
EnvVar(EnvVarError),
}
#[derive(Debug)]
pub struct UnexpectedResponseError {
pub status: StatusCode,
pub body: String,
pub request_id: Option<String>,
}
impl std::fmt::Display for UnexpectedResponseError {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(
f,
"unexpected status {}: {}{}",
self.status,
self.body,
self.request_id
.as_ref()
.map(|id| format!(", request id: {id}"))
.unwrap_or_default()
)
}
}
impl std::error::Error for UnexpectedResponseError {}
#[derive(Debug)]
pub struct RetryLimitReachedError {
pub status: StatusCode,
pub request_id: Option<String>,
}
impl std::fmt::Display for RetryLimitReachedError {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(
f,
"exceeded retry limit, last status: {}{}",
self.status,
self.request_id
.as_ref()
.map(|id| format!(", request id: {id}"))
.unwrap_or_default()
)
}
}
#[derive(Debug)]
pub struct UsageLimitReachedError {
pub(crate) plan_type: Option<PlanType>,

View File

@@ -1,101 +0,0 @@
use std::collections::HashMap;
use std::env;
use async_trait::async_trait;
use crate::CODEX_APPLY_PATCH_ARG1;
use crate::apply_patch::ApplyPatchExec;
use crate::exec::ExecParams;
use crate::function_tool::FunctionCallError;
pub(crate) enum ExecutionMode {
Shell,
ApplyPatch(ApplyPatchExec),
}
#[async_trait]
/// Backend-specific hooks that prepare and post-process execution requests for a
/// given [`ExecutionMode`].
pub(crate) trait ExecutionBackend: Send + Sync {
fn prepare(
&self,
params: ExecParams,
// Required for downcasting the apply_patch.
mode: &ExecutionMode,
) -> Result<ExecParams, FunctionCallError>;
fn stream_stdout(&self, _mode: &ExecutionMode) -> bool {
true
}
}
static SHELL_BACKEND: ShellBackend = ShellBackend;
static APPLY_PATCH_BACKEND: ApplyPatchBackend = ApplyPatchBackend;
pub(crate) fn backend_for_mode(mode: &ExecutionMode) -> &'static dyn ExecutionBackend {
match mode {
ExecutionMode::Shell => &SHELL_BACKEND,
ExecutionMode::ApplyPatch(_) => &APPLY_PATCH_BACKEND,
}
}
struct ShellBackend;
#[async_trait]
impl ExecutionBackend for ShellBackend {
fn prepare(
&self,
params: ExecParams,
mode: &ExecutionMode,
) -> Result<ExecParams, FunctionCallError> {
match mode {
ExecutionMode::Shell => Ok(params),
_ => Err(FunctionCallError::RespondToModel(
"shell backend invoked with non-shell mode".to_string(),
)),
}
}
}
struct ApplyPatchBackend;
#[async_trait]
impl ExecutionBackend for ApplyPatchBackend {
fn prepare(
&self,
params: ExecParams,
mode: &ExecutionMode,
) -> Result<ExecParams, FunctionCallError> {
match mode {
ExecutionMode::ApplyPatch(exec) => {
let path_to_codex = env::current_exe()
.ok()
.map(|p| p.to_string_lossy().to_string())
.ok_or_else(|| {
FunctionCallError::RespondToModel(
"failed to determine path to codex executable".to_string(),
)
})?;
let patch = exec.action.patch.clone();
Ok(ExecParams {
command: vec![path_to_codex, CODEX_APPLY_PATCH_ARG1.to_string(), patch],
cwd: exec.action.cwd.clone(),
timeout_ms: params.timeout_ms,
// Run apply_patch with a minimal environment for determinism and to
// avoid leaking host environment variables into the patch process.
env: HashMap::new(),
with_escalated_permissions: params.with_escalated_permissions,
justification: params.justification,
})
}
ExecutionMode::Shell => Err(FunctionCallError::RespondToModel(
"apply_patch backend invoked without patch context".to_string(),
)),
}
}
fn stream_stdout(&self, _mode: &ExecutionMode) -> bool {
false
}
}

View File

@@ -1,51 +0,0 @@
use std::collections::HashSet;
use std::sync::Arc;
use std::sync::Mutex;
#[derive(Clone, Debug, Default)]
/// Thread-safe store of user approvals so repeated commands can reuse
/// previously granted trust.
pub(crate) struct ApprovalCache {
inner: Arc<Mutex<HashSet<Vec<String>>>>,
}
impl ApprovalCache {
pub(crate) fn insert(&self, command: Vec<String>) {
if command.is_empty() {
return;
}
if let Ok(mut guard) = self.inner.lock() {
guard.insert(command);
}
}
pub(crate) fn snapshot(&self) -> HashSet<Vec<String>> {
self.inner.lock().map(|g| g.clone()).unwrap_or_default()
}
}
#[cfg(test)]
mod tests {
use super::*;
use pretty_assertions::assert_eq;
#[test]
fn insert_ignores_empty_and_dedupes() {
let cache = ApprovalCache::default();
// Empty should be ignored
cache.insert(vec![]);
assert!(cache.snapshot().is_empty());
// Insert a command and verify snapshot contains it
let cmd = vec!["foo".to_string(), "bar".to_string()];
cache.insert(cmd.clone());
let snap1 = cache.snapshot();
assert!(snap1.contains(&cmd));
// Reinserting should not create duplicates
cache.insert(cmd);
let snap2 = cache.snapshot();
assert_eq!(snap1, snap2);
}
}

View File

@@ -1,64 +0,0 @@
mod backends;
mod cache;
mod runner;
mod sandbox;
pub(crate) use backends::ExecutionMode;
pub(crate) use runner::ExecutionRequest;
pub(crate) use runner::Executor;
pub(crate) use runner::ExecutorConfig;
pub(crate) use runner::normalize_exec_result;
pub(crate) mod linkers {
use crate::codex::ExecCommandContext;
use crate::exec::ExecParams;
use crate::exec::StdoutStream;
use crate::executor::backends::ExecutionMode;
use crate::executor::runner::ExecutionRequest;
pub struct PreparedExec {
pub(crate) context: ExecCommandContext,
pub(crate) request: ExecutionRequest,
}
impl PreparedExec {
pub fn new(
context: ExecCommandContext,
params: ExecParams,
approval_command: Vec<String>,
mode: ExecutionMode,
stdout_stream: Option<StdoutStream>,
use_shell_profile: bool,
) -> Self {
let request = ExecutionRequest {
params,
approval_command,
mode,
stdout_stream,
use_shell_profile,
};
Self { context, request }
}
}
}
pub mod errors {
use crate::error::CodexErr;
use crate::function_tool::FunctionCallError;
use thiserror::Error;
#[derive(Debug, Error)]
pub enum ExecError {
#[error(transparent)]
Function(#[from] FunctionCallError),
#[error(transparent)]
Codex(#[from] CodexErr),
}
impl ExecError {
pub(crate) fn rejection(msg: impl Into<String>) -> Self {
FunctionCallError::RespondToModel(msg.into()).into()
}
}
}

View File

@@ -1,387 +0,0 @@
use std::path::PathBuf;
use std::sync::Arc;
use std::sync::RwLock;
use std::time::Duration;
use super::backends::ExecutionMode;
use super::backends::backend_for_mode;
use super::cache::ApprovalCache;
use crate::codex::ExecCommandContext;
use crate::codex::Session;
use crate::error::CodexErr;
use crate::error::SandboxErr;
use crate::error::get_error_message_ui;
use crate::exec::ExecParams;
use crate::exec::ExecToolCallOutput;
use crate::exec::SandboxType;
use crate::exec::StdoutStream;
use crate::exec::StreamOutput;
use crate::exec::process_exec_tool_call;
use crate::executor::errors::ExecError;
use crate::executor::sandbox::select_sandbox;
use crate::function_tool::FunctionCallError;
use crate::protocol::AskForApproval;
use crate::protocol::ReviewDecision;
use crate::protocol::SandboxPolicy;
use crate::shell;
use codex_otel::otel_event_manager::ToolDecisionSource;
#[derive(Clone, Debug)]
pub(crate) struct ExecutorConfig {
pub(crate) sandbox_policy: SandboxPolicy,
pub(crate) sandbox_cwd: PathBuf,
codex_linux_sandbox_exe: Option<PathBuf>,
}
impl ExecutorConfig {
pub(crate) fn new(
sandbox_policy: SandboxPolicy,
sandbox_cwd: PathBuf,
codex_linux_sandbox_exe: Option<PathBuf>,
) -> Self {
Self {
sandbox_policy,
sandbox_cwd,
codex_linux_sandbox_exe,
}
}
}
/// Coordinates sandbox selection, backend-specific preparation, and command
/// execution for tool calls requested by the model.
pub(crate) struct Executor {
approval_cache: ApprovalCache,
config: Arc<RwLock<ExecutorConfig>>,
}
impl Executor {
pub(crate) fn new(config: ExecutorConfig) -> Self {
Self {
approval_cache: ApprovalCache::default(),
config: Arc::new(RwLock::new(config)),
}
}
/// Updates the sandbox policy and working directory used for future
/// executions without recreating the executor.
pub(crate) fn update_environment(&self, sandbox_policy: SandboxPolicy, sandbox_cwd: PathBuf) {
if let Ok(mut cfg) = self.config.write() {
cfg.sandbox_policy = sandbox_policy;
cfg.sandbox_cwd = sandbox_cwd;
}
}
/// Runs a prepared execution request end-to-end: prepares parameters, decides on
/// sandbox placement (prompting the user when necessary), launches the command,
/// and lets the backend post-process the final output.
pub(crate) async fn run(
&self,
mut request: ExecutionRequest,
session: &Session,
approval_policy: AskForApproval,
context: &ExecCommandContext,
) -> Result<ExecToolCallOutput, ExecError> {
if matches!(request.mode, ExecutionMode::Shell) {
request.params =
maybe_translate_shell_command(request.params, session, request.use_shell_profile);
}
// Step 1: Normalise parameters via the selected backend.
let backend = backend_for_mode(&request.mode);
let stdout_stream = if backend.stream_stdout(&request.mode) {
request.stdout_stream.clone()
} else {
None
};
request.params = backend
.prepare(request.params, &request.mode)
.map_err(ExecError::from)?;
// Step 2: Snapshot sandbox configuration so it stays stable for this run.
let config = self
.config
.read()
.map_err(|_| ExecError::rejection("executor config poisoned"))?
.clone();
// Step 3: Decide sandbox placement, prompting for approval when needed.
let sandbox_decision = select_sandbox(
&request,
approval_policy,
self.approval_cache.snapshot(),
&config,
session,
&context.sub_id,
&context.call_id,
&context.otel_event_manager,
)
.await?;
if sandbox_decision.record_session_approval {
self.approval_cache.insert(request.approval_command.clone());
}
// Step 4: Launch the command within the chosen sandbox.
let first_attempt = self
.spawn(
request.params.clone(),
sandbox_decision.initial_sandbox,
&config,
stdout_stream.clone(),
)
.await;
// Step 5: Handle sandbox outcomes, optionally escalating to an unsandboxed retry.
match first_attempt {
Ok(output) => Ok(output),
Err(CodexErr::Sandbox(SandboxErr::Timeout { output })) => {
Err(CodexErr::Sandbox(SandboxErr::Timeout { output }).into())
}
Err(CodexErr::Sandbox(error)) => {
if sandbox_decision.escalate_on_failure {
self.retry_without_sandbox(
&request,
&config,
session,
context,
stdout_stream,
error,
)
.await
} else {
Err(ExecError::rejection(format!(
"failed in sandbox {:?} with execution error: {error:?}",
sandbox_decision.initial_sandbox
)))
}
}
Err(err) => Err(err.into()),
}
}
/// Fallback path invoked when a sandboxed run is denied so the user can
/// approve rerunning without isolation.
async fn retry_without_sandbox(
&self,
request: &ExecutionRequest,
config: &ExecutorConfig,
session: &Session,
context: &ExecCommandContext,
stdout_stream: Option<StdoutStream>,
sandbox_error: SandboxErr,
) -> Result<ExecToolCallOutput, ExecError> {
session
.notify_background_event(
&context.sub_id,
format!("Execution failed: {sandbox_error}"),
)
.await;
let decision = session
.request_command_approval(
context.sub_id.to_string(),
context.call_id.to_string(),
request.approval_command.clone(),
request.params.cwd.clone(),
Some("command failed; retry without sandbox?".to_string()),
)
.await;
context.otel_event_manager.tool_decision(
&context.tool_name,
&context.call_id,
decision,
ToolDecisionSource::User,
);
match decision {
ReviewDecision::Approved | ReviewDecision::ApprovedForSession => {
if matches!(decision, ReviewDecision::ApprovedForSession) {
self.approval_cache.insert(request.approval_command.clone());
}
session
.notify_background_event(&context.sub_id, "retrying command without sandbox")
.await;
let retry_output = self
.spawn(
request.params.clone(),
SandboxType::None,
config,
stdout_stream,
)
.await?;
Ok(retry_output)
}
ReviewDecision::Denied | ReviewDecision::Abort => {
Err(ExecError::rejection("exec command rejected by user"))
}
}
}
async fn spawn(
&self,
params: ExecParams,
sandbox: SandboxType,
config: &ExecutorConfig,
stdout_stream: Option<StdoutStream>,
) -> Result<ExecToolCallOutput, CodexErr> {
process_exec_tool_call(
params,
sandbox,
&config.sandbox_policy,
&config.sandbox_cwd,
&config.codex_linux_sandbox_exe,
stdout_stream,
)
.await
}
}
fn maybe_translate_shell_command(
params: ExecParams,
session: &Session,
use_shell_profile: bool,
) -> ExecParams {
let should_translate =
matches!(session.user_shell(), shell::Shell::PowerShell(_)) || use_shell_profile;
if should_translate
&& let Some(command) = session
.user_shell()
.format_default_shell_invocation(params.command.clone())
{
return ExecParams { command, ..params };
}
params
}
pub(crate) struct ExecutionRequest {
pub params: ExecParams,
pub approval_command: Vec<String>,
pub mode: ExecutionMode,
pub stdout_stream: Option<StdoutStream>,
pub use_shell_profile: bool,
}
pub(crate) struct NormalizedExecOutput<'a> {
borrowed: Option<&'a ExecToolCallOutput>,
synthetic: Option<ExecToolCallOutput>,
}
impl<'a> NormalizedExecOutput<'a> {
pub(crate) fn event_output(&'a self) -> &'a ExecToolCallOutput {
match (self.borrowed, self.synthetic.as_ref()) {
(Some(output), _) => output,
(None, Some(output)) => output,
(None, None) => unreachable!("normalized exec output missing data"),
}
}
}
/// Converts a raw execution result into a uniform view that always exposes an
/// [`ExecToolCallOutput`], synthesizing error output when the command fails
/// before producing a response.
pub(crate) fn normalize_exec_result(
result: &Result<ExecToolCallOutput, ExecError>,
) -> NormalizedExecOutput<'_> {
match result {
Ok(output) => NormalizedExecOutput {
borrowed: Some(output),
synthetic: None,
},
Err(ExecError::Codex(CodexErr::Sandbox(SandboxErr::Timeout { output }))) => {
NormalizedExecOutput {
borrowed: Some(output.as_ref()),
synthetic: None,
}
}
Err(err) => {
let message = match err {
ExecError::Function(FunctionCallError::RespondToModel(msg)) => msg.clone(),
ExecError::Codex(e) => get_error_message_ui(e),
};
let synthetic = ExecToolCallOutput {
exit_code: -1,
stdout: StreamOutput::new(String::new()),
stderr: StreamOutput::new(message.clone()),
aggregated_output: StreamOutput::new(message),
duration: Duration::default(),
timed_out: false,
};
NormalizedExecOutput {
borrowed: None,
synthetic: Some(synthetic),
}
}
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::error::CodexErr;
use crate::error::EnvVarError;
use crate::error::SandboxErr;
use crate::exec::StreamOutput;
use pretty_assertions::assert_eq;
fn make_output(text: &str) -> ExecToolCallOutput {
ExecToolCallOutput {
exit_code: 1,
stdout: StreamOutput::new(String::new()),
stderr: StreamOutput::new(String::new()),
aggregated_output: StreamOutput::new(text.to_string()),
duration: Duration::from_millis(123),
timed_out: false,
}
}
#[test]
fn normalize_success_borrows() {
let out = make_output("ok");
let result: Result<ExecToolCallOutput, ExecError> = Ok(out);
let normalized = normalize_exec_result(&result);
assert_eq!(normalized.event_output().aggregated_output.text, "ok");
}
#[test]
fn normalize_timeout_borrows_embedded_output() {
let out = make_output("timed out payload");
let err = CodexErr::Sandbox(SandboxErr::Timeout {
output: Box::new(out),
});
let result: Result<ExecToolCallOutput, ExecError> = Err(ExecError::Codex(err));
let normalized = normalize_exec_result(&result);
assert_eq!(
normalized.event_output().aggregated_output.text,
"timed out payload"
);
}
#[test]
fn normalize_function_error_synthesizes_payload() {
let err = FunctionCallError::RespondToModel("boom".to_string());
let result: Result<ExecToolCallOutput, ExecError> = Err(ExecError::Function(err));
let normalized = normalize_exec_result(&result);
assert_eq!(normalized.event_output().aggregated_output.text, "boom");
}
#[test]
fn normalize_codex_error_synthesizes_user_message() {
// Use a simple EnvVar error which formats to a clear message
let e = CodexErr::EnvVar(EnvVarError {
var: "FOO".to_string(),
instructions: Some("set it".to_string()),
});
let result: Result<ExecToolCallOutput, ExecError> = Err(ExecError::Codex(e));
let normalized = normalize_exec_result(&result);
assert!(
normalized
.event_output()
.aggregated_output
.text
.contains("Missing environment variable: `FOO`"),
"expected synthesized user-friendly message"
);
}
}

View File

@@ -1,405 +0,0 @@
use crate::apply_patch::ApplyPatchExec;
use crate::codex::Session;
use crate::exec::SandboxType;
use crate::executor::ExecutionMode;
use crate::executor::ExecutionRequest;
use crate::executor::ExecutorConfig;
use crate::executor::errors::ExecError;
use crate::safety::SafetyCheck;
use crate::safety::assess_command_safety;
use crate::safety::assess_patch_safety;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_otel::otel_event_manager::ToolDecisionSource;
use codex_protocol::protocol::AskForApproval;
use codex_protocol::protocol::ReviewDecision;
use std::collections::HashSet;
/// Sandbox placement options selected for an execution run, including whether
/// to escalate after failures and whether approvals should persist.
pub(crate) struct SandboxDecision {
pub(crate) initial_sandbox: SandboxType,
pub(crate) escalate_on_failure: bool,
pub(crate) record_session_approval: bool,
}
impl SandboxDecision {
fn auto(sandbox: SandboxType, escalate_on_failure: bool) -> Self {
Self {
initial_sandbox: sandbox,
escalate_on_failure,
record_session_approval: false,
}
}
fn user_override(record_session_approval: bool) -> Self {
Self {
initial_sandbox: SandboxType::None,
escalate_on_failure: false,
record_session_approval,
}
}
}
fn should_escalate_on_failure(approval: AskForApproval, sandbox: SandboxType) -> bool {
matches!(
(approval, sandbox),
(
AskForApproval::UnlessTrusted | AskForApproval::OnFailure,
SandboxType::MacosSeatbelt | SandboxType::LinuxSeccomp
)
)
}
/// Determines how a command should be sandboxed, prompting the user when
/// policy requires explicit approval.
#[allow(clippy::too_many_arguments)]
pub async fn select_sandbox(
request: &ExecutionRequest,
approval_policy: AskForApproval,
approval_cache: HashSet<Vec<String>>,
config: &ExecutorConfig,
session: &Session,
sub_id: &str,
call_id: &str,
otel_event_manager: &OtelEventManager,
) -> Result<SandboxDecision, ExecError> {
match &request.mode {
ExecutionMode::Shell => {
select_shell_sandbox(
request,
approval_policy,
approval_cache,
config,
session,
sub_id,
call_id,
otel_event_manager,
)
.await
}
ExecutionMode::ApplyPatch(exec) => {
select_apply_patch_sandbox(exec, approval_policy, config)
}
}
}
#[allow(clippy::too_many_arguments)]
async fn select_shell_sandbox(
request: &ExecutionRequest,
approval_policy: AskForApproval,
approved_snapshot: HashSet<Vec<String>>,
config: &ExecutorConfig,
session: &Session,
sub_id: &str,
call_id: &str,
otel_event_manager: &OtelEventManager,
) -> Result<SandboxDecision, ExecError> {
let command_for_safety = if request.approval_command.is_empty() {
request.params.command.clone()
} else {
request.approval_command.clone()
};
let safety = assess_command_safety(
&command_for_safety,
approval_policy,
&config.sandbox_policy,
&approved_snapshot,
request.params.with_escalated_permissions.unwrap_or(false),
);
match safety {
SafetyCheck::AutoApprove {
sandbox_type,
user_explicitly_approved,
} => {
let mut decision = SandboxDecision::auto(
sandbox_type,
should_escalate_on_failure(approval_policy, sandbox_type),
);
if user_explicitly_approved {
decision.record_session_approval = true;
}
let (decision_for_event, source) = if user_explicitly_approved {
(ReviewDecision::ApprovedForSession, ToolDecisionSource::User)
} else {
(ReviewDecision::Approved, ToolDecisionSource::Config)
};
otel_event_manager.tool_decision("local_shell", call_id, decision_for_event, source);
Ok(decision)
}
SafetyCheck::AskUser => {
let decision = session
.request_command_approval(
sub_id.to_string(),
call_id.to_string(),
request.approval_command.clone(),
request.params.cwd.clone(),
request.params.justification.clone(),
)
.await;
otel_event_manager.tool_decision(
"local_shell",
call_id,
decision,
ToolDecisionSource::User,
);
match decision {
ReviewDecision::Approved => Ok(SandboxDecision::user_override(false)),
ReviewDecision::ApprovedForSession => Ok(SandboxDecision::user_override(true)),
ReviewDecision::Denied | ReviewDecision::Abort => {
Err(ExecError::rejection("exec command rejected by user"))
}
}
}
SafetyCheck::Reject { reason } => Err(ExecError::rejection(format!(
"exec command rejected: {reason}"
))),
}
}
fn select_apply_patch_sandbox(
exec: &ApplyPatchExec,
approval_policy: AskForApproval,
config: &ExecutorConfig,
) -> Result<SandboxDecision, ExecError> {
if exec.user_explicitly_approved_this_action {
return Ok(SandboxDecision::user_override(false));
}
match assess_patch_safety(
&exec.action,
approval_policy,
&config.sandbox_policy,
&config.sandbox_cwd,
) {
SafetyCheck::AutoApprove { sandbox_type, .. } => Ok(SandboxDecision::auto(
sandbox_type,
should_escalate_on_failure(approval_policy, sandbox_type),
)),
SafetyCheck::AskUser => Err(ExecError::rejection(
"patch requires approval but none was recorded",
)),
SafetyCheck::Reject { reason } => {
Err(ExecError::rejection(format!("patch rejected: {reason}")))
}
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::codex::make_session_and_context;
use crate::exec::ExecParams;
use crate::function_tool::FunctionCallError;
use crate::protocol::SandboxPolicy;
use codex_apply_patch::ApplyPatchAction;
use pretty_assertions::assert_eq;
#[tokio::test]
async fn select_apply_patch_user_override_when_explicit() {
let (session, ctx) = make_session_and_context();
let tmp = tempfile::tempdir().expect("tmp");
let p = tmp.path().join("a.txt");
let action = ApplyPatchAction::new_add_for_test(&p, "hello".to_string());
let exec = ApplyPatchExec {
action,
user_explicitly_approved_this_action: true,
};
let cfg = ExecutorConfig::new(SandboxPolicy::ReadOnly, std::env::temp_dir(), None);
let request = ExecutionRequest {
params: ExecParams {
command: vec!["apply_patch".into()],
cwd: std::env::temp_dir(),
timeout_ms: None,
env: std::collections::HashMap::new(),
with_escalated_permissions: None,
justification: None,
},
approval_command: vec!["apply_patch".into()],
mode: ExecutionMode::ApplyPatch(exec),
stdout_stream: None,
use_shell_profile: false,
};
let otel_event_manager = ctx.client.get_otel_event_manager();
let decision = select_sandbox(
&request,
AskForApproval::OnRequest,
Default::default(),
&cfg,
&session,
"sub",
"call",
&otel_event_manager,
)
.await
.expect("ok");
// Explicit user override runs without sandbox
assert_eq!(decision.initial_sandbox, SandboxType::None);
assert_eq!(decision.escalate_on_failure, false);
}
#[tokio::test]
async fn select_apply_patch_autoapprove_in_danger() {
let (session, ctx) = make_session_and_context();
let tmp = tempfile::tempdir().expect("tmp");
let p = tmp.path().join("a.txt");
let action = ApplyPatchAction::new_add_for_test(&p, "hello".to_string());
let exec = ApplyPatchExec {
action,
user_explicitly_approved_this_action: false,
};
let cfg = ExecutorConfig::new(SandboxPolicy::DangerFullAccess, std::env::temp_dir(), None);
let request = ExecutionRequest {
params: ExecParams {
command: vec!["apply_patch".into()],
cwd: std::env::temp_dir(),
timeout_ms: None,
env: std::collections::HashMap::new(),
with_escalated_permissions: None,
justification: None,
},
approval_command: vec!["apply_patch".into()],
mode: ExecutionMode::ApplyPatch(exec),
stdout_stream: None,
use_shell_profile: false,
};
let otel_event_manager = ctx.client.get_otel_event_manager();
let decision = select_sandbox(
&request,
AskForApproval::OnRequest,
Default::default(),
&cfg,
&session,
"sub",
"call",
&otel_event_manager,
)
.await
.expect("ok");
// On platforms with a sandbox, DangerFullAccess still prefers it
let expected = crate::safety::get_platform_sandbox().unwrap_or(SandboxType::None);
assert_eq!(decision.initial_sandbox, expected);
assert_eq!(decision.escalate_on_failure, false);
}
#[tokio::test]
async fn select_apply_patch_requires_approval_on_unless_trusted() {
let (session, ctx) = make_session_and_context();
let tempdir = tempfile::tempdir().expect("tmpdir");
let p = tempdir.path().join("a.txt");
let action = ApplyPatchAction::new_add_for_test(&p, "hello".to_string());
let exec = ApplyPatchExec {
action,
user_explicitly_approved_this_action: false,
};
let cfg = ExecutorConfig::new(SandboxPolicy::ReadOnly, std::env::temp_dir(), None);
let request = ExecutionRequest {
params: ExecParams {
command: vec!["apply_patch".into()],
cwd: std::env::temp_dir(),
timeout_ms: None,
env: std::collections::HashMap::new(),
with_escalated_permissions: None,
justification: None,
},
approval_command: vec!["apply_patch".into()],
mode: ExecutionMode::ApplyPatch(exec),
stdout_stream: None,
use_shell_profile: false,
};
let otel_event_manager = ctx.client.get_otel_event_manager();
let result = select_sandbox(
&request,
AskForApproval::UnlessTrusted,
Default::default(),
&cfg,
&session,
"sub",
"call",
&otel_event_manager,
)
.await;
match result {
Ok(_) => panic!("expected error"),
Err(ExecError::Function(FunctionCallError::RespondToModel(msg))) => {
assert!(msg.contains("requires approval"))
}
Err(other) => panic!("unexpected error: {other:?}"),
}
}
#[tokio::test]
async fn select_shell_autoapprove_in_danger_mode() {
let (session, ctx) = make_session_and_context();
let cfg = ExecutorConfig::new(SandboxPolicy::DangerFullAccess, std::env::temp_dir(), None);
let request = ExecutionRequest {
params: ExecParams {
command: vec!["some-unknown".into()],
cwd: std::env::temp_dir(),
timeout_ms: None,
env: std::collections::HashMap::new(),
with_escalated_permissions: None,
justification: None,
},
approval_command: vec!["some-unknown".into()],
mode: ExecutionMode::Shell,
stdout_stream: None,
use_shell_profile: false,
};
let otel_event_manager = ctx.client.get_otel_event_manager();
let decision = select_sandbox(
&request,
AskForApproval::OnRequest,
Default::default(),
&cfg,
&session,
"sub",
"call",
&otel_event_manager,
)
.await
.expect("ok");
assert_eq!(decision.initial_sandbox, SandboxType::None);
assert_eq!(decision.escalate_on_failure, false);
}
#[cfg(any(target_os = "macos", target_os = "linux"))]
#[tokio::test]
async fn select_shell_escalates_on_failure_with_platform_sandbox() {
let (session, ctx) = make_session_and_context();
let cfg = ExecutorConfig::new(SandboxPolicy::ReadOnly, std::env::temp_dir(), None);
let request = ExecutionRequest {
params: ExecParams {
// Unknown command => untrusted but not flagged dangerous
command: vec!["some-unknown".into()],
cwd: std::env::temp_dir(),
timeout_ms: None,
env: std::collections::HashMap::new(),
with_escalated_permissions: None,
justification: None,
},
approval_command: vec!["some-unknown".into()],
mode: ExecutionMode::Shell,
stdout_stream: None,
use_shell_profile: false,
};
let otel_event_manager = ctx.client.get_otel_event_manager();
let decision = select_sandbox(
&request,
AskForApproval::OnFailure,
Default::default(),
&cfg,
&session,
"sub",
"call",
&otel_event_manager,
)
.await
.expect("ok");
// On macOS/Linux we should have a platform sandbox and escalate on failure
assert_ne!(decision.initial_sandbox, SandboxType::None);
assert_eq!(decision.escalate_on_failure, true);
}
}

View File

@@ -2,7 +2,7 @@ use std::collections::HashSet;
use std::path::Path;
use std::path::PathBuf;
use codex_app_server_protocol::GitSha;
use codex_protocol::mcp_protocol::GitSha;
use codex_protocol::protocol::GitInfo;
use futures::future::join_all;
use serde::Deserialize;

View File

@@ -27,7 +27,6 @@ pub mod error;
pub mod exec;
mod exec_command;
pub mod exec_env;
pub mod executor;
mod flags;
pub mod git_info;
pub mod landlock;
@@ -104,5 +103,3 @@ pub use codex_protocol::models::LocalShellExecAction;
pub use codex_protocol::models::LocalShellStatus;
pub use codex_protocol::models::ReasoningItemContent;
pub use codex_protocol::models::ResponseItem;
pub mod otel_init;

View File

@@ -196,7 +196,8 @@ impl McpConnectionManager {
// Validate server name before spawning
if !is_valid_mcp_server_name(&server_name) {
let error = anyhow::anyhow!(
"invalid server name '{server_name}': must match pattern ^[a-zA-Z0-9_-]+$"
"invalid server name '{}': must match pattern ^[a-zA-Z0-9_-]+$",
server_name
);
errors.insert(server_name, error);
continue;

View File

@@ -30,7 +30,7 @@ use tokio::io::AsyncReadExt;
use crate::config::Config;
use crate::config_types::HistoryPersistence;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::ConversationId;
#[cfg(unix)]
use std::os::unix::fs::OpenOptionsExt;
#[cfg(unix)]

View File

@@ -6,7 +6,7 @@
//! key. These override or extend the defaults at runtime.
use crate::CodexAuth;
use codex_app_server_protocol::AuthMode;
use codex_protocol::mcp_protocol::AuthMode;
use serde::Deserialize;
use serde::Serialize;
use std::collections::HashMap;

View File

@@ -1,61 +0,0 @@
use crate::config::Config;
use crate::config_types::OtelExporterKind as Kind;
use crate::config_types::OtelHttpProtocol as Protocol;
use crate::default_client::originator;
use codex_otel::config::OtelExporter;
use codex_otel::config::OtelHttpProtocol;
use codex_otel::config::OtelSettings;
use codex_otel::otel_provider::OtelProvider;
use std::error::Error;
/// Build an OpenTelemetry provider from the app Config.
///
/// Returns `None` when OTEL export is disabled.
pub fn build_provider(
config: &Config,
service_version: &str,
) -> Result<Option<OtelProvider>, Box<dyn Error>> {
let exporter = match &config.otel.exporter {
Kind::None => OtelExporter::None,
Kind::OtlpHttp {
endpoint,
headers,
protocol,
} => {
let protocol = match protocol {
Protocol::Json => OtelHttpProtocol::Json,
Protocol::Binary => OtelHttpProtocol::Binary,
};
OtelExporter::OtlpHttp {
endpoint: endpoint.clone(),
headers: headers
.iter()
.map(|(k, v)| (k.clone(), v.clone()))
.collect(),
protocol,
}
}
Kind::OtlpGrpc { endpoint, headers } => OtelExporter::OtlpGrpc {
endpoint: endpoint.clone(),
headers: headers
.iter()
.map(|(k, v)| (k.clone(), v.clone()))
.collect(),
},
};
OtelProvider::from(&OtelSettings {
service_name: originator().value.to_owned(),
service_version: service_version.to_string(),
codex_home: config.codex_home.clone(),
environment: config.otel.environment.to_string(),
exporter,
})
}
/// Filter predicate for exporting only Codex-owned events via OTEL.
/// Keeps events that originated from codex_otel module
pub fn codex_export_filter(meta: &tracing::Metadata<'_>) -> bool {
meta.target().starts_with("codex_otel")
}

View File

@@ -1,7 +1,6 @@
//! Project-level documentation discovery.
//!
//! Project-level documentation is primarily stored in files named `AGENTS.md`.
//! Additional fallback filenames can be configured via `project_doc_fallback_filenames`.
//! Project-level documentation can be stored in files named `AGENTS.md`.
//! We include the concatenation of all files found along the path from the
//! repository root to the current working directory as follows:
//!
@@ -14,13 +13,12 @@
//! 3. We do **not** walk past the Git root.
use crate::config::Config;
use dunce::canonicalize as normalize_path;
use std::path::PathBuf;
use tokio::io::AsyncReadExt;
use tracing::error;
/// Default filename scanned for project-level docs.
pub const DEFAULT_PROJECT_DOC_FILENAME: &str = "AGENTS.md";
/// Currently, we only match the filename `AGENTS.md` exactly.
const CANDIDATE_FILENAMES: &[&str] = &["AGENTS.md"];
/// When both `Config::instructions` and the project doc are present, they will
/// be concatenated with the following separator.
@@ -110,7 +108,7 @@ pub async fn read_project_docs(config: &Config) -> std::io::Result<Option<String
/// is zero, returns an empty list.
pub fn discover_project_doc_paths(config: &Config) -> std::io::Result<Vec<PathBuf>> {
let mut dir = config.cwd.clone();
if let Ok(canon) = normalize_path(&dir) {
if let Ok(canon) = dir.canonicalize() {
dir = canon;
}
@@ -154,9 +152,8 @@ pub fn discover_project_doc_paths(config: &Config) -> std::io::Result<Vec<PathBu
};
let mut found: Vec<PathBuf> = Vec::new();
let candidate_filenames = candidate_filenames(config);
for d in search_dirs {
for name in &candidate_filenames {
for name in CANDIDATE_FILENAMES {
let candidate = d.join(name);
match std::fs::symlink_metadata(&candidate) {
Ok(md) => {
@@ -176,22 +173,6 @@ pub fn discover_project_doc_paths(config: &Config) -> std::io::Result<Vec<PathBu
Ok(found)
}
fn candidate_filenames<'a>(config: &'a Config) -> Vec<&'a str> {
let mut names: Vec<&'a str> =
Vec::with_capacity(1 + config.project_doc_fallback_filenames.len());
names.push(DEFAULT_PROJECT_DOC_FILENAME);
for candidate in &config.project_doc_fallback_filenames {
let candidate = candidate.as_str();
if candidate.is_empty() {
continue;
}
if !names.contains(&candidate) {
names.push(candidate);
}
}
names
}
#[cfg(test)]
mod tests {
use super::*;
@@ -221,20 +202,6 @@ mod tests {
config
}
fn make_config_with_fallback(
root: &TempDir,
limit: usize,
instructions: Option<&str>,
fallbacks: &[&str],
) -> Config {
let mut config = make_config(root, limit, instructions);
config.project_doc_fallback_filenames = fallbacks
.iter()
.map(std::string::ToString::to_string)
.collect();
config
}
/// AGENTS.md missing should yield `None`.
#[tokio::test]
async fn no_doc_file_returns_none() {
@@ -380,45 +347,4 @@ mod tests {
let res = get_user_instructions(&cfg).await.expect("doc expected");
assert_eq!(res, "root doc\n\ncrate doc");
}
/// When AGENTS.md is absent but a configured fallback exists, the fallback is used.
#[tokio::test]
async fn uses_configured_fallback_when_agents_missing() {
let tmp = tempfile::tempdir().expect("tempdir");
fs::write(tmp.path().join("EXAMPLE.md"), "example instructions").unwrap();
let cfg = make_config_with_fallback(&tmp, 4096, None, &["EXAMPLE.md"]);
let res = get_user_instructions(&cfg)
.await
.expect("fallback doc expected");
assert_eq!(res, "example instructions");
}
/// AGENTS.md remains preferred when both AGENTS.md and fallbacks are present.
#[tokio::test]
async fn agents_md_preferred_over_fallbacks() {
let tmp = tempfile::tempdir().expect("tempdir");
fs::write(tmp.path().join("AGENTS.md"), "primary").unwrap();
fs::write(tmp.path().join("EXAMPLE.md"), "secondary").unwrap();
let cfg = make_config_with_fallback(&tmp, 4096, None, &["EXAMPLE.md", ".example.md"]);
let res = get_user_instructions(&cfg)
.await
.expect("AGENTS.md should win");
assert_eq!(res, "primary");
let discovery = discover_project_doc_paths(&cfg).expect("discover paths");
assert_eq!(discovery.len(), 1);
assert!(
discovery[0]
.file_name()
.unwrap()
.to_string_lossy()
.eq(DEFAULT_PROJECT_DOC_FILENAME)
);
}
}

View File

@@ -36,30 +36,13 @@ pub struct ConversationsPage {
pub struct ConversationItem {
/// Absolute path to the rollout file.
pub path: PathBuf,
/// First up to `HEAD_RECORD_LIMIT` JSONL records parsed as JSON (includes meta line).
/// First up to 5 JSONL records parsed as JSON (includes meta line).
pub head: Vec<serde_json::Value>,
/// Last up to `TAIL_RECORD_LIMIT` JSONL response records parsed as JSON.
pub tail: Vec<serde_json::Value>,
/// RFC3339 timestamp string for when the session was created, if available.
pub created_at: Option<String>,
/// RFC3339 timestamp string for the most recent response in the tail, if available.
pub updated_at: Option<String>,
}
#[derive(Default)]
struct HeadTailSummary {
head: Vec<serde_json::Value>,
tail: Vec<serde_json::Value>,
saw_session_meta: bool,
saw_user_event: bool,
created_at: Option<String>,
updated_at: Option<String>,
}
/// Hard cap to bound worstcase work per request.
const MAX_SCAN_FILES: usize = 10000;
const MAX_SCAN_FILES: usize = 100;
const HEAD_RECORD_LIMIT: usize = 10;
const TAIL_RECORD_LIMIT: usize = 10;
/// Pagination cursor identifying a file by timestamp and UUID.
#[derive(Debug, Clone, PartialEq, Eq)]
@@ -193,26 +176,13 @@ async fn traverse_directories_for_paths(
}
// Read head and simultaneously detect message events within the same
// first N JSONL records to avoid a second file read.
let summary = read_head_and_tail(&path, HEAD_RECORD_LIMIT, TAIL_RECORD_LIMIT)
.await
.unwrap_or_default();
let (head, saw_session_meta, saw_user_event) =
read_head_and_flags(&path, HEAD_RECORD_LIMIT)
.await
.unwrap_or((Vec::new(), false, false));
// Apply filters: must have session meta and at least one user message event
if summary.saw_session_meta && summary.saw_user_event {
let HeadTailSummary {
head,
tail,
created_at,
mut updated_at,
..
} = summary;
updated_at = updated_at.or_else(|| created_at.clone());
items.push(ConversationItem {
path,
head,
tail,
created_at,
updated_at,
});
if saw_session_meta && saw_user_event {
items.push(ConversationItem { path, head });
}
}
}
@@ -316,19 +286,20 @@ fn parse_timestamp_uuid_from_filename(name: &str) -> Option<(OffsetDateTime, Uui
Some((ts, uuid))
}
async fn read_head_and_tail(
async fn read_head_and_flags(
path: &Path,
head_limit: usize,
tail_limit: usize,
) -> io::Result<HeadTailSummary> {
max_records: usize,
) -> io::Result<(Vec<serde_json::Value>, bool, bool)> {
use tokio::io::AsyncBufReadExt;
let file = tokio::fs::File::open(path).await?;
let reader = tokio::io::BufReader::new(file);
let mut lines = reader.lines();
let mut summary = HeadTailSummary::default();
let mut head: Vec<serde_json::Value> = Vec::new();
let mut saw_session_meta = false;
let mut saw_user_event = false;
while summary.head.len() < head_limit {
while head.len() < max_records {
let line_opt = lines.next_line().await?;
let Some(line) = line_opt else { break };
let trimmed = line.trim();
@@ -341,22 +312,14 @@ async fn read_head_and_tail(
match rollout_line.item {
RolloutItem::SessionMeta(session_meta_line) => {
summary.created_at = summary
.created_at
.clone()
.or_else(|| Some(rollout_line.timestamp.clone()));
if let Ok(val) = serde_json::to_value(session_meta_line) {
summary.head.push(val);
summary.saw_session_meta = true;
head.push(val);
saw_session_meta = true;
}
}
RolloutItem::ResponseItem(item) => {
summary.created_at = summary
.created_at
.clone()
.or_else(|| Some(rollout_line.timestamp.clone()));
if let Ok(val) = serde_json::to_value(item) {
summary.head.push(val);
head.push(val);
}
}
RolloutItem::TurnContext(_) => {
@@ -367,104 +330,13 @@ async fn read_head_and_tail(
}
RolloutItem::EventMsg(ev) => {
if matches!(ev, EventMsg::UserMessage(_)) {
summary.saw_user_event = true;
saw_user_event = true;
}
}
}
}
if tail_limit != 0 {
let (tail, updated_at) = read_tail_records(path, tail_limit).await?;
summary.tail = tail;
summary.updated_at = updated_at;
}
Ok(summary)
}
async fn read_tail_records(
path: &Path,
max_records: usize,
) -> io::Result<(Vec<serde_json::Value>, Option<String>)> {
use std::io::SeekFrom;
use tokio::io::AsyncReadExt;
use tokio::io::AsyncSeekExt;
if max_records == 0 {
return Ok((Vec::new(), None));
}
const CHUNK_SIZE: usize = 8192;
let mut file = tokio::fs::File::open(path).await?;
let mut pos = file.seek(SeekFrom::End(0)).await?;
if pos == 0 {
return Ok((Vec::new(), None));
}
let mut buffer: Vec<u8> = Vec::new();
let mut latest_timestamp: Option<String> = None;
loop {
let slice_start = match (pos > 0, buffer.iter().position(|&b| b == b'\n')) {
(true, Some(idx)) => idx + 1,
_ => 0,
};
let (tail, newest_ts) = collect_last_response_values(&buffer[slice_start..], max_records);
if latest_timestamp.is_none() {
latest_timestamp = newest_ts.clone();
}
if tail.len() >= max_records || pos == 0 {
return Ok((tail, latest_timestamp.or(newest_ts)));
}
let read_size = CHUNK_SIZE.min(pos as usize);
if read_size == 0 {
return Ok((tail, latest_timestamp.or(newest_ts)));
}
pos -= read_size as u64;
file.seek(SeekFrom::Start(pos)).await?;
let mut chunk = vec![0; read_size];
file.read_exact(&mut chunk).await?;
chunk.extend_from_slice(&buffer);
buffer = chunk;
}
}
fn collect_last_response_values(
buffer: &[u8],
max_records: usize,
) -> (Vec<serde_json::Value>, Option<String>) {
use std::borrow::Cow;
if buffer.is_empty() || max_records == 0 {
return (Vec::new(), None);
}
let text: Cow<'_, str> = String::from_utf8_lossy(buffer);
let mut collected_rev: Vec<serde_json::Value> = Vec::new();
let mut latest_timestamp: Option<String> = None;
for line in text.lines().rev() {
let trimmed = line.trim();
if trimmed.is_empty() {
continue;
}
let parsed: serde_json::Result<RolloutLine> = serde_json::from_str(trimmed);
let Ok(rollout_line) = parsed else { continue };
let RolloutLine { timestamp, item } = rollout_line;
if let RolloutItem::ResponseItem(item) = item
&& let Ok(val) = serde_json::to_value(&item)
{
if latest_timestamp.is_none() {
latest_timestamp = Some(timestamp.clone());
}
collected_rev.push(val);
if collected_rev.len() == max_records {
break;
}
}
}
collected_rev.reverse();
(collected_rev, latest_timestamp)
Ok((head, saw_session_meta, saw_user_event))
}
/// Locate a recorded conversation rollout file by its UUID string using the existing

View File

@@ -6,7 +6,7 @@ use std::io::Error as IoError;
use std::path::Path;
use std::path::PathBuf;
use codex_protocol::ConversationId;
use codex_protocol::mcp_protocol::ConversationId;
use serde_json::Value;
use time::OffsetDateTime;
use time::format_description::FormatItem;
@@ -24,7 +24,7 @@ use super::list::Cursor;
use super::list::get_conversations;
use super::policy::is_persisted_response_item;
use crate::config::Config;
use crate::default_client::originator;
use crate::default_client::ORIGINATOR;
use crate::git_info::collect_git_info;
use codex_protocol::protocol::InitialHistory;
use codex_protocol::protocol::ResumedHistory;
@@ -124,7 +124,7 @@ impl RolloutRecorder {
id: session_id,
timestamp,
cwd: config.cwd.clone(),
originator: originator().value.clone(),
originator: ORIGINATOR.value.clone(),
cli_version: env!("CARGO_PKG_VERSION").to_string(),
instructions,
}),

View File

@@ -17,18 +17,6 @@ use crate::rollout::list::ConversationsPage;
use crate::rollout::list::Cursor;
use crate::rollout::list::get_conversation;
use crate::rollout::list::get_conversations;
use anyhow::Result;
use codex_protocol::ConversationId;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ResponseItem;
use codex_protocol::protocol::CompactedItem;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::InputMessageKind;
use codex_protocol::protocol::RolloutItem;
use codex_protocol::protocol::RolloutLine;
use codex_protocol::protocol::SessionMeta;
use codex_protocol::protocol::SessionMetaLine;
use codex_protocol::protocol::UserMessageEvent;
fn write_session_file(
root: &Path,
@@ -158,23 +146,14 @@ async fn test_list_conversations_latest_first() {
ConversationItem {
path: p1,
head: head_3,
tail: Vec::new(),
created_at: Some("2025-01-03T12-00-00".into()),
updated_at: Some("2025-01-03T12-00-00".into()),
},
ConversationItem {
path: p2,
head: head_2,
tail: Vec::new(),
created_at: Some("2025-01-02T12-00-00".into()),
updated_at: Some("2025-01-02T12-00-00".into()),
},
ConversationItem {
path: p3,
head: head_1,
tail: Vec::new(),
created_at: Some("2025-01-01T12-00-00".into()),
updated_at: Some("2025-01-01T12-00-00".into()),
},
],
next_cursor: Some(expected_cursor),
@@ -240,16 +219,10 @@ async fn test_pagination_cursor() {
ConversationItem {
path: p5,
head: head_5,
tail: Vec::new(),
created_at: Some("2025-03-05T09-00-00".into()),
updated_at: Some("2025-03-05T09-00-00".into()),
},
ConversationItem {
path: p4,
head: head_4,
tail: Vec::new(),
created_at: Some("2025-03-04T09-00-00".into()),
updated_at: Some("2025-03-04T09-00-00".into()),
},
],
next_cursor: Some(expected_cursor1.clone()),
@@ -296,16 +269,10 @@ async fn test_pagination_cursor() {
ConversationItem {
path: p3,
head: head_3,
tail: Vec::new(),
created_at: Some("2025-03-03T09-00-00".into()),
updated_at: Some("2025-03-03T09-00-00".into()),
},
ConversationItem {
path: p2,
head: head_2,
tail: Vec::new(),
created_at: Some("2025-03-02T09-00-00".into()),
updated_at: Some("2025-03-02T09-00-00".into()),
},
],
next_cursor: Some(expected_cursor2.clone()),
@@ -337,9 +304,6 @@ async fn test_pagination_cursor() {
items: vec![ConversationItem {
path: p1,
head: head_1,
tail: Vec::new(),
created_at: Some("2025-03-01T09-00-00".into()),
updated_at: Some("2025-03-01T09-00-00".into()),
}],
next_cursor: Some(expected_cursor3),
num_scanned_files: 5, // scanned 05, 04 (anchor), 03, 02 (anchor), 01
@@ -382,9 +346,6 @@ async fn test_get_conversation_contents() {
items: vec![ConversationItem {
path: expected_path,
head: expected_head,
tail: Vec::new(),
created_at: Some(ts.into()),
updated_at: Some(ts.into()),
}],
next_cursor: Some(expected_cursor),
num_scanned_files: 1,
@@ -405,269 +366,6 @@ async fn test_get_conversation_contents() {
assert_eq!(content, expected_content);
}
#[tokio::test]
async fn test_tail_includes_last_response_items() -> Result<()> {
let temp = TempDir::new().unwrap();
let home = temp.path();
let ts = "2025-06-01T08-00-00";
let uuid = Uuid::from_u128(42);
let day_dir = home.join("sessions").join("2025").join("06").join("01");
fs::create_dir_all(&day_dir)?;
let file_path = day_dir.join(format!("rollout-{ts}-{uuid}.jsonl"));
let mut file = File::create(&file_path)?;
let conversation_id = ConversationId::from_string(&uuid.to_string())?;
let meta_line = RolloutLine {
timestamp: ts.to_string(),
item: RolloutItem::SessionMeta(SessionMetaLine {
meta: SessionMeta {
id: conversation_id,
timestamp: ts.to_string(),
instructions: None,
cwd: ".".into(),
originator: "test_originator".into(),
cli_version: "test_version".into(),
},
git: None,
}),
};
writeln!(file, "{}", serde_json::to_string(&meta_line)?)?;
let user_event_line = RolloutLine {
timestamp: ts.to_string(),
item: RolloutItem::EventMsg(EventMsg::UserMessage(UserMessageEvent {
message: "hello".into(),
kind: Some(InputMessageKind::Plain),
images: None,
})),
};
writeln!(file, "{}", serde_json::to_string(&user_event_line)?)?;
let total_messages = 12usize;
for idx in 0..total_messages {
let response_line = RolloutLine {
timestamp: format!("{ts}-{idx:02}"),
item: RolloutItem::ResponseItem(ResponseItem::Message {
id: None,
role: "assistant".into(),
content: vec![ContentItem::OutputText {
text: format!("reply-{idx}"),
}],
}),
};
writeln!(file, "{}", serde_json::to_string(&response_line)?)?;
}
drop(file);
let page = get_conversations(home, 1, None).await?;
let item = page.items.first().expect("conversation item");
let tail_len = item.tail.len();
assert_eq!(tail_len, 10usize.min(total_messages));
let expected: Vec<serde_json::Value> = (total_messages - tail_len..total_messages)
.map(|idx| {
serde_json::json!({
"type": "message",
"role": "assistant",
"content": [
{
"type": "output_text",
"text": format!("reply-{idx}"),
}
],
})
})
.collect();
assert_eq!(item.tail, expected);
assert_eq!(item.created_at.as_deref(), Some(ts));
let expected_updated = format!("{ts}-{last:02}", last = total_messages - 1);
assert_eq!(item.updated_at.as_deref(), Some(expected_updated.as_str()));
Ok(())
}
#[tokio::test]
async fn test_tail_handles_short_sessions() -> Result<()> {
let temp = TempDir::new().unwrap();
let home = temp.path();
let ts = "2025-06-02T08-30-00";
let uuid = Uuid::from_u128(7);
let day_dir = home.join("sessions").join("2025").join("06").join("02");
fs::create_dir_all(&day_dir)?;
let file_path = day_dir.join(format!("rollout-{ts}-{uuid}.jsonl"));
let mut file = File::create(&file_path)?;
let conversation_id = ConversationId::from_string(&uuid.to_string())?;
let meta_line = RolloutLine {
timestamp: ts.to_string(),
item: RolloutItem::SessionMeta(SessionMetaLine {
meta: SessionMeta {
id: conversation_id,
timestamp: ts.to_string(),
instructions: None,
cwd: ".".into(),
originator: "test_originator".into(),
cli_version: "test_version".into(),
},
git: None,
}),
};
writeln!(file, "{}", serde_json::to_string(&meta_line)?)?;
let user_event_line = RolloutLine {
timestamp: ts.to_string(),
item: RolloutItem::EventMsg(EventMsg::UserMessage(UserMessageEvent {
message: "hi".into(),
kind: Some(InputMessageKind::Plain),
images: None,
})),
};
writeln!(file, "{}", serde_json::to_string(&user_event_line)?)?;
for idx in 0..3 {
let response_line = RolloutLine {
timestamp: format!("{ts}-{idx:02}"),
item: RolloutItem::ResponseItem(ResponseItem::Message {
id: None,
role: "assistant".into(),
content: vec![ContentItem::OutputText {
text: format!("short-{idx}"),
}],
}),
};
writeln!(file, "{}", serde_json::to_string(&response_line)?)?;
}
drop(file);
let page = get_conversations(home, 1, None).await?;
let tail = &page.items.first().expect("conversation item").tail;
assert_eq!(tail.len(), 3);
let expected: Vec<serde_json::Value> = (0..3)
.map(|idx| {
serde_json::json!({
"type": "message",
"role": "assistant",
"content": [
{
"type": "output_text",
"text": format!("short-{idx}"),
}
],
})
})
.collect();
assert_eq!(tail, &expected);
let expected_updated = format!("{ts}-{last:02}", last = 2);
assert_eq!(
page.items[0].updated_at.as_deref(),
Some(expected_updated.as_str())
);
Ok(())
}
#[tokio::test]
async fn test_tail_skips_trailing_non_responses() -> Result<()> {
let temp = TempDir::new().unwrap();
let home = temp.path();
let ts = "2025-06-03T10-00-00";
let uuid = Uuid::from_u128(11);
let day_dir = home.join("sessions").join("2025").join("06").join("03");
fs::create_dir_all(&day_dir)?;
let file_path = day_dir.join(format!("rollout-{ts}-{uuid}.jsonl"));
let mut file = File::create(&file_path)?;
let conversation_id = ConversationId::from_string(&uuid.to_string())?;
let meta_line = RolloutLine {
timestamp: ts.to_string(),
item: RolloutItem::SessionMeta(SessionMetaLine {
meta: SessionMeta {
id: conversation_id,
timestamp: ts.to_string(),
instructions: None,
cwd: ".".into(),
originator: "test_originator".into(),
cli_version: "test_version".into(),
},
git: None,
}),
};
writeln!(file, "{}", serde_json::to_string(&meta_line)?)?;
let user_event_line = RolloutLine {
timestamp: ts.to_string(),
item: RolloutItem::EventMsg(EventMsg::UserMessage(UserMessageEvent {
message: "hello".into(),
kind: Some(InputMessageKind::Plain),
images: None,
})),
};
writeln!(file, "{}", serde_json::to_string(&user_event_line)?)?;
for idx in 0..4 {
let response_line = RolloutLine {
timestamp: format!("{ts}-{idx:02}"),
item: RolloutItem::ResponseItem(ResponseItem::Message {
id: None,
role: "assistant".into(),
content: vec![ContentItem::OutputText {
text: format!("response-{idx}"),
}],
}),
};
writeln!(file, "{}", serde_json::to_string(&response_line)?)?;
}
let compacted_line = RolloutLine {
timestamp: format!("{ts}-compacted"),
item: RolloutItem::Compacted(CompactedItem {
message: "compacted".into(),
}),
};
writeln!(file, "{}", serde_json::to_string(&compacted_line)?)?;
let shutdown_event = RolloutLine {
timestamp: format!("{ts}-shutdown"),
item: RolloutItem::EventMsg(EventMsg::ShutdownComplete),
};
writeln!(file, "{}", serde_json::to_string(&shutdown_event)?)?;
drop(file);
let page = get_conversations(home, 1, None).await?;
let tail = &page.items.first().expect("conversation item").tail;
let expected: Vec<serde_json::Value> = (0..4)
.map(|idx| {
serde_json::json!({
"type": "message",
"role": "assistant",
"content": [
{
"type": "output_text",
"text": format!("response-{idx}"),
}
],
})
})
.collect();
assert_eq!(tail, &expected);
let expected_updated = format!("{ts}-{last:02}", last = 3);
assert_eq!(
page.items[0].updated_at.as_deref(),
Some(expected_updated.as_str())
);
Ok(())
}
#[tokio::test]
async fn test_stable_ordering_same_second_pagination() {
let temp = TempDir::new().unwrap();
@@ -712,16 +410,10 @@ async fn test_stable_ordering_same_second_pagination() {
ConversationItem {
path: p3,
head: head(u3),
tail: Vec::new(),
created_at: Some(ts.to_string()),
updated_at: Some(ts.to_string()),
},
ConversationItem {
path: p2,
head: head(u2),
tail: Vec::new(),
created_at: Some(ts.to_string()),
updated_at: Some(ts.to_string()),
},
],
next_cursor: Some(expected_cursor1.clone()),
@@ -744,9 +436,6 @@ async fn test_stable_ordering_same_second_pagination() {
items: vec![ConversationItem {
path: p1,
head: head(u1),
tail: Vec::new(),
created_at: Some(ts.to_string()),
updated_at: Some(ts.to_string()),
}],
next_cursor: Some(expected_cursor2),
num_scanned_files: 3, // scanned u3, u2 (anchor), u1

View File

@@ -15,14 +15,9 @@ use crate::protocol::SandboxPolicy;
#[derive(Debug, PartialEq)]
pub enum SafetyCheck {
AutoApprove {
sandbox_type: SandboxType,
user_explicitly_approved: bool,
},
AutoApprove { sandbox_type: SandboxType },
AskUser,
Reject {
reason: String,
},
Reject { reason: String },
}
pub fn assess_patch_safety(
@@ -59,16 +54,12 @@ pub fn assess_patch_safety(
// fall back to asking the user because the patch may touch arbitrary
// paths outside the project.
match get_platform_sandbox() {
Some(sandbox_type) => SafetyCheck::AutoApprove {
sandbox_type,
user_explicitly_approved: false,
},
Some(sandbox_type) => SafetyCheck::AutoApprove { sandbox_type },
None if sandbox_policy == &SandboxPolicy::DangerFullAccess => {
// If the user has explicitly requested DangerFullAccess, then
// we can auto-approve even without a sandbox.
SafetyCheck::AutoApprove {
sandbox_type: SandboxType::None,
user_explicitly_approved: false,
}
}
None => SafetyCheck::AskUser,
@@ -125,10 +116,8 @@ pub fn assess_command_safety(
// the session _because_ they know it needs to run outside a sandbox.
if is_known_safe_command(command) || approved.contains(command) {
let user_explicitly_approved = approved.contains(command);
return SafetyCheck::AutoApprove {
sandbox_type: SandboxType::None,
user_explicitly_approved,
};
}
@@ -154,17 +143,13 @@ pub(crate) fn assess_safety_for_untrusted_command(
| (Never, DangerFullAccess)
| (OnRequest, DangerFullAccess) => SafetyCheck::AutoApprove {
sandbox_type: SandboxType::None,
user_explicitly_approved: false,
},
(OnRequest, ReadOnly) | (OnRequest, WorkspaceWrite { .. }) => {
if with_escalated_permissions {
SafetyCheck::AskUser
} else {
match get_platform_sandbox() {
Some(sandbox_type) => SafetyCheck::AutoApprove {
sandbox_type,
user_explicitly_approved: false,
},
Some(sandbox_type) => SafetyCheck::AutoApprove { sandbox_type },
// Fall back to asking since the command is untrusted and
// we do not have a sandbox available
None => SafetyCheck::AskUser,
@@ -176,10 +161,7 @@ pub(crate) fn assess_safety_for_untrusted_command(
| (OnFailure, ReadOnly)
| (OnFailure, WorkspaceWrite { .. }) => {
match get_platform_sandbox() {
Some(sandbox_type) => SafetyCheck::AutoApprove {
sandbox_type,
user_explicitly_approved: false,
},
Some(sandbox_type) => SafetyCheck::AutoApprove { sandbox_type },
None => {
if matches!(approval_policy, OnFailure) {
// Since the command is not trusted, even though the
@@ -380,8 +362,7 @@ mod tests {
assert_eq!(
safety_check,
SafetyCheck::AutoApprove {
sandbox_type: SandboxType::None,
user_explicitly_approved: true,
sandbox_type: SandboxType::None
}
);
}
@@ -428,10 +409,7 @@ mod tests {
);
let expected = match get_platform_sandbox() {
Some(sandbox_type) => SafetyCheck::AutoApprove {
sandbox_type,
user_explicitly_approved: false,
},
Some(sandbox_type) => SafetyCheck::AutoApprove { sandbox_type },
None => SafetyCheck::AskUser,
};
assert_eq!(safety_check, expected);

View File

@@ -1,9 +1,9 @@
use crate::RolloutRecorder;
use crate::exec_command::ExecSessionManager;
use crate::executor::Executor;
use crate::mcp_connection_manager::McpConnectionManager;
use crate::unified_exec::UnifiedExecSessionManager;
use crate::user_notification::UserNotifier;
use std::path::PathBuf;
use tokio::sync::Mutex;
pub(crate) struct SessionServices {
@@ -12,7 +12,7 @@ pub(crate) struct SessionServices {
pub(crate) unified_exec_manager: UnifiedExecSessionManager,
pub(crate) notifier: UserNotifier,
pub(crate) rollout: Mutex<Option<RolloutRecorder>>,
pub(crate) codex_linux_sandbox_exe: Option<PathBuf>,
pub(crate) user_shell: crate::shell::Shell,
pub(crate) show_raw_agent_reasoning: bool,
pub(crate) executor: Executor,
}

Some files were not shown because too many files have changed in this diff Show More