Compare commits

..

38 Commits

Author SHA1 Message Date
Ahmed Ibrahim
093406c67f feedback 2025-10-17 08:06:14 -07:00
Ahmed Ibrahim
daf762f607 tests 2025-10-16 22:23:57 -07:00
Ahmed Ibrahim
cafd08ab41 tests 2025-10-16 17:23:20 -07:00
Ahmed Ibrahim
a4e042b671 tests 2025-10-16 17:22:54 -07:00
Ahmed Ibrahim
2994b63fe5 tests 2025-10-16 17:19:36 -07:00
Ahmed Ibrahim
0cf9bd6aa7 tests 2025-10-16 17:12:14 -07:00
Ahmed Ibrahim
f2444893ca tests 2025-10-16 17:05:34 -07:00
Ahmed Ibrahim
d4e59dedd8 fix_compact 2025-10-16 16:47:42 -07:00
Ahmed Ibrahim
e0f7c32217 fix_compact 2025-10-16 16:27:15 -07:00
Ahmed Ibrahim
578a6bc9e1 fix_compact 2025-10-16 16:09:56 -07:00
joshka-oai
18d00e36b9 feat(tui): warn high effort rate use (#5035)
Highlight that selecting a high reasoning level will hit Plus plan rate
limits faster.
2025-10-15 14:57:05 -07:00
Jeremy Rose
17550fee9e add ^Y and kill-buffer to textarea (#5075)
## Summary
- add a kill buffer to the text area and wire Ctrl+Y to yank it
- capture text from Ctrl+W, Ctrl+U, and Ctrl+K operations into the kill
buffer
- add regression coverage ensuring the last kill can be yanked back

Fixes #5017


------
https://chatgpt.com/codex/tasks/task_i_68e95bf06c48832cbf3d2ba8fa2035d2
2025-10-15 14:39:55 -07:00
Michael Bolin
995f5c3614 feat: add Vec<ParsedCommand> to ExecApprovalRequestEvent (#5222)
This adds `parsed_cmd: Vec<ParsedCommand>` to `ExecApprovalRequestEvent`
in the core protocol (`protocol/src/protocol.rs`), which is also what
this field is named on `ExecCommandBeginEvent`. Honestly, I don't love
the name (it sounds like a single command, but it is actually a list of
them), but I don't want to get distracted by a naming discussion right
now.

This also adds `parsed_cmd` to `ExecCommandApprovalParams` in
`codex-rs/app-server-protocol/src/protocol.rs`, so it will be available
via `codex app-server`, as well.

For consistency, I also updated `ExecApprovalElicitRequestParams` in
`codex-rs/mcp-server/src/exec_approval.rs` to include this field under
the name `codex_parsed_cmd`, as that struct already has a number of
special `codex_*` fields. Note this is the code for when Codex is used
as an MCP _server_ and therefore has to conform to the official spec for
an MCP elicitation type.
2025-10-15 13:58:40 -07:00
Jeremy Rose
9b53a306e3 Keep backtrack Esc hint gated on empty composer (#5076)
## Summary
- only prime backtrack and show the ESC hint when the composer is empty
- keep the composer-side ESC hint unchanged when drafts or attachments
exist and cover it with a regression test

Fixes #5030

------
https://chatgpt.com/codex/tasks/task_i_68e95ba59cd8832caec8e72ae2efeb55
2025-10-15 13:57:50 -07:00
Jeremy Rose
0016346dfb tui: ^C in prompt area resets history navigation cursor (#5078)
^C resets the history navigation, similar to zsh/bash.

Fixes #4834

------
https://chatgpt.com/codex/tasks/task_i_68e9674b6ac8832c8212bff6cba75e87
2025-10-15 13:57:44 -07:00
Michael Bolin
f38ad65254 chore: standardize on ParsedCommand from codex_protocol (#5218)
Note these two types were identical, so it seems clear to standardize on the one in `codex_protocol` and eliminate the `Into` stuff.

---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/5218).
* #5222
* __->__ #5218
2025-10-15 13:00:22 -07:00
jif-oai
774892c6d7 feat: add auto-approval for codex exec (#5043) 2025-10-15 19:03:54 +01:00
jif-oai
897d4d5f17 feat: agent override file (#5215)
Add a file that overrides `AGENTS.md` but is not versioned (for local
devs)
2025-10-15 17:46:01 +01:00
Gabriel Peal
8a281cd1f4 [MCP] Prompt mcp login when adding a streamable HTTP server that supports oauth (#5193)
1. If Codex detects that a `codex mcp add -url …` server supports oauth,
it will auto-initiate the login flow.
2. If the TUI starts and a MCP server supports oauth but isn't logged
in, it will give the user an explicit warning telling them to log in.
2025-10-15 12:27:40 -04:00
Shijie Rao
e8863b233b feat: updated github issue template (#5191)
### Update github issue template for bug submission. 
* Add subscription field
* Require codex cli/extension version
* Require subscription plan
* Require error message with added context
2025-10-15 07:27:24 -07:00
jif-oai
8fed0b53c4 test: reduce time dependency on test harness (#5053)
Tightened the CLI integration tests to stop relying on wall-clock
sleeps—new fs watcher helper waits for session files instead of timing
out, and SSE mocks/fixtures make the flows deterministic.
2025-10-15 09:56:59 +01:00
Dylan
00debb6399 fix(core) use regex for all shell_serialization tests (#5189)
## Summary
Thought I switched all of these to using a regex instead, but missed 2.
This should address our [flakey test
problem](https://github.com/openai/codex/actions/runs/18511206616/job/52752341520?pr=5185).

## Test Plan
- [x] Only updated unit tests
2025-10-14 16:29:02 -07:00
Dylan
0a0a10d8b3 fix: apply_patch shell_serialization tests (#4786)
## Summary
Adds additional shell_serialization tests specifically for apply_patch
and other cases.

## Test Plan
- [x] These are all tests
2025-10-14 13:00:49 -07:00
Javi
13035561cd feat: pass codex thread ID in notifier metadata (#4582) 2025-10-14 11:55:10 -07:00
Jeremy Rose
9be704a934 tui: reserve 1 cell right margin for composer and user history (#5026)
keep a 1 cell margin at the right edge of the screen in the composer
(and in the user message in history).

this lets us print clear-to-EOL 1 char before the end of the line in
history, so that resizing the terminal will keep the background color
(at least in iterm/terminal.app). it also stops the cursor in the
textarea from floating off the right edge.

---------

Co-authored-by: joshka-oai <joshka@openai.com>
2025-10-14 18:02:11 +00:00
jif-oai
f7b4e29609 feat: feature flag (#4948)
Add proper feature flag instead of having custom flags for everything.
This is just for experimental/wip part of the code
It can be used through CLI:
```bash
codex --enable unified_exec --disable view_image_tool
```

Or in the `config.toml`
```toml
# Global toggles applied to every profile unless overridden.
[features]
apply_patch_freeform = true
view_image_tool = false
```

Follow-up:
In a following PR, the goal is to have a default have `bundles` of
features that we can associate to a model
2025-10-14 17:50:00 +00:00
Jeremy Rose
d6c5df9a0a detect Bun installs in CLI update banner (#5074)
## Summary
- detect Bun-managed installs in the JavaScript launcher and set a
dedicated environment flag
- show a Bun-specific upgrade command in the update banner when that
flag is present

Fixes #5012

------
https://chatgpt.com/codex/tasks/task_i_68e95c439494832c835bdf34b3b1774e

---------

Co-authored-by: Michael Bolin <mbolin@openai.com>
2025-10-14 17:49:44 +00:00
Jeremy Rose
8662162f45 cloud: codex cloud exec (#5060)
By analogy to `codex exec`, this kicks off a task in codex cloud
noninteractively.
2025-10-14 10:49:17 -07:00
jif-oai
57584d6f34 fix: the 7 omitted lines issue (#5141)
Before, the CLI was always showing `... +7 lines` (with the 7 constant)
due to a double truncation

<img width="263" height="127" alt="Screenshot 2025-10-13 at 10 28 11"
src="https://github.com/user-attachments/assets/49a92d2b-c28a-4e2f-96d1-1818955470b8"
/>
2025-10-14 18:15:47 +01:00
jif-oai
268a10f917 feat: add header for task kind (#5142)
Add a header in the responses API request for the task kind (compact,
review, ...) for observability purpose
The header name is `codex-task-type`
2025-10-14 15:17:00 +00:00
jif-oai
5346cc422d feat: discard prompt starting with a slash (#5048)
This is does not consider lines starting with a space or containing
multiple `/` as commands
<img width="550" height="362" alt="Screenshot 2025-10-13 at 10 00 08"
src="https://github.com/user-attachments/assets/17f7347f-db24-47cb-9845-b0eb6fb139cb"
/>
2025-10-14 09:47:20 +01:00
Shijie Rao
26f7c46856 fixes #5011: update mcp server doc (#5014) 2025-10-10 17:23:41 -07:00
Jeremy Rose
90af046c5c tui: include the image name in the textarea placeholder (#5056)
Fixes #5013
2025-10-10 09:56:18 -07:00
jif-oai
961ed31901 feat: make shortcut works even with capslock (#5049)
Shortcut where not working in caps-lock. Fixing this
2025-10-10 14:35:28 +00:00
jif-oai
85e7357973 fix: workflow cache (#5050)
Decouple cache saving to fix the `verify` steps never being run due to a
cache saving issue
2025-10-10 15:57:47 +02:00
jif-oai
f98fa85b44 feat: message when stream get correctly resumed (#4988)
<img width="366" height="109" alt="Screenshot 2025-10-09 at 17 44 16"
src="https://github.com/user-attachments/assets/26bc6f60-11bc-4fc6-a1cc-430ca1203969"
/>
2025-10-10 09:07:14 +00:00
Jeremy Rose
ddcaf3dccd tui: fix crash when alt+bksp past unicode nbsp (#5016)
notably, screenshot filenames on macOS by default contain U+202F right
before the "AM/PM" part of the filename.
2025-10-09 15:07:04 -07:00
Jeremy Rose
56296cad82 tui: /diff mode wraps long lines (#4891)
fixes a regression that stopped long lines from being wrapped when
viewing `/diff`.
2025-10-09 14:01:45 -07:00
78 changed files with 3460 additions and 677 deletions

View File

@@ -20,6 +20,14 @@ body:
attributes:
label: What version of Codex is running?
description: Copy the output of `codex --version`
validations:
required: true
- type: input
id: plan
attributes:
label: What subscription do you have?
validations:
required: true
- type: input
id: model
attributes:
@@ -32,11 +40,18 @@ body:
description: |
For MacOS and Linux: copy the output of `uname -mprs`
For Windows: copy the output of `"$([Environment]::OSVersion | ForEach-Object VersionString) $(if ([Environment]::Is64BitOperatingSystem) { "x64" } else { "x86" })"` in the PowerShell console
- type: textarea
id: actual
attributes:
label: What issue are you seeing?
description: Please include the full error messages and prompts with PII redacted. If possible, please provide text instead of a screenshot.
validations:
required: true
- type: textarea
id: steps
attributes:
label: What steps can reproduce the bug?
description: Explain the bug and provide a code snippet that can reproduce it.
description: Explain the bug and provide a code snippet that can reproduce it. Please include session id, token limit usage, context window usage if applicable.
validations:
required: true
- type: textarea
@@ -44,11 +59,6 @@ body:
attributes:
label: What is the expected behavior?
description: If possible, please provide text instead of a screenshot.
- type: textarea
id: actual
attributes:
label: What do you see instead?
description: If possible, please provide text instead of a screenshot.
- type: textarea
id: notes
attributes:

View File

@@ -14,11 +14,21 @@ body:
id: version
attributes:
label: What version of the VS Code extension are you using?
validations:
required: true
- type: input
id: plan
attributes:
label: What subscription do you have?
validations:
required: true
- type: input
id: ide
attributes:
label: Which IDE are you using?
description: Like `VS Code`, `Cursor`, `Windsurf`, etc.
validations:
required: true
- type: input
id: platform
attributes:
@@ -26,11 +36,18 @@ body:
description: |
For MacOS and Linux: copy the output of `uname -mprs`
For Windows: copy the output of `"$([Environment]::OSVersion | ForEach-Object VersionString) $(if ([Environment]::Is64BitOperatingSystem) { "x64" } else { "x86" })"` in the PowerShell console
- type: textarea
id: actual
attributes:
label: What issue are you seeing?
description: Please include the full error messages and prompts with PII redacted. If possible, please provide text instead of a screenshot.
validations:
required: true
- type: textarea
id: steps
attributes:
label: What steps can reproduce the bug?
description: Explain the bug and provide a code snippet that can reproduce it.
description: Explain the bug and provide a code snippet that can reproduce it. Please include session id, token limit usage, context window usage if applicable.
validations:
required: true
- type: textarea
@@ -38,11 +55,6 @@ body:
attributes:
label: What is the expected behavior?
description: If possible, please provide text instead of a screenshot.
- type: textarea
id: actual
attributes:
label: What do you see instead?
description: If possible, please provide text instead of a screenshot.
- type: textarea
id: notes
attributes:

View File

@@ -148,15 +148,26 @@ jobs:
targets: ${{ matrix.target }}
components: clippy
- uses: actions/cache@v4
# Explicit cache restore: split cargo home vs target, so we can
# avoid caching the large target dir on the gnu-dev job.
- name: Restore cargo home cache
id: cache_cargo_home_restore
uses: actions/cache/restore@v4
with:
path: |
~/.cargo/bin/
~/.cargo/registry/index/
~/.cargo/registry/cache/
~/.cargo/git/db/
${{ github.workspace }}/codex-rs/target/
key: cargo-${{ matrix.runner }}-${{ matrix.target }}-${{ matrix.profile }}-${{ hashFiles('**/Cargo.lock') }}
key: cargo-home-${{ matrix.runner }}-${{ matrix.target }}-${{ matrix.profile }}-${{ hashFiles('**/Cargo.lock') }}
- name: Restore target cache (except gnu-dev)
id: cache_target_restore
if: ${{ !(matrix.target == 'x86_64-unknown-linux-gnu' && matrix.profile != 'release') }}
uses: actions/cache/restore@v4
with:
path: ${{ github.workspace }}/codex-rs/target/
key: cargo-target-${{ matrix.runner }}-${{ matrix.target }}-${{ matrix.profile }}-${{ hashFiles('**/Cargo.lock') }}
- if: ${{ matrix.target == 'x86_64-unknown-linux-musl' || matrix.target == 'aarch64-unknown-linux-musl'}}
name: Install musl build tools
@@ -194,6 +205,31 @@ jobs:
env:
RUST_BACKTRACE: 1
# Save caches explicitly; make non-fatal so cache packaging
# never fails the overall job. Only save when key wasn't hit.
- name: Save cargo home cache
if: always() && !cancelled() && steps.cache_cargo_home_restore.outputs.cache-hit != 'true'
continue-on-error: true
uses: actions/cache/save@v4
with:
path: |
~/.cargo/bin/
~/.cargo/registry/index/
~/.cargo/registry/cache/
~/.cargo/git/db/
key: cargo-home-${{ matrix.runner }}-${{ matrix.target }}-${{ matrix.profile }}-${{ hashFiles('**/Cargo.lock') }}
- name: Save target cache (except gnu-dev)
if: >-
always() && !cancelled() &&
(steps.cache_target_restore.outputs.cache-hit != 'true') &&
!(matrix.target == 'x86_64-unknown-linux-gnu' && matrix.profile != 'release')
continue-on-error: true
uses: actions/cache/save@v4
with:
path: ${{ github.workspace }}/codex-rs/target/
key: cargo-target-${{ matrix.runner }}-${{ matrix.target }}-${{ matrix.profile }}-${{ hashFiles('**/Cargo.lock') }}
# Fail the job if any of the previous steps failed.
- name: verify all steps passed
if: |

1
.gitignore vendored
View File

@@ -30,6 +30,7 @@ result
# cli tools
CLAUDE.md
.claude/
AGENTS.override.md
# caches
.cache/

35
codex-cli/bin/codex.js Executable file → Normal file
View File

@@ -80,6 +80,32 @@ function getUpdatedPath(newDirs) {
return updatedPath;
}
/**
* Use heuristics to detect the package manager that was used to install Codex
* in order to give the user a hint about how to update it.
*/
function detectPackageManager() {
const userAgent = process.env.npm_config_user_agent || "";
if (/\bbun\//.test(userAgent)) {
return "bun";
}
const execPath = process.env.npm_execpath || "";
if (execPath.includes("bun")) {
return "bun";
}
if (
process.env.BUN_INSTALL ||
process.env.BUN_INSTALL_GLOBAL_DIR ||
process.env.BUN_INSTALL_BIN_DIR
) {
return "bun";
}
return userAgent ? "npm" : null;
}
const additionalDirs = [];
const pathDir = path.join(archRoot, "path");
if (existsSync(pathDir)) {
@@ -87,9 +113,16 @@ if (existsSync(pathDir)) {
}
const updatedPath = getUpdatedPath(additionalDirs);
const env = { ...process.env, PATH: updatedPath };
const packageManagerEnvVar =
detectPackageManager() === "bun"
? "CODEX_MANAGED_BY_BUN"
: "CODEX_MANAGED_BY_NPM";
env[packageManagerEnvVar] = "1";
const child = spawn(binaryPath, process.argv.slice(2), {
stdio: "inherit",
env: { ...process.env, PATH: updatedPath, CODEX_MANAGED_BY_NPM: "1" },
env,
});
child.on("error", (err) => {

75
codex-rs/Cargo.lock generated
View File

@@ -1577,10 +1577,12 @@ dependencies = [
"anyhow",
"assert_cmd",
"codex-core",
"notify",
"regex-lite",
"serde_json",
"tempfile",
"tokio",
"walkdir",
"wiremock",
]
@@ -2371,6 +2373,15 @@ dependencies = [
"percent-encoding",
]
[[package]]
name = "fsevent-sys"
version = "4.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "76ee7a02da4d231650c7cea31349b889be2f45ddb3ef3032d2ec8185f6313fd2"
dependencies = [
"libc",
]
[[package]]
name = "futures"
version = "0.3.31"
@@ -3057,6 +3068,26 @@ version = "2.0.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f4c7245a08504955605670dbf141fceab975f15ca21570696aebe9d2e71576bd"
[[package]]
name = "inotify"
version = "0.11.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f37dccff2791ab604f9babef0ba14fbe0be30bd368dc541e2b08d07c8aa908f3"
dependencies = [
"bitflags 2.9.1",
"inotify-sys",
"libc",
]
[[package]]
name = "inotify-sys"
version = "0.1.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e05c02b5e89bff3b946cedeca278abc628fe811e604f027c45a8aa3cf793d0eb"
dependencies = [
"libc",
]
[[package]]
name = "inout"
version = "0.1.4"
@@ -3257,6 +3288,26 @@ dependencies = [
"zeroize",
]
[[package]]
name = "kqueue"
version = "1.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "eac30106d7dce88daf4a3fcb4879ea939476d5074a9b7ddd0fb97fa4bed5596a"
dependencies = [
"kqueue-sys",
"libc",
]
[[package]]
name = "kqueue-sys"
version = "1.0.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ed9625ffda8729b85e45cf04090035ac368927b8cebc34898e7c120f52e4838b"
dependencies = [
"bitflags 1.3.2",
"libc",
]
[[package]]
name = "lalrpop"
version = "0.19.12"
@@ -3656,6 +3707,30 @@ version = "0.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "61807f77802ff30975e01f4f071c8ba10c022052f98b3294119f3e615d13e5be"
[[package]]
name = "notify"
version = "8.2.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4d3d07927151ff8575b7087f245456e549fea62edf0ec4e565a5ee50c8402bc3"
dependencies = [
"bitflags 2.9.1",
"fsevent-sys",
"inotify",
"kqueue",
"libc",
"log",
"mio",
"notify-types",
"walkdir",
"windows-sys 0.60.2",
]
[[package]]
name = "notify-types"
version = "2.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5e0826a989adedc2a244799e823aece04662b66609d96af8dff7ac6df9a8925d"
[[package]]
name = "nu-ansi-term"
version = "0.50.1"

View File

@@ -37,7 +37,7 @@ members = [
resolver = "2"
[workspace.package]
version = "0.47.0-alpha.1"
version = "0.0.0"
# Track the edition for all workspace crates in one place. Individual
# crates can still override this value, but keeping it here means new
# crates created with `cargo new -w ...` automatically inherit the 2024
@@ -122,6 +122,7 @@ log = "0.4"
maplit = "1.0.2"
mime_guess = "2.0.5"
multimap = "0.10.0"
notify = "8.2.0"
nucleo-matcher = "0.3.1"
openssl-sys = "*"
opentelemetry = "0.30.0"

View File

@@ -9,6 +9,7 @@ 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::parse_command::ParsedCommand;
use codex_protocol::protocol::AskForApproval;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::FileChange;
@@ -697,6 +698,7 @@ pub struct ExecCommandApprovalParams {
pub cwd: PathBuf,
#[serde(skip_serializing_if = "Option::is_none")]
pub reason: Option<String>,
pub parsed_cmd: Vec<ParsedCommand>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, TS)]
@@ -904,6 +906,9 @@ mod tests {
command: vec!["echo".to_string(), "hello".to_string()],
cwd: PathBuf::from("/tmp"),
reason: Some("because tests".to_string()),
parsed_cmd: vec![ParsedCommand::Unknown {
cmd: "echo hello".to_string(),
}],
};
let request = ServerRequest::ExecCommandApproval {
request_id: RequestId::Integer(7),
@@ -920,6 +925,12 @@ mod tests {
"command": ["echo", "hello"],
"cwd": "/tmp",
"reason": "because tests",
"parsedCmd": [
{
"type": "unknown",
"cmd": "echo hello"
}
]
}
}),
serde_json::to_value(&request)?,

View File

@@ -1284,6 +1284,7 @@ async fn apply_bespoke_event_handling(
command,
cwd,
reason,
parsed_cmd,
}) => {
let params = ExecCommandApprovalParams {
conversation_id,
@@ -1291,6 +1292,7 @@ async fn apply_bespoke_event_handling(
command,
cwd,
reason,
parsed_cmd,
};
let rx = outgoing
.send_request(ServerRequestPayload::ExecCommandApproval(params))

View File

@@ -27,6 +27,7 @@ use codex_core::protocol_config_types::ReasoningEffort;
use codex_core::protocol_config_types::ReasoningSummary;
use codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR;
use codex_protocol::config_types::SandboxMode;
use codex_protocol::parse_command::ParsedCommand;
use codex_protocol::protocol::Event;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::InputMessageKind;
@@ -311,6 +312,9 @@ async fn test_send_user_turn_changes_approval_policy_behavior() {
],
cwd: working_directory.clone(),
reason: None,
parsed_cmd: vec![ParsedCommand::Unknown {
cmd: "python3 -c 'print(42)'".to_string()
}],
},
params
);

View File

@@ -26,6 +26,8 @@ use supports_color::Stream;
mod mcp_cmd;
use crate::mcp_cmd::McpCli;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
/// Codex CLI
///
@@ -45,6 +47,9 @@ struct MultitoolCli {
#[clap(flatten)]
pub config_overrides: CliConfigOverrides,
#[clap(flatten)]
pub feature_toggles: FeatureToggles,
#[clap(flatten)]
interactive: TuiCli,
@@ -97,6 +102,9 @@ enum Subcommand {
/// Internal: run the responses API proxy.
#[clap(hide = true)]
ResponsesApiProxy(ResponsesApiProxyArgs),
/// Inspect feature flags.
Features(FeaturesCli),
}
#[derive(Debug, Parser)]
@@ -231,6 +239,53 @@ fn print_exit_messages(exit_info: AppExitInfo) {
}
}
#[derive(Debug, Default, Parser, Clone)]
struct FeatureToggles {
/// Enable a feature (repeatable). Equivalent to `-c features.<name>=true`.
#[arg(long = "enable", value_name = "FEATURE", action = clap::ArgAction::Append, global = true)]
enable: Vec<String>,
/// Disable a feature (repeatable). Equivalent to `-c features.<name>=false`.
#[arg(long = "disable", value_name = "FEATURE", action = clap::ArgAction::Append, global = true)]
disable: Vec<String>,
}
impl FeatureToggles {
fn to_overrides(&self) -> Vec<String> {
let mut v = Vec::new();
for k in &self.enable {
v.push(format!("features.{k}=true"));
}
for k in &self.disable {
v.push(format!("features.{k}=false"));
}
v
}
}
#[derive(Debug, Parser)]
struct FeaturesCli {
#[command(subcommand)]
sub: FeaturesSubcommand,
}
#[derive(Debug, Parser)]
enum FeaturesSubcommand {
/// List known features with their stage and effective state.
List,
}
fn stage_str(stage: codex_core::features::Stage) -> &'static str {
use codex_core::features::Stage;
match stage {
Stage::Experimental => "experimental",
Stage::Beta => "beta",
Stage::Stable => "stable",
Stage::Deprecated => "deprecated",
Stage::Removed => "removed",
}
}
/// As early as possible in the process lifecycle, apply hardening measures. We
/// skip this in debug builds to avoid interfering with debugging.
#[ctor::ctor]
@@ -248,11 +303,17 @@ fn main() -> anyhow::Result<()> {
async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()> {
let MultitoolCli {
config_overrides: root_config_overrides,
config_overrides: mut root_config_overrides,
feature_toggles,
mut interactive,
subcommand,
} = MultitoolCli::parse();
// Fold --enable/--disable into config overrides so they flow to all subcommands.
root_config_overrides
.raw_overrides
.extend(feature_toggles.to_overrides());
match subcommand {
None => {
prepend_config_flags(
@@ -381,6 +442,30 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
Some(Subcommand::GenerateTs(gen_cli)) => {
codex_protocol_ts::generate_ts(&gen_cli.out_dir, gen_cli.prettier.as_deref())?;
}
Some(Subcommand::Features(FeaturesCli { sub })) => match sub {
FeaturesSubcommand::List => {
// Respect root-level `-c` overrides plus top-level flags like `--profile`.
let cli_kv_overrides = root_config_overrides
.parse_overrides()
.map_err(|e| anyhow::anyhow!(e))?;
// Thread through relevant top-level flags (at minimum, `--profile`).
// Also honor `--search` since it maps to a feature toggle.
let overrides = ConfigOverrides {
config_profile: interactive.config_profile.clone(),
tools_web_search_request: interactive.web_search.then_some(true),
..Default::default()
};
let config = Config::load_with_cli_overrides(cli_kv_overrides, overrides).await?;
for def in codex_core::features::FEATURES.iter() {
let name = def.key;
let stage = stage_str(def.stage);
let enabled = config.features.enabled(def.id);
println!("{name}\t{stage}\t{enabled}");
}
}
},
}
Ok(())
@@ -484,6 +569,7 @@ mod tests {
interactive,
config_overrides: root_overrides,
subcommand,
feature_toggles: _,
} = cli;
let Subcommand::Resume(ResumeCommand {

View File

@@ -13,10 +13,12 @@ use codex_core::config::load_global_mcp_servers;
use codex_core::config::write_global_mcp_servers;
use codex_core::config_types::McpServerConfig;
use codex_core::config_types::McpServerTransportConfig;
use codex_core::features::Feature;
use codex_core::mcp::auth::compute_auth_statuses;
use codex_core::protocol::McpAuthStatus;
use codex_rmcp_client::delete_oauth_tokens;
use codex_rmcp_client::perform_oauth_login;
use codex_rmcp_client::supports_oauth_login;
/// [experimental] Launch Codex as an MCP server or manage configured MCP servers.
///
@@ -189,7 +191,10 @@ impl McpCli {
async fn run_add(config_overrides: &CliConfigOverrides, add_args: AddArgs) -> Result<()> {
// Validate any provided overrides even though they are not currently applied.
config_overrides.parse_overrides().map_err(|e| anyhow!(e))?;
let overrides = config_overrides.parse_overrides().map_err(|e| anyhow!(e))?;
let config = Config::load_with_cli_overrides(overrides, ConfigOverrides::default())
.await
.context("failed to load configuration")?;
let AddArgs {
name,
@@ -225,17 +230,21 @@ async fn run_add(config_overrides: &CliConfigOverrides, add_args: AddArgs) -> Re
}
}
AddMcpTransportArgs {
streamable_http: Some(streamable_http),
streamable_http:
Some(AddMcpStreamableHttpArgs {
url,
bearer_token_env_var,
}),
..
} => McpServerTransportConfig::StreamableHttp {
url: streamable_http.url,
bearer_token_env_var: streamable_http.bearer_token_env_var,
url,
bearer_token_env_var,
},
AddMcpTransportArgs { .. } => bail!("exactly one of --command or --url must be provided"),
};
let new_entry = McpServerConfig {
transport,
transport: transport.clone(),
enabled: true,
startup_timeout_sec: None,
tool_timeout_sec: None,
@@ -248,6 +257,17 @@ async fn run_add(config_overrides: &CliConfigOverrides, add_args: AddArgs) -> Re
println!("Added global MCP server '{name}'.");
if let McpServerTransportConfig::StreamableHttp {
url,
bearer_token_env_var: None,
} = transport
&& matches!(supports_oauth_login(&url).await, Ok(true))
{
println!("Detected OAuth support. Starting OAuth flow…");
perform_oauth_login(&name, &url, config.mcp_oauth_credentials_store_mode).await?;
println!("Successfully logged in.");
}
Ok(())
}
@@ -285,7 +305,7 @@ async fn run_login(config_overrides: &CliConfigOverrides, login_args: LoginArgs)
.await
.context("failed to load configuration")?;
if !config.use_experimental_use_rmcp_client {
if !config.features.enabled(Feature::RmcpClient) {
bail!(
"OAuth login is only supported when experimental_use_rmcp_client is true in config.toml."
);

View File

@@ -1,3 +1,4 @@
use clap::Args;
use clap::Parser;
use codex_common::CliConfigOverrides;
@@ -6,4 +7,43 @@ use codex_common::CliConfigOverrides;
pub struct Cli {
#[clap(skip)]
pub config_overrides: CliConfigOverrides,
#[command(subcommand)]
pub command: Option<Command>,
}
#[derive(Debug, clap::Subcommand)]
pub enum Command {
/// Submit a new Codex Cloud task without launching the TUI.
Exec(ExecCommand),
}
#[derive(Debug, Args)]
pub struct ExecCommand {
/// Task prompt to run in Codex Cloud.
#[arg(value_name = "QUERY")]
pub query: Option<String>,
/// Target environment identifier (see `codex cloud` to browse).
#[arg(long = "env", value_name = "ENV_ID")]
pub environment: String,
/// Number of assistant attempts (best-of-N).
#[arg(
long = "attempts",
default_value_t = 1usize,
value_parser = parse_attempts
)]
pub attempts: usize,
}
fn parse_attempts(input: &str) -> Result<usize, String> {
let value: usize = input
.parse()
.map_err(|_| "attempts must be an integer between 1 and 4".to_string())?;
if (1..=4).contains(&value) {
Ok(value)
} else {
Err("attempts must be between 1 and 4".to_string())
}
}

View File

@@ -7,7 +7,9 @@ mod ui;
pub mod util;
pub use cli::Cli;
use anyhow::anyhow;
use std::io::IsTerminal;
use std::io::Read;
use std::path::PathBuf;
use std::sync::Arc;
use std::time::Duration;
@@ -23,6 +25,175 @@ struct ApplyJob {
diff_override: Option<String>,
}
struct BackendContext {
backend: Arc<dyn codex_cloud_tasks_client::CloudBackend>,
base_url: String,
}
async fn init_backend(user_agent_suffix: &str) -> anyhow::Result<BackendContext> {
let use_mock = matches!(
std::env::var("CODEX_CLOUD_TASKS_MODE").ok().as_deref(),
Some("mock") | Some("MOCK")
);
let base_url = std::env::var("CODEX_CLOUD_TASKS_BASE_URL")
.unwrap_or_else(|_| "https://chatgpt.com/backend-api".to_string());
set_user_agent_suffix(user_agent_suffix);
if use_mock {
return Ok(BackendContext {
backend: Arc::new(codex_cloud_tasks_client::MockClient),
base_url,
});
}
let ua = codex_core::default_client::get_codex_user_agent();
let mut http = codex_cloud_tasks_client::HttpClient::new(base_url.clone())?.with_user_agent(ua);
let style = if base_url.contains("/backend-api") {
"wham"
} else {
"codex-api"
};
append_error_log(format!("startup: base_url={base_url} path_style={style}"));
let auth = match codex_core::config::find_codex_home()
.ok()
.map(|home| codex_login::AuthManager::new(home, false))
.and_then(|am| am.auth())
{
Some(auth) => auth,
None => {
eprintln!(
"Not signed in. Please run 'codex login' to sign in with ChatGPT, then re-run 'codex cloud'."
);
std::process::exit(1);
}
};
if let Some(acc) = auth.get_account_id() {
append_error_log(format!("auth: mode=ChatGPT account_id={acc}"));
}
let token = match auth.get_token().await {
Ok(t) if !t.is_empty() => t,
_ => {
eprintln!(
"Not signed in. Please run 'codex login' to sign in with ChatGPT, then re-run 'codex cloud'."
);
std::process::exit(1);
}
};
http = http.with_bearer_token(token.clone());
if let Some(acc) = auth
.get_account_id()
.or_else(|| util::extract_chatgpt_account_id(&token))
{
append_error_log(format!("auth: set ChatGPT-Account-Id header: {acc}"));
http = http.with_chatgpt_account_id(acc);
}
Ok(BackendContext {
backend: Arc::new(http),
base_url,
})
}
async fn run_exec_command(args: crate::cli::ExecCommand) -> anyhow::Result<()> {
let crate::cli::ExecCommand {
query,
environment,
attempts,
} = args;
let ctx = init_backend("codex_cloud_tasks_exec").await?;
let prompt = resolve_query_input(query)?;
let env_id = resolve_environment_id(&ctx, &environment).await?;
let created = codex_cloud_tasks_client::CloudBackend::create_task(
&*ctx.backend,
&env_id,
&prompt,
"main",
false,
attempts,
)
.await?;
let url = util::task_url(&ctx.base_url, &created.id.0);
println!("{url}");
Ok(())
}
async fn resolve_environment_id(ctx: &BackendContext, requested: &str) -> anyhow::Result<String> {
let trimmed = requested.trim();
if trimmed.is_empty() {
return Err(anyhow!("environment id must not be empty"));
}
let normalized = util::normalize_base_url(&ctx.base_url);
let headers = util::build_chatgpt_headers().await;
let environments = crate::env_detect::list_environments(&normalized, &headers).await?;
if environments.is_empty() {
return Err(anyhow!(
"no cloud environments are available for this workspace"
));
}
if let Some(row) = environments.iter().find(|row| row.id == trimmed) {
return Ok(row.id.clone());
}
let label_matches = environments
.iter()
.filter(|row| {
row.label
.as_deref()
.map(|label| label.eq_ignore_ascii_case(trimmed))
.unwrap_or(false)
})
.collect::<Vec<_>>();
match label_matches.as_slice() {
[] => Err(anyhow!(
"environment '{trimmed}' not found; run `codex cloud` to list available environments"
)),
[single] => Ok(single.id.clone()),
[first, rest @ ..] => {
let first_id = &first.id;
if rest.iter().all(|row| row.id == *first_id) {
Ok(first_id.clone())
} else {
Err(anyhow!(
"environment label '{trimmed}' is ambiguous; run `codex cloud` to pick the desired environment id"
))
}
}
}
}
fn resolve_query_input(query_arg: Option<String>) -> anyhow::Result<String> {
match query_arg {
Some(q) if q != "-" => Ok(q),
maybe_dash => {
let force_stdin = matches!(maybe_dash.as_deref(), Some("-"));
if std::io::stdin().is_terminal() && !force_stdin {
return Err(anyhow!(
"no query provided. Pass one as an argument or pipe it via stdin."
));
}
if !force_stdin {
eprintln!("Reading query from stdin...");
}
let mut buffer = String::new();
std::io::stdin()
.read_to_string(&mut buffer)
.map_err(|e| anyhow!("failed to read query from stdin: {e}"))?;
if buffer.trim().is_empty() {
return Err(anyhow!(
"no query provided via stdin (received empty input)."
));
}
Ok(buffer)
}
}
}
fn level_from_status(status: codex_cloud_tasks_client::ApplyStatus) -> app::ApplyResultLevel {
match status {
codex_cloud_tasks_client::ApplyStatus::Success => app::ApplyResultLevel::Success,
@@ -148,7 +319,14 @@ fn spawn_apply(
// (no standalone patch summarizer needed UI displays raw diffs)
/// Entry point for the `codex cloud` subcommand.
pub async fn run_main(_cli: Cli, _codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()> {
pub async fn run_main(cli: Cli, _codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()> {
if let Some(command) = cli.command {
return match command {
crate::cli::Command::Exec(args) => run_exec_command(args).await,
};
}
let Cli { .. } = cli;
// Very minimal logging setup; mirrors other crates' pattern.
let default_level = "error";
let _ = tracing_subscriber::fmt()
@@ -162,72 +340,8 @@ pub async fn run_main(_cli: Cli, _codex_linux_sandbox_exe: Option<PathBuf>) -> a
.try_init();
info!("Launching Cloud Tasks list UI");
set_user_agent_suffix("codex_cloud_tasks_tui");
// Default to online unless explicitly configured to use mock.
let use_mock = matches!(
std::env::var("CODEX_CLOUD_TASKS_MODE").ok().as_deref(),
Some("mock") | Some("MOCK")
);
let backend: Arc<dyn codex_cloud_tasks_client::CloudBackend> = if use_mock {
Arc::new(codex_cloud_tasks_client::MockClient)
} else {
// Build an HTTP client against the configured (or default) base URL.
let base_url = std::env::var("CODEX_CLOUD_TASKS_BASE_URL")
.unwrap_or_else(|_| "https://chatgpt.com/backend-api".to_string());
let ua = codex_core::default_client::get_codex_user_agent();
let mut http =
codex_cloud_tasks_client::HttpClient::new(base_url.clone())?.with_user_agent(ua);
// Log which base URL and path style we're going to use.
let style = if base_url.contains("/backend-api") {
"wham"
} else {
"codex-api"
};
append_error_log(format!("startup: base_url={base_url} path_style={style}"));
// Require ChatGPT login (SWIC). Exit with a clear message if missing.
let _token = match codex_core::config::find_codex_home()
.ok()
.map(|home| codex_login::AuthManager::new(home, false))
.and_then(|am| am.auth())
{
Some(auth) => {
// Log account context for debugging workspace selection.
if let Some(acc) = auth.get_account_id() {
append_error_log(format!("auth: mode=ChatGPT account_id={acc}"));
}
match auth.get_token().await {
Ok(t) if !t.is_empty() => {
// Attach token and ChatGPT-Account-Id header if available
http = http.with_bearer_token(t.clone());
if let Some(acc) = auth
.get_account_id()
.or_else(|| util::extract_chatgpt_account_id(&t))
{
append_error_log(format!("auth: set ChatGPT-Account-Id header: {acc}"));
http = http.with_chatgpt_account_id(acc);
}
t
}
_ => {
eprintln!(
"Not signed in. Please run 'codex login' to sign in with ChatGPT, then re-run 'codex cloud'."
);
std::process::exit(1);
}
}
}
None => {
eprintln!(
"Not signed in. Please run 'codex login' to sign in with ChatGPT, then re-run 'codex cloud'."
);
std::process::exit(1);
}
};
Arc::new(http)
};
let BackendContext { backend, .. } = init_backend("codex_cloud_tasks_tui").await?;
let backend = backend;
// Terminal setup
use crossterm::ExecutableCommand;

View File

@@ -91,3 +91,18 @@ pub async fn build_chatgpt_headers() -> HeaderMap {
}
headers
}
/// Construct a browser-friendly task URL for the given backend base URL.
pub fn task_url(base_url: &str, task_id: &str) -> String {
let normalized = normalize_base_url(base_url);
if let Some(root) = normalized.strip_suffix("/backend-api") {
return format!("{root}/codex/tasks/{task_id}");
}
if let Some(root) = normalized.strip_suffix("/api/codex") {
return format!("{root}/codex/tasks/{task_id}");
}
if normalized.ends_with("/codex") {
return format!("{normalized}/tasks/{task_id}");
}
format!("{normalized}/codex/tasks/{task_id}")
}

View File

@@ -47,6 +47,7 @@ use crate::openai_tools::create_tools_json_for_responses_api;
use crate::protocol::RateLimitSnapshot;
use crate::protocol::RateLimitWindow;
use crate::protocol::TokenUsage;
use crate::state::TaskKind;
use crate::token_data::PlanType;
use crate::util::backoff;
use codex_otel::otel_event_manager::OtelEventManager;
@@ -123,8 +124,16 @@ impl ModelClient {
/// the provider config. Public callers always invoke `stream()` the
/// specialised helpers are private to avoid accidental misuse.
pub async fn stream(&self, prompt: &Prompt) -> Result<ResponseStream> {
self.stream_with_task_kind(prompt, TaskKind::Regular).await
}
pub(crate) async fn stream_with_task_kind(
&self,
prompt: &Prompt,
task_kind: TaskKind,
) -> Result<ResponseStream> {
match self.provider.wire_api {
WireApi::Responses => self.stream_responses(prompt).await,
WireApi::Responses => self.stream_responses(prompt, task_kind).await,
WireApi::Chat => {
// Create the raw streaming connection first.
let response_stream = stream_chat_completions(
@@ -165,7 +174,11 @@ impl ModelClient {
}
/// Implementation for the OpenAI *Responses* experimental API.
async fn stream_responses(&self, prompt: &Prompt) -> Result<ResponseStream> {
async fn stream_responses(
&self,
prompt: &Prompt,
task_kind: TaskKind,
) -> Result<ResponseStream> {
if let Some(path) = &*CODEX_RS_SSE_FIXTURE {
// short circuit for tests
warn!(path, "Streaming from fixture");
@@ -244,7 +257,7 @@ 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(attempt, &payload_json, &auth_manager, task_kind)
.await
{
Ok(stream) => {
@@ -272,6 +285,7 @@ impl ModelClient {
attempt: u64,
payload_json: &Value,
auth_manager: &Option<Arc<AuthManager>>,
task_kind: TaskKind,
) -> std::result::Result<ResponseStream, StreamAttemptError> {
// Always fetch the latest auth in case a prior attempt refreshed the token.
let auth = auth_manager.as_ref().and_then(|m| m.auth());
@@ -294,6 +308,7 @@ impl ModelClient {
.header("conversation_id", self.conversation_id.to_string())
.header("session_id", self.conversation_id.to_string())
.header(reqwest::header::ACCEPT, "text/event-stream")
.header("Codex-Task-Type", task_kind.header_value())
.json(payload_json);
if let Some(auth) = auth.as_ref()

View File

@@ -17,6 +17,7 @@ use codex_apply_patch::ApplyPatchAction;
use codex_protocol::ConversationId;
use codex_protocol::protocol::ConversationPathResponseEvent;
use codex_protocol::protocol::ExitedReviewModeEvent;
use codex_protocol::protocol::McpAuthStatus;
use codex_protocol::protocol::ReviewRequest;
use codex_protocol::protocol::RolloutItem;
use codex_protocol::protocol::SessionSource;
@@ -99,6 +100,7 @@ use crate::rollout::RolloutRecorderParams;
use crate::shell;
use crate::state::ActiveTurn;
use crate::state::SessionServices;
use crate::state::TaskKind;
use crate::tasks::CompactTask;
use crate::tasks::RegularTask;
use crate::tasks::ReviewTask;
@@ -364,15 +366,32 @@ impl Session {
let mcp_fut = McpConnectionManager::new(
config.mcp_servers.clone(),
config.use_experimental_use_rmcp_client,
config
.features
.enabled(crate::features::Feature::RmcpClient),
config.mcp_oauth_credentials_store_mode,
);
let default_shell_fut = shell::default_user_shell();
let history_meta_fut = crate::message_history::history_metadata(&config);
let auth_statuses_fut = compute_auth_statuses(
config.mcp_servers.iter(),
config.mcp_oauth_credentials_store_mode,
);
// Join all independent futures.
let (rollout_recorder, mcp_res, default_shell, (history_log_id, history_entry_count)) =
tokio::join!(rollout_fut, mcp_fut, default_shell_fut, history_meta_fut);
let (
rollout_recorder,
mcp_res,
default_shell,
(history_log_id, history_entry_count),
auth_statuses,
) = tokio::join!(
rollout_fut,
mcp_fut,
default_shell_fut,
history_meta_fut,
auth_statuses_fut
);
let rollout_recorder = rollout_recorder.map_err(|e| {
error!("failed to initialize rollout recorder: {e:#}");
@@ -399,11 +418,24 @@ impl Session {
// Surface individual client start-up failures to the user.
if !failed_clients.is_empty() {
for (server_name, err) in failed_clients {
let message = format!("MCP client for `{server_name}` failed to start: {err:#}");
error!("{message}");
let log_message =
format!("MCP client for `{server_name}` failed to start: {err:#}");
error!("{log_message}");
let display_message = if matches!(
auth_statuses.get(&server_name),
Some(McpAuthStatus::NotLoggedIn)
) {
format!(
"The {server_name} MCP server is not logged in. Run `codex mcp login {server_name}` to log in."
)
} else {
log_message
};
post_session_configured_error_events.push(Event {
id: INITIAL_SUBMIT_ID.to_owned(),
msg: EventMsg::Error(ErrorEvent { message }),
msg: EventMsg::Error(ErrorEvent {
message: display_message,
}),
});
}
}
@@ -446,12 +478,7 @@ impl Session {
client,
tools_config: ToolsConfig::new(&ToolsConfigParams {
model_family: &config.model_family,
include_plan_tool: config.include_plan_tool,
include_apply_patch_tool: config.include_apply_patch_tool,
include_web_search_request: config.tools_web_search_request,
use_streamable_shell_tool: config.use_experimental_streamable_shell_tool,
include_view_image_tool: config.include_view_image_tool,
experimental_unified_exec_tool: config.use_experimental_unified_exec_tool,
features: &config.features,
}),
user_instructions,
base_instructions,
@@ -593,6 +620,7 @@ impl Session {
warn!("Overwriting existing pending approval for sub_id: {event_id}");
}
let parsed_cmd = parse_command(&command);
let event = Event {
id: event_id,
msg: EventMsg::ExecApprovalRequest(ExecApprovalRequestEvent {
@@ -600,6 +628,7 @@ impl Session {
command,
cwd,
reason,
parsed_cmd,
}),
};
self.send_event(event).await;
@@ -855,10 +884,7 @@ impl Session {
call_id,
command: command_for_display.clone(),
cwd,
parsed_cmd: parse_command(&command_for_display)
.into_iter()
.map(Into::into)
.collect(),
parsed_cmd: parse_command(&command_for_display),
}),
};
let event = Event {
@@ -1195,12 +1221,7 @@ async fn submission_loop(
let tools_config = ToolsConfig::new(&ToolsConfigParams {
model_family: &effective_family,
include_plan_tool: config.include_plan_tool,
include_apply_patch_tool: config.include_apply_patch_tool,
include_web_search_request: config.tools_web_search_request,
use_streamable_shell_tool: config.use_experimental_streamable_shell_tool,
include_view_image_tool: config.include_view_image_tool,
experimental_unified_exec_tool: config.use_experimental_unified_exec_tool,
features: &config.features,
});
let new_turn_context = TurnContext {
@@ -1297,14 +1318,7 @@ async fn submission_loop(
client,
tools_config: ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: config.include_plan_tool,
include_apply_patch_tool: config.include_apply_patch_tool,
include_web_search_request: config.tools_web_search_request,
use_streamable_shell_tool: config
.use_experimental_streamable_shell_tool,
include_view_image_tool: config.include_view_image_tool,
experimental_unified_exec_tool: config
.use_experimental_unified_exec_tool,
features: &config.features,
}),
user_instructions: turn_context.user_instructions.clone(),
base_instructions: turn_context.base_instructions.clone(),
@@ -1536,14 +1550,15 @@ async fn spawn_review_thread(
let model = config.review_model.clone();
let review_model_family = find_family_for_model(&model)
.unwrap_or_else(|| parent_turn_context.client.get_model_family());
// For reviews, disable plan, web_search, view_image regardless of global settings.
let mut review_features = config.features.clone();
review_features.disable(crate::features::Feature::PlanTool);
review_features.disable(crate::features::Feature::WebSearchRequest);
review_features.disable(crate::features::Feature::ViewImageTool);
review_features.disable(crate::features::Feature::StreamableShell);
let tools_config = ToolsConfig::new(&ToolsConfigParams {
model_family: &review_model_family,
include_plan_tool: false,
include_apply_patch_tool: config.include_apply_patch_tool,
include_web_search_request: false,
use_streamable_shell_tool: false,
include_view_image_tool: false,
experimental_unified_exec_tool: config.use_experimental_unified_exec_tool,
features: &review_features,
});
let base_instructions = REVIEW_PROMPT.to_string();
@@ -1634,6 +1649,7 @@ pub(crate) async fn run_task(
turn_context: Arc<TurnContext>,
sub_id: String,
input: Vec<InputItem>,
task_kind: TaskKind,
) -> Option<String> {
if input.is_empty() {
return None;
@@ -1717,6 +1733,7 @@ pub(crate) async fn run_task(
Arc::clone(&turn_diff_tracker),
sub_id.clone(),
turn_input,
task_kind,
)
.await
{
@@ -1869,6 +1886,7 @@ pub(crate) async fn run_task(
);
sess.notifier()
.notify(&UserNotification::AgentTurnComplete {
thread_id: sess.conversation_id.to_string(),
turn_id: sub_id.clone(),
input_messages: turn_input_messages,
last_assistant_message: last_agent_message.clone(),
@@ -1942,6 +1960,7 @@ async fn run_turn(
turn_diff_tracker: SharedTurnDiffTracker,
sub_id: String,
input: Vec<ResponseItem>,
task_kind: TaskKind,
) -> CodexResult<TurnRunResult> {
let mcp_tools = sess.services.mcp_connection_manager.list_all_tools();
let router = Arc::new(ToolRouter::from_config(
@@ -1971,6 +1990,7 @@ async fn run_turn(
Arc::clone(&turn_diff_tracker),
&sub_id,
&prompt,
task_kind,
)
.await
{
@@ -2008,9 +2028,7 @@ async fn run_turn(
// at a seemingly frozen screen.
sess.notify_stream_error(
&sub_id,
format!(
"stream error: {e}; retrying {retries}/{max_retries} in {delay:?}"
),
format!("Re-connecting... {retries}/{max_retries}"),
)
.await;
@@ -2046,6 +2064,7 @@ async fn try_run_turn(
turn_diff_tracker: SharedTurnDiffTracker,
sub_id: &str,
prompt: &Prompt,
task_kind: TaskKind,
) -> CodexResult<TurnRunResult> {
// call_ids that are part of this response.
let completed_call_ids = prompt
@@ -2111,7 +2130,11 @@ async fn try_run_turn(
summary: turn_context.client.get_reasoning_summary(),
});
sess.persist_rollout_items(&[rollout_item]).await;
let mut stream = turn_context.client.clone().stream(&prompt).await?;
let mut stream = turn_context
.client
.clone()
.stream_with_task_kind(prompt.as_ref(), task_kind)
.await?;
let tool_runtime = ToolCallRuntime::new(
Arc::clone(&router),
@@ -2750,12 +2773,7 @@ mod tests {
);
let tools_config = ToolsConfig::new(&ToolsConfigParams {
model_family: &config.model_family,
include_plan_tool: config.include_plan_tool,
include_apply_patch_tool: config.include_apply_patch_tool,
include_web_search_request: config.tools_web_search_request,
use_streamable_shell_tool: config.use_experimental_streamable_shell_tool,
include_view_image_tool: config.include_view_image_tool,
experimental_unified_exec_tool: config.use_experimental_unified_exec_tool,
features: &config.features,
});
let turn_context = TurnContext {
client,
@@ -2823,12 +2841,7 @@ mod tests {
);
let tools_config = ToolsConfig::new(&ToolsConfigParams {
model_family: &config.model_family,
include_plan_tool: config.include_plan_tool,
include_apply_patch_tool: config.include_apply_patch_tool,
include_web_search_request: config.tools_web_search_request,
use_streamable_shell_tool: config.use_experimental_streamable_shell_tool,
include_view_image_tool: config.include_view_image_tool,
experimental_unified_exec_tool: config.use_experimental_unified_exec_tool,
features: &config.features,
});
let turn_context = Arc::new(TurnContext {
client,

View File

@@ -16,6 +16,7 @@ use crate::protocol::InputItem;
use crate::protocol::InputMessageKind;
use crate::protocol::TaskStartedEvent;
use crate::protocol::TurnContextItem;
use crate::state::TaskKind;
use crate::truncate::truncate_middle;
use crate::util::backoff;
use askama::Template;
@@ -70,13 +71,15 @@ async fn run_compact_task_inner(
input: Vec<InputItem>,
) {
let initial_input_for_turn: ResponseInputItem = ResponseInputItem::from(input);
let mut turn_input = sess
.turn_input_with_history(vec![initial_input_for_turn.clone().into()])
.await;
// Track the items we append for this compact prompt so trimming does not drop them.
let extra_items: Vec<ResponseItem> = vec![initial_input_for_turn.clone().into()];
let mut turn_input = sess.turn_input_with_history(extra_items.clone()).await;
let mut truncated_count = 0usize;
let mut trimmed_tails: Vec<Vec<ResponseItem>> = Vec::new();
let max_retries = turn_context.client.get_provider().stream_max_retries();
let mut retries = 0;
let mut context_retries = 0;
let mut stream_retries = 0;
let rollout_item = RolloutItem::TurnContext(TurnContextItem {
cwd: turn_context.cwd.clone(),
@@ -113,11 +116,32 @@ async fn run_compact_task_inner(
return;
}
Err(e @ CodexErr::ContextWindowExceeded) => {
if turn_input.len() > 1 {
turn_input.remove(0);
truncated_count += 1;
retries = 0;
continue;
// Drop the most recent user turn (its message plus ensuing traffic) and retry.
if turn_input.len() > extra_items.len() {
let history_len = turn_input.len() - extra_items.len();
let mut prompt_items = turn_input.split_off(history_len);
let trimmed = trim_recent_history_to_previous_user_message(&mut turn_input);
turn_input.append(&mut prompt_items);
if !trimmed.is_empty() {
truncated_count += trimmed.len();
trimmed_tails.push(trimmed);
if context_retries >= max_retries {
sess.set_total_tokens_full(&sub_id, turn_context.as_ref())
.await;
let event = Event {
id: sub_id.clone(),
msg: EventMsg::Error(ErrorEvent {
message: e.to_string(),
}),
};
sess.send_event(event).await;
return;
}
context_retries += 1;
stream_retries = 0;
// Keep stream retry budget untouched; we trimmed context successfully.
continue;
}
}
sess.set_total_tokens_full(&sub_id, turn_context.as_ref())
.await;
@@ -131,14 +155,12 @@ async fn run_compact_task_inner(
return;
}
Err(e) => {
if retries < max_retries {
retries += 1;
let delay = backoff(retries);
if stream_retries < max_retries {
stream_retries += 1;
let delay = backoff(stream_retries);
sess.notify_stream_error(
&sub_id,
format!(
"stream error: {e}; retrying {retries}/{max_retries} in {delay:?}"
),
format!("Re-connecting... {stream_retries}/{max_retries}"),
)
.await;
tokio::time::sleep(delay).await;
@@ -161,7 +183,10 @@ async fn run_compact_task_inner(
let summary_text = get_last_assistant_message_from_turn(&history_snapshot).unwrap_or_default();
let user_messages = collect_user_messages(&history_snapshot);
let initial_context = sess.build_initial_context(turn_context.as_ref());
let new_history = build_compacted_history(initial_context, &user_messages, &summary_text);
let mut new_history = build_compacted_history(initial_context, &user_messages, &summary_text);
for mut trimmed in trimmed_tails.into_iter().rev() {
new_history.append(&mut trimmed);
}
sess.replace_history(new_history).await;
let rollout_item = RolloutItem::Compacted(CompactedItem {
@@ -178,6 +203,27 @@ async fn run_compact_task_inner(
sess.send_event(event).await;
}
/// Trim conversation history back to the previous user message boundary, removing that user turn.
///
/// Returns the removed items in their original order so they can be restored later.
fn trim_recent_history_to_previous_user_message(
turn_input: &mut Vec<ResponseItem>,
) -> Vec<ResponseItem> {
if turn_input.is_empty() {
return Vec::new();
}
if let Some(last_user_index) = turn_input.iter().rposition(|item| {
matches!(
item,
ResponseItem::Message { role, .. } if role == "user"
)
}) {
turn_input.split_off(last_user_index)
} else {
std::mem::take(turn_input)
}
}
pub fn content_items_to_text(content: &[ContentItem]) -> Option<String> {
let mut pieces = Vec::new();
for item in content {
@@ -260,7 +306,11 @@ async fn drain_to_completed(
sub_id: &str,
prompt: &Prompt,
) -> CodexResult<()> {
let mut stream = turn_context.client.clone().stream(prompt).await?;
let mut stream = turn_context
.client
.clone()
.stream_with_task_kind(prompt, TaskKind::Compact)
.await?;
loop {
let maybe_event = stream.next().await;
let Some(event) = maybe_event else {

View File

@@ -17,6 +17,10 @@ use crate::config_types::ShellEnvironmentPolicy;
use crate::config_types::ShellEnvironmentPolicyToml;
use crate::config_types::Tui;
use crate::config_types::UriBasedFileOpener;
use crate::features::Feature;
use crate::features::FeatureOverrides;
use crate::features::Features;
use crate::features::FeaturesToml;
use crate::git_info::resolve_root_git_project_for_trust;
use crate::model_family::ModelFamily;
use crate::model_family::derive_default_model_family;
@@ -24,6 +28,8 @@ use crate::model_family::find_family_for_model;
use crate::model_provider_info::ModelProviderInfo;
use crate::model_provider_info::built_in_model_providers;
use crate::openai_model_info::get_model_info;
use crate::project_doc::DEFAULT_PROJECT_DOC_FILENAME;
use crate::project_doc::LOCAL_PROJECT_DOC_FILENAME;
use crate::protocol::AskForApproval;
use crate::protocol::SandboxPolicy;
use anyhow::Context;
@@ -218,6 +224,9 @@ pub struct Config {
/// Include the `view_image` tool that lets the agent attach a local image path to context.
pub include_view_image_tool: bool,
/// Centralized feature flags; source of truth for feature gating.
pub features: Features,
/// The active profile name used to derive this `Config` (if any).
pub active_profile: Option<String>,
@@ -794,19 +803,15 @@ pub struct ConfigToml {
/// Base URL for requests to ChatGPT (as opposed to the OpenAI API).
pub chatgpt_base_url: Option<String>,
/// Experimental path to a file whose contents replace the built-in BASE_INSTRUCTIONS.
pub experimental_instructions_file: Option<PathBuf>,
pub experimental_use_exec_command_tool: Option<bool>,
pub experimental_use_unified_exec_tool: Option<bool>,
pub experimental_use_rmcp_client: Option<bool>,
pub experimental_use_freeform_apply_patch: Option<bool>,
pub projects: Option<HashMap<String, ProjectConfig>>,
/// Nested tools section for feature toggles
pub tools: Option<ToolsToml>,
/// Centralized feature flags (new). Prefer this over individual toggles.
#[serde(default)]
pub features: Option<FeaturesToml>,
/// When true, disables burst-paste detection for typed input entirely.
/// All characters are inserted as they are received, and no buffering
/// or placeholder replacement will occur for fast keypress bursts.
@@ -817,6 +822,13 @@ pub struct ConfigToml {
/// Tracks whether the Windows onboarding screen has been acknowledged.
pub windows_wsl_setup_acknowledged: Option<bool>,
/// Legacy, now use features
pub experimental_instructions_file: Option<PathBuf>,
pub experimental_use_exec_command_tool: Option<bool>,
pub experimental_use_unified_exec_tool: Option<bool>,
pub experimental_use_rmcp_client: Option<bool>,
pub experimental_use_freeform_apply_patch: Option<bool>,
}
impl From<ConfigToml> for UserSavedConfig {
@@ -980,9 +992,9 @@ impl Config {
config_profile: config_profile_key,
codex_linux_sandbox_exe,
base_instructions,
include_plan_tool,
include_apply_patch_tool,
include_view_image_tool,
include_plan_tool: include_plan_tool_override,
include_apply_patch_tool: include_apply_patch_tool_override,
include_view_image_tool: include_view_image_tool_override,
show_raw_agent_reasoning,
tools_web_search_request: override_tools_web_search_request,
} = overrides;
@@ -1005,6 +1017,15 @@ impl Config {
None => ConfigProfile::default(),
};
let feature_overrides = FeatureOverrides {
include_plan_tool: include_plan_tool_override,
include_apply_patch_tool: include_apply_patch_tool_override,
include_view_image_tool: include_view_image_tool_override,
web_search_request: override_tools_web_search_request,
};
let features = Features::from_config(&cfg, &config_profile, feature_overrides);
let sandbox_policy = cfg.derive_sandbox_policy(sandbox_mode);
let mut model_providers = built_in_model_providers();
@@ -1050,13 +1071,13 @@ impl Config {
let history = cfg.history.unwrap_or_default();
let tools_web_search_request = override_tools_web_search_request
.or(cfg.tools.as_ref().and_then(|t| t.web_search))
.unwrap_or(false);
let include_view_image_tool = include_view_image_tool
.or(cfg.tools.as_ref().and_then(|t| t.view_image))
.unwrap_or(true);
let include_plan_tool_flag = features.enabled(Feature::PlanTool);
let include_apply_patch_tool_flag = features.enabled(Feature::ApplyPatchFreeform);
let include_view_image_tool_flag = features.enabled(Feature::ViewImageTool);
let tools_web_search_request = features.enabled(Feature::WebSearchRequest);
let use_experimental_streamable_shell_tool = features.enabled(Feature::StreamableShell);
let use_experimental_unified_exec_tool = features.enabled(Feature::UnifiedExec);
let use_experimental_use_rmcp_client = features.enabled(Feature::RmcpClient);
let model = model
.or(config_profile.model)
@@ -1104,6 +1125,15 @@ impl Config {
.or(cfg.review_model)
.unwrap_or_else(default_review_model);
let mut approval_policy = approval_policy
.or(config_profile.approval_policy)
.or(cfg.approval_policy)
.unwrap_or_else(AskForApproval::default);
if features.enabled(Feature::ApproveAll) {
approval_policy = AskForApproval::OnRequest;
}
let config = Self {
model,
review_model,
@@ -1114,10 +1144,7 @@ impl Config {
model_provider_id,
model_provider,
cwd: resolved_cwd,
approval_policy: approval_policy
.or(config_profile.approval_policy)
.or(cfg.approval_policy)
.unwrap_or_else(AskForApproval::default),
approval_policy,
sandbox_policy,
shell_environment_policy,
notify: cfg.notify,
@@ -1164,19 +1191,14 @@ 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(false),
include_apply_patch_tool: include_apply_patch_tool
.or(cfg.experimental_use_freeform_apply_patch)
.unwrap_or(false),
include_plan_tool: include_plan_tool_flag,
include_apply_patch_tool: include_apply_patch_tool_flag,
tools_web_search_request,
use_experimental_streamable_shell_tool: cfg
.experimental_use_exec_command_tool
.unwrap_or(false),
use_experimental_unified_exec_tool: cfg
.experimental_use_unified_exec_tool
.unwrap_or(false),
use_experimental_use_rmcp_client: cfg.experimental_use_rmcp_client.unwrap_or(false),
include_view_image_tool,
use_experimental_streamable_shell_tool,
use_experimental_unified_exec_tool,
use_experimental_use_rmcp_client,
include_view_image_tool: include_view_image_tool_flag,
features,
active_profile: active_profile_name,
windows_wsl_setup_acknowledged: cfg.windows_wsl_setup_acknowledged.unwrap_or(false),
disable_paste_burst: cfg.disable_paste_burst.unwrap_or(false),
@@ -1203,20 +1225,18 @@ impl Config {
}
fn load_instructions(codex_dir: Option<&Path>) -> Option<String> {
let mut p = match codex_dir {
Some(p) => p.to_path_buf(),
None => return None,
};
p.push("AGENTS.md");
std::fs::read_to_string(&p).ok().and_then(|s| {
let s = s.trim();
if s.is_empty() {
None
} else {
Some(s.to_string())
let base = codex_dir?;
for candidate in [LOCAL_PROJECT_DOC_FILENAME, DEFAULT_PROJECT_DOC_FILENAME] {
let mut path = base.to_path_buf();
path.push(candidate);
if let Ok(contents) = std::fs::read_to_string(&path) {
let trimmed = contents.trim();
if !trimmed.is_empty() {
return Some(trimmed.to_string());
}
}
})
}
None
}
fn get_base_instructions(
@@ -1309,6 +1329,7 @@ pub fn log_dir(cfg: &Config) -> std::io::Result<PathBuf> {
mod tests {
use crate::config_types::HistoryPersistence;
use crate::config_types::Notifications;
use crate::features::Feature;
use super::*;
use pretty_assertions::assert_eq;
@@ -1417,6 +1438,26 @@ exclude_slash_tmp = true
);
}
#[test]
fn approve_all_feature_forces_on_request_policy() -> std::io::Result<()> {
let cfg = r#"
[features]
approve_all = true
"#;
let parsed = toml::from_str::<ConfigToml>(cfg)
.expect("TOML deserialization should succeed for approve_all feature");
let temp_dir = TempDir::new()?;
let config = Config::load_from_base_config_with_overrides(
parsed,
ConfigOverrides::default(),
temp_dir.path().to_path_buf(),
)?;
assert!(config.features.enabled(Feature::ApproveAll));
assert_eq!(config.approval_policy, AskForApproval::OnRequest);
Ok(())
}
#[test]
fn config_defaults_to_auto_oauth_store_mode() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
@@ -1436,6 +1477,93 @@ exclude_slash_tmp = true
Ok(())
}
#[test]
fn profile_legacy_toggles_override_base() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
let mut profiles = HashMap::new();
profiles.insert(
"work".to_string(),
ConfigProfile {
include_plan_tool: Some(true),
include_view_image_tool: Some(false),
..Default::default()
},
);
let cfg = ConfigToml {
profiles,
profile: Some("work".to_string()),
..Default::default()
};
let config = Config::load_from_base_config_with_overrides(
cfg,
ConfigOverrides::default(),
codex_home.path().to_path_buf(),
)?;
assert!(config.features.enabled(Feature::PlanTool));
assert!(!config.features.enabled(Feature::ViewImageTool));
assert!(config.include_plan_tool);
assert!(!config.include_view_image_tool);
Ok(())
}
#[test]
fn feature_table_overrides_legacy_flags() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
let mut entries = BTreeMap::new();
entries.insert("plan_tool".to_string(), false);
entries.insert("apply_patch_freeform".to_string(), false);
let cfg = ConfigToml {
features: Some(crate::features::FeaturesToml { entries }),
..Default::default()
};
let config = Config::load_from_base_config_with_overrides(
cfg,
ConfigOverrides::default(),
codex_home.path().to_path_buf(),
)?;
assert!(!config.features.enabled(Feature::PlanTool));
assert!(!config.features.enabled(Feature::ApplyPatchFreeform));
assert!(!config.include_plan_tool);
assert!(!config.include_apply_patch_tool);
Ok(())
}
#[test]
fn legacy_toggles_map_to_features() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
let cfg = ConfigToml {
experimental_use_exec_command_tool: Some(true),
experimental_use_unified_exec_tool: Some(true),
experimental_use_rmcp_client: Some(true),
experimental_use_freeform_apply_patch: Some(true),
..Default::default()
};
let config = Config::load_from_base_config_with_overrides(
cfg,
ConfigOverrides::default(),
codex_home.path().to_path_buf(),
)?;
assert!(config.features.enabled(Feature::ApplyPatchFreeform));
assert!(config.features.enabled(Feature::StreamableShell));
assert!(config.features.enabled(Feature::UnifiedExec));
assert!(config.features.enabled(Feature::RmcpClient));
assert!(config.include_apply_patch_tool);
assert!(config.use_experimental_streamable_shell_tool);
assert!(config.use_experimental_unified_exec_tool);
assert!(config.use_experimental_use_rmcp_client);
Ok(())
}
#[test]
fn config_honors_explicit_file_oauth_store_mode() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
@@ -2120,6 +2248,7 @@ model_verbosity = "high"
use_experimental_unified_exec_tool: false,
use_experimental_use_rmcp_client: false,
include_view_image_tool: true,
features: Features::with_defaults(),
active_profile: Some("o3".to_string()),
windows_wsl_setup_acknowledged: false,
disable_paste_burst: false,
@@ -2183,6 +2312,7 @@ model_verbosity = "high"
use_experimental_unified_exec_tool: false,
use_experimental_use_rmcp_client: false,
include_view_image_tool: true,
features: Features::with_defaults(),
active_profile: Some("gpt3".to_string()),
windows_wsl_setup_acknowledged: false,
disable_paste_burst: false,
@@ -2261,6 +2391,7 @@ model_verbosity = "high"
use_experimental_unified_exec_tool: false,
use_experimental_use_rmcp_client: false,
include_view_image_tool: true,
features: Features::with_defaults(),
active_profile: Some("zdr".to_string()),
windows_wsl_setup_acknowledged: false,
disable_paste_burst: false,
@@ -2325,6 +2456,7 @@ model_verbosity = "high"
use_experimental_unified_exec_tool: false,
use_experimental_use_rmcp_client: false,
include_view_image_tool: true,
features: Features::with_defaults(),
active_profile: Some("gpt5".to_string()),
windows_wsl_setup_acknowledged: false,
disable_paste_burst: false,

View File

@@ -20,6 +20,18 @@ pub struct ConfigProfile {
pub model_verbosity: Option<Verbosity>,
pub chatgpt_base_url: Option<String>,
pub experimental_instructions_file: Option<PathBuf>,
pub include_plan_tool: Option<bool>,
pub include_apply_patch_tool: Option<bool>,
pub include_view_image_tool: Option<bool>,
pub experimental_use_unified_exec_tool: Option<bool>,
pub experimental_use_exec_command_tool: Option<bool>,
pub experimental_use_rmcp_client: Option<bool>,
pub experimental_use_freeform_apply_patch: Option<bool>,
pub tools_web_search: Option<bool>,
pub tools_view_image: Option<bool>,
/// Optional feature toggles scoped to this profile.
#[serde(default)]
pub features: Option<crate::features::FeaturesToml>,
}
impl From<ConfigProfile> for codex_app_server_protocol::Profile {

View File

@@ -6,6 +6,7 @@ use async_trait::async_trait;
use crate::CODEX_APPLY_PATCH_ARG1;
use crate::apply_patch::ApplyPatchExec;
use crate::exec::ExecParams;
use crate::executor::ExecutorConfig;
use crate::function_tool::FunctionCallError;
pub(crate) enum ExecutionMode {
@@ -22,6 +23,7 @@ pub(crate) trait ExecutionBackend: Send + Sync {
params: ExecParams,
// Required for downcasting the apply_patch.
mode: &ExecutionMode,
config: &ExecutorConfig,
) -> Result<ExecParams, FunctionCallError>;
fn stream_stdout(&self, _mode: &ExecutionMode) -> bool {
@@ -47,6 +49,7 @@ impl ExecutionBackend for ShellBackend {
&self,
params: ExecParams,
mode: &ExecutionMode,
_config: &ExecutorConfig,
) -> Result<ExecParams, FunctionCallError> {
match mode {
ExecutionMode::Shell => Ok(params),
@@ -65,17 +68,22 @@ impl ExecutionBackend for ApplyPatchBackend {
&self,
params: ExecParams,
mode: &ExecutionMode,
config: &ExecutorConfig,
) -> 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 path_to_codex = if let Some(exe_path) = &config.codex_exe {
exe_path.to_string_lossy().to_string()
} else {
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 {

View File

@@ -30,19 +30,19 @@ use codex_otel::otel_event_manager::ToolDecisionSource;
pub(crate) struct ExecutorConfig {
pub(crate) sandbox_policy: SandboxPolicy,
pub(crate) sandbox_cwd: PathBuf,
codex_linux_sandbox_exe: Option<PathBuf>,
pub(crate) codex_exe: Option<PathBuf>,
}
impl ExecutorConfig {
pub(crate) fn new(
sandbox_policy: SandboxPolicy,
sandbox_cwd: PathBuf,
codex_linux_sandbox_exe: Option<PathBuf>,
codex_exe: Option<PathBuf>,
) -> Self {
Self {
sandbox_policy,
sandbox_cwd,
codex_linux_sandbox_exe,
codex_exe,
}
}
}
@@ -86,7 +86,14 @@ impl Executor {
maybe_translate_shell_command(request.params, session, request.use_shell_profile);
}
// Step 1: Normalise parameters via the selected backend.
// Step 1: Snapshot sandbox configuration so it stays stable for this run.
let config = self
.config
.read()
.map_err(|_| ExecError::rejection("executor config poisoned"))?
.clone();
// Step 2: 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()
@@ -94,16 +101,9 @@ impl Executor {
None
};
request.params = backend
.prepare(request.params, &request.mode)
.prepare(request.params, &request.mode, &config)
.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,
@@ -227,7 +227,7 @@ impl Executor {
sandbox,
&config.sandbox_policy,
&config.sandbox_cwd,
&config.codex_linux_sandbox_exe,
&config.codex_exe,
stdout_stream,
)
.await

View File

@@ -0,0 +1,258 @@
//! Centralized feature flags and metadata.
//!
//! This module defines a small set of toggles that gate experimental and
//! optional behavior across the codebase. Instead of wiring individual
//! booleans through multiple types, call sites consult a single `Features`
//! container attached to `Config`.
use crate::config::ConfigToml;
use crate::config_profile::ConfigProfile;
use serde::Deserialize;
use std::collections::BTreeMap;
use std::collections::BTreeSet;
mod legacy;
pub(crate) use legacy::LegacyFeatureToggles;
/// High-level lifecycle stage for a feature.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum Stage {
Experimental,
Beta,
Stable,
Deprecated,
Removed,
}
/// Unique features toggled via configuration.
#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash)]
pub enum Feature {
/// Use the single unified PTY-backed exec tool.
UnifiedExec,
/// Use the streamable exec-command/write-stdin tool pair.
StreamableShell,
/// Use the official Rust MCP client (rmcp).
RmcpClient,
/// Include the plan tool.
PlanTool,
/// Include the freeform apply_patch tool.
ApplyPatchFreeform,
/// Include the view_image tool.
ViewImageTool,
/// Allow the model to request web searches.
WebSearchRequest,
/// Automatically approve all approval requests from the harness.
ApproveAll,
}
impl Feature {
pub fn key(self) -> &'static str {
self.info().key
}
pub fn stage(self) -> Stage {
self.info().stage
}
pub fn default_enabled(self) -> bool {
self.info().default_enabled
}
fn info(self) -> &'static FeatureSpec {
FEATURES
.iter()
.find(|spec| spec.id == self)
.unwrap_or_else(|| unreachable!("missing FeatureSpec for {:?}", self))
}
}
/// Holds the effective set of enabled features.
#[derive(Debug, Clone, Default, PartialEq)]
pub struct Features {
enabled: BTreeSet<Feature>,
}
#[derive(Debug, Clone, Default)]
pub struct FeatureOverrides {
pub include_plan_tool: Option<bool>,
pub include_apply_patch_tool: Option<bool>,
pub include_view_image_tool: Option<bool>,
pub web_search_request: Option<bool>,
}
impl FeatureOverrides {
fn apply(self, features: &mut Features) {
LegacyFeatureToggles {
include_plan_tool: self.include_plan_tool,
include_apply_patch_tool: self.include_apply_patch_tool,
include_view_image_tool: self.include_view_image_tool,
tools_web_search: self.web_search_request,
..Default::default()
}
.apply(features);
}
}
impl Features {
/// Starts with built-in defaults.
pub fn with_defaults() -> Self {
let mut set = BTreeSet::new();
for spec in FEATURES {
if spec.default_enabled {
set.insert(spec.id);
}
}
Self { enabled: set }
}
pub fn enabled(&self, f: Feature) -> bool {
self.enabled.contains(&f)
}
pub fn enable(&mut self, f: Feature) {
self.enabled.insert(f);
}
pub fn disable(&mut self, f: Feature) {
self.enabled.remove(&f);
}
/// Apply a table of key -> bool toggles (e.g. from TOML).
pub fn apply_map(&mut self, m: &BTreeMap<String, bool>) {
for (k, v) in m {
match feature_for_key(k) {
Some(feat) => {
if *v {
self.enable(feat);
} else {
self.disable(feat);
}
}
None => {
tracing::warn!("unknown feature key in config: {k}");
}
}
}
}
pub fn from_config(
cfg: &ConfigToml,
config_profile: &ConfigProfile,
overrides: FeatureOverrides,
) -> Self {
let mut features = Features::with_defaults();
let base_legacy = LegacyFeatureToggles {
experimental_use_freeform_apply_patch: cfg.experimental_use_freeform_apply_patch,
experimental_use_exec_command_tool: cfg.experimental_use_exec_command_tool,
experimental_use_unified_exec_tool: cfg.experimental_use_unified_exec_tool,
experimental_use_rmcp_client: cfg.experimental_use_rmcp_client,
tools_web_search: cfg.tools.as_ref().and_then(|t| t.web_search),
tools_view_image: cfg.tools.as_ref().and_then(|t| t.view_image),
..Default::default()
};
base_legacy.apply(&mut features);
if let Some(base_features) = cfg.features.as_ref() {
features.apply_map(&base_features.entries);
}
let profile_legacy = LegacyFeatureToggles {
include_plan_tool: config_profile.include_plan_tool,
include_apply_patch_tool: config_profile.include_apply_patch_tool,
include_view_image_tool: config_profile.include_view_image_tool,
experimental_use_freeform_apply_patch: config_profile
.experimental_use_freeform_apply_patch,
experimental_use_exec_command_tool: config_profile.experimental_use_exec_command_tool,
experimental_use_unified_exec_tool: config_profile.experimental_use_unified_exec_tool,
experimental_use_rmcp_client: config_profile.experimental_use_rmcp_client,
tools_web_search: config_profile.tools_web_search,
tools_view_image: config_profile.tools_view_image,
};
profile_legacy.apply(&mut features);
if let Some(profile_features) = config_profile.features.as_ref() {
features.apply_map(&profile_features.entries);
}
overrides.apply(&mut features);
features
}
}
/// Keys accepted in `[features]` tables.
fn feature_for_key(key: &str) -> Option<Feature> {
for spec in FEATURES {
if spec.key == key {
return Some(spec.id);
}
}
legacy::feature_for_key(key)
}
/// Deserializable features table for TOML.
#[derive(Deserialize, Debug, Clone, Default, PartialEq)]
pub struct FeaturesToml {
#[serde(flatten)]
pub entries: BTreeMap<String, bool>,
}
/// Single, easy-to-read registry of all feature definitions.
#[derive(Debug, Clone, Copy)]
pub struct FeatureSpec {
pub id: Feature,
pub key: &'static str,
pub stage: Stage,
pub default_enabled: bool,
}
pub const FEATURES: &[FeatureSpec] = &[
FeatureSpec {
id: Feature::UnifiedExec,
key: "unified_exec",
stage: Stage::Experimental,
default_enabled: false,
},
FeatureSpec {
id: Feature::StreamableShell,
key: "streamable_shell",
stage: Stage::Experimental,
default_enabled: false,
},
FeatureSpec {
id: Feature::RmcpClient,
key: "rmcp_client",
stage: Stage::Experimental,
default_enabled: false,
},
FeatureSpec {
id: Feature::PlanTool,
key: "plan_tool",
stage: Stage::Stable,
default_enabled: false,
},
FeatureSpec {
id: Feature::ApplyPatchFreeform,
key: "apply_patch_freeform",
stage: Stage::Beta,
default_enabled: false,
},
FeatureSpec {
id: Feature::ViewImageTool,
key: "view_image_tool",
stage: Stage::Stable,
default_enabled: true,
},
FeatureSpec {
id: Feature::WebSearchRequest,
key: "web_search_request",
stage: Stage::Stable,
default_enabled: false,
},
FeatureSpec {
id: Feature::ApproveAll,
key: "approve_all",
stage: Stage::Experimental,
default_enabled: false,
},
];

View File

@@ -0,0 +1,158 @@
use super::Feature;
use super::Features;
use tracing::info;
#[derive(Clone, Copy)]
struct Alias {
legacy_key: &'static str,
feature: Feature,
}
const ALIASES: &[Alias] = &[
Alias {
legacy_key: "experimental_use_unified_exec_tool",
feature: Feature::UnifiedExec,
},
Alias {
legacy_key: "experimental_use_exec_command_tool",
feature: Feature::StreamableShell,
},
Alias {
legacy_key: "experimental_use_rmcp_client",
feature: Feature::RmcpClient,
},
Alias {
legacy_key: "experimental_use_freeform_apply_patch",
feature: Feature::ApplyPatchFreeform,
},
Alias {
legacy_key: "include_apply_patch_tool",
feature: Feature::ApplyPatchFreeform,
},
Alias {
legacy_key: "include_plan_tool",
feature: Feature::PlanTool,
},
Alias {
legacy_key: "include_view_image_tool",
feature: Feature::ViewImageTool,
},
Alias {
legacy_key: "web_search",
feature: Feature::WebSearchRequest,
},
];
pub(crate) fn feature_for_key(key: &str) -> Option<Feature> {
ALIASES
.iter()
.find(|alias| alias.legacy_key == key)
.map(|alias| {
log_alias(alias.legacy_key, alias.feature);
alias.feature
})
}
#[derive(Debug, Default)]
pub struct LegacyFeatureToggles {
pub include_plan_tool: Option<bool>,
pub include_apply_patch_tool: Option<bool>,
pub include_view_image_tool: Option<bool>,
pub experimental_use_freeform_apply_patch: Option<bool>,
pub experimental_use_exec_command_tool: Option<bool>,
pub experimental_use_unified_exec_tool: Option<bool>,
pub experimental_use_rmcp_client: Option<bool>,
pub tools_web_search: Option<bool>,
pub tools_view_image: Option<bool>,
}
impl LegacyFeatureToggles {
pub fn apply(self, features: &mut Features) {
set_if_some(
features,
Feature::PlanTool,
self.include_plan_tool,
"include_plan_tool",
);
set_if_some(
features,
Feature::ApplyPatchFreeform,
self.include_apply_patch_tool,
"include_apply_patch_tool",
);
set_if_some(
features,
Feature::ApplyPatchFreeform,
self.experimental_use_freeform_apply_patch,
"experimental_use_freeform_apply_patch",
);
set_if_some(
features,
Feature::StreamableShell,
self.experimental_use_exec_command_tool,
"experimental_use_exec_command_tool",
);
set_if_some(
features,
Feature::UnifiedExec,
self.experimental_use_unified_exec_tool,
"experimental_use_unified_exec_tool",
);
set_if_some(
features,
Feature::RmcpClient,
self.experimental_use_rmcp_client,
"experimental_use_rmcp_client",
);
set_if_some(
features,
Feature::WebSearchRequest,
self.tools_web_search,
"tools.web_search",
);
set_if_some(
features,
Feature::ViewImageTool,
self.include_view_image_tool,
"include_view_image_tool",
);
set_if_some(
features,
Feature::ViewImageTool,
self.tools_view_image,
"tools.view_image",
);
}
}
fn set_if_some(
features: &mut Features,
feature: Feature,
maybe_value: Option<bool>,
alias_key: &'static str,
) {
if let Some(enabled) = maybe_value {
set_feature(features, feature, enabled);
log_alias(alias_key, feature);
}
}
fn set_feature(features: &mut Features, feature: Feature, enabled: bool) {
if enabled {
features.enable(feature);
} else {
features.disable(feature);
}
}
fn log_alias(alias: &str, feature: Feature) {
let canonical = feature.key();
if alias == canonical {
return;
}
info!(
%alias,
canonical,
"legacy feature toggle detected; prefer `[features].{canonical}`"
);
}

View File

@@ -29,6 +29,7 @@ pub mod exec;
mod exec_command;
pub mod exec_env;
pub mod executor;
pub mod features;
mod flags;
pub mod git_info;
pub mod landlock;

View File

@@ -1,44 +1,9 @@
use crate::bash::try_parse_bash;
use crate::bash::try_parse_word_only_commands_sequence;
use serde::Deserialize;
use serde::Serialize;
use codex_protocol::parse_command::ParsedCommand;
use shlex::split as shlex_split;
use shlex::try_join as shlex_try_join;
#[derive(Debug, Clone, PartialEq, Eq, Deserialize, Serialize)]
pub enum ParsedCommand {
Read {
cmd: String,
name: String,
},
ListFiles {
cmd: String,
path: Option<String>,
},
Search {
cmd: String,
query: Option<String>,
path: Option<String>,
},
Unknown {
cmd: String,
},
}
// Convert core's parsed command enum into the protocol's simplified type so
// events can carry the canonical representation across process boundaries.
impl From<ParsedCommand> for codex_protocol::parse_command::ParsedCommand {
fn from(v: ParsedCommand) -> Self {
use codex_protocol::parse_command::ParsedCommand as P;
match v {
ParsedCommand::Read { cmd, name } => P::Read { cmd, name },
ParsedCommand::ListFiles { cmd, path } => P::ListFiles { cmd, path },
ParsedCommand::Search { cmd, query, path } => P::Search { cmd, query, path },
ParsedCommand::Unknown { cmd } => P::Unknown { cmd },
}
}
}
fn shlex_join(tokens: &[String]) -> String {
shlex_try_join(tokens.iter().map(String::as_str))
.unwrap_or_else(|_| "<command included NUL byte>".to_string())

View File

@@ -21,6 +21,8 @@ use tracing::error;
/// Default filename scanned for project-level docs.
pub const DEFAULT_PROJECT_DOC_FILENAME: &str = "AGENTS.md";
/// Preferred local override for project-level docs.
pub const LOCAL_PROJECT_DOC_FILENAME: &str = "AGENTS.override.md";
/// When both `Config::instructions` and the project doc are present, they will
/// be concatenated with the following separator.
@@ -178,7 +180,8 @@ pub fn discover_project_doc_paths(config: &Config) -> std::io::Result<Vec<PathBu
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());
Vec::with_capacity(2 + config.project_doc_fallback_filenames.len());
names.push(LOCAL_PROJECT_DOC_FILENAME);
names.push(DEFAULT_PROJECT_DOC_FILENAME);
for candidate in &config.project_doc_fallback_filenames {
let candidate = candidate.as_str();
@@ -381,6 +384,29 @@ mod tests {
assert_eq!(res, "root doc\n\ncrate doc");
}
/// AGENTS.override.md is preferred over AGENTS.md when both are present.
#[tokio::test]
async fn agents_local_md_preferred() {
let tmp = tempfile::tempdir().expect("tempdir");
fs::write(tmp.path().join(DEFAULT_PROJECT_DOC_FILENAME), "versioned").unwrap();
fs::write(tmp.path().join(LOCAL_PROJECT_DOC_FILENAME), "local").unwrap();
let cfg = make_config(&tmp, 4096, None);
let res = get_user_instructions(&cfg)
.await
.expect("local doc expected");
assert_eq!(res, "local");
let discovery = discover_project_doc_paths(&cfg).expect("discover paths");
assert_eq!(discovery.len(), 1);
assert_eq!(
discovery[0].file_name().unwrap().to_string_lossy(),
LOCAL_PROJECT_DOC_FILENAME
);
}
/// When AGENTS.md is absent but a configured fallback exists, the fallback is used.
#[tokio::test]
async fn uses_configured_fallback_when_agents_missing() {

View File

@@ -34,6 +34,16 @@ pub(crate) enum TaskKind {
Compact,
}
impl TaskKind {
pub(crate) fn header_value(self) -> &'static str {
match self {
TaskKind::Regular => "standard",
TaskKind::Review => "review",
TaskKind::Compact => "compact",
}
}
}
#[derive(Clone)]
pub(crate) struct RunningTask {
pub(crate) handle: AbortHandle,
@@ -113,3 +123,15 @@ impl ActiveTurn {
}
}
}
#[cfg(test)]
mod tests {
use super::TaskKind;
#[test]
fn header_value_matches_expected_labels() {
assert_eq!(TaskKind::Regular.header_value(), "standard");
assert_eq!(TaskKind::Review.header_value(), "review");
assert_eq!(TaskKind::Compact.header_value(), "compact");
}
}

View File

@@ -27,6 +27,6 @@ impl SessionTask for RegularTask {
input: Vec<InputItem>,
) -> Option<String> {
let sess = session.clone_session();
run_task(sess, ctx, sub_id, input).await
run_task(sess, ctx, sub_id, input, TaskKind::Regular).await
}
}

View File

@@ -28,7 +28,7 @@ impl SessionTask for ReviewTask {
input: Vec<InputItem>,
) -> Option<String> {
let sess = session.clone_session();
run_task(sess, ctx, sub_id, input).await
run_task(sess, ctx, sub_id, input, TaskKind::Review).await
}
async fn abort(&self, session: Arc<SessionTaskContext>, sub_id: &str) {

View File

@@ -1,5 +1,7 @@
use crate::client_common::tools::ResponsesApiTool;
use crate::client_common::tools::ToolSpec;
use crate::features::Feature;
use crate::features::Features;
use crate::model_family::ModelFamily;
use crate::tools::handlers::PLAN_TOOL;
use crate::tools::handlers::apply_patch::ApplyPatchToolType;
@@ -33,26 +35,23 @@ pub(crate) struct ToolsConfig {
pub(crate) struct ToolsConfigParams<'a> {
pub(crate) model_family: &'a ModelFamily,
pub(crate) include_plan_tool: bool,
pub(crate) include_apply_patch_tool: bool,
pub(crate) include_web_search_request: bool,
pub(crate) use_streamable_shell_tool: bool,
pub(crate) include_view_image_tool: bool,
pub(crate) experimental_unified_exec_tool: bool,
pub(crate) features: &'a Features,
}
impl ToolsConfig {
pub fn new(params: &ToolsConfigParams) -> Self {
let ToolsConfigParams {
model_family,
include_plan_tool,
include_apply_patch_tool,
include_web_search_request,
use_streamable_shell_tool,
include_view_image_tool,
experimental_unified_exec_tool,
features,
} = params;
let shell_type = if *use_streamable_shell_tool {
let use_streamable_shell_tool = features.enabled(Feature::StreamableShell);
let experimental_unified_exec_tool = features.enabled(Feature::UnifiedExec);
let include_plan_tool = features.enabled(Feature::PlanTool);
let include_apply_patch_tool = features.enabled(Feature::ApplyPatchFreeform);
let include_web_search_request = features.enabled(Feature::WebSearchRequest);
let include_view_image_tool = features.enabled(Feature::ViewImageTool);
let shell_type = if use_streamable_shell_tool {
ConfigShellToolType::Streamable
} else if model_family.uses_local_shell_tool {
ConfigShellToolType::Local
@@ -64,7 +63,7 @@ impl ToolsConfig {
Some(ApplyPatchToolType::Freeform) => Some(ApplyPatchToolType::Freeform),
Some(ApplyPatchToolType::Function) => Some(ApplyPatchToolType::Function),
None => {
if *include_apply_patch_tool {
if include_apply_patch_tool {
Some(ApplyPatchToolType::Freeform)
} else {
None
@@ -74,11 +73,11 @@ impl ToolsConfig {
Self {
shell_type,
plan_tool: *include_plan_tool,
plan_tool: include_plan_tool,
apply_patch_tool_type,
web_search_request: *include_web_search_request,
include_view_image_tool: *include_view_image_tool,
experimental_unified_exec_tool: *experimental_unified_exec_tool,
web_search_request: include_web_search_request,
include_view_image_tool,
experimental_unified_exec_tool,
experimental_supported_tools: model_family.experimental_supported_tools.clone(),
}
}
@@ -906,14 +905,13 @@ mod tests {
fn test_build_specs() {
let model_family = find_family_for_model("codex-mini-latest")
.expect("codex-mini-latest should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::PlanTool);
features.enable(Feature::WebSearchRequest);
features.enable(Feature::UnifiedExec);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: true,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(&config, Some(HashMap::new())).build();
@@ -926,14 +924,13 @@ mod tests {
#[test]
fn test_build_specs_default_shell() {
let model_family = find_family_for_model("o3").expect("o3 should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::PlanTool);
features.enable(Feature::WebSearchRequest);
features.enable(Feature::UnifiedExec);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: true,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(&config, Some(HashMap::new())).build();
@@ -948,14 +945,12 @@ mod tests {
fn test_parallel_support_flags() {
let model_family = find_family_for_model("gpt-5-codex")
.expect("codex-mini-latest should be a valid model family");
let mut features = Features::with_defaults();
features.disable(Feature::ViewImageTool);
features.enable(Feature::UnifiedExec);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: false,
use_streamable_shell_tool: false,
include_view_image_tool: false,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(&config, None).build();
@@ -969,14 +964,11 @@ mod tests {
fn test_test_model_family_includes_sync_tool() {
let model_family = find_family_for_model("test-gpt-5-codex")
.expect("test-gpt-5-codex should be a valid model family");
let mut features = Features::with_defaults();
features.disable(Feature::ViewImageTool);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: false,
use_streamable_shell_tool: false,
include_view_image_tool: false,
experimental_unified_exec_tool: false,
features: &features,
});
let (tools, _) = build_specs(&config, None).build();
@@ -1001,14 +993,12 @@ mod tests {
#[test]
fn test_build_specs_mcp_tools() {
let model_family = find_family_for_model("o3").expect("o3 should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
features.enable(Feature::WebSearchRequest);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(
&config,
@@ -1106,14 +1096,11 @@ mod tests {
#[test]
fn test_build_specs_mcp_tools_sorted_by_name() {
let model_family = find_family_for_model("o3").expect("o3 should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: false,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
// Intentionally construct a map with keys that would sort alphabetically.
@@ -1183,14 +1170,12 @@ mod tests {
fn test_mcp_tool_property_missing_type_defaults_to_string() {
let model_family = find_family_for_model("gpt-5-codex")
.expect("gpt-5-codex should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
features.enable(Feature::WebSearchRequest);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(
@@ -1252,14 +1237,12 @@ mod tests {
fn test_mcp_tool_integer_normalized_to_number() {
let model_family = find_family_for_model("gpt-5-codex")
.expect("gpt-5-codex should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
features.enable(Feature::WebSearchRequest);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(
@@ -1316,14 +1299,13 @@ mod tests {
fn test_mcp_tool_array_without_items_gets_default_string_items() {
let model_family = find_family_for_model("gpt-5-codex")
.expect("gpt-5-codex should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
features.enable(Feature::WebSearchRequest);
features.enable(Feature::ApplyPatchFreeform);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: true,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(
@@ -1383,14 +1365,12 @@ mod tests {
fn test_mcp_tool_anyof_defaults_to_string() {
let model_family = find_family_for_model("gpt-5-codex")
.expect("gpt-5-codex should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
features.enable(Feature::WebSearchRequest);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(
@@ -1462,14 +1442,12 @@ mod tests {
fn test_get_openai_tools_mcp_tools_with_additional_properties_schema() {
let model_family = find_family_for_model("gpt-5-codex")
.expect("gpt-5-codex should be a valid model family");
let mut features = Features::with_defaults();
features.enable(Feature::UnifiedExec);
features.enable(Feature::WebSearchRequest);
let config = ToolsConfig::new(&ToolsConfigParams {
model_family: &model_family,
include_plan_tool: false,
include_apply_patch_tool: false,
include_web_search_request: true,
use_streamable_shell_tool: false,
include_view_image_tool: true,
experimental_unified_exec_tool: true,
features: &features,
});
let (tools, _) = build_specs(
&config,

View File

@@ -49,6 +49,7 @@ impl UserNotifier {
pub(crate) enum UserNotification {
#[serde(rename_all = "kebab-case")]
AgentTurnComplete {
thread_id: String,
turn_id: String,
/// Messages that the user sent to the agent to initiate the turn.
@@ -67,6 +68,7 @@ mod tests {
#[test]
fn test_user_notification() -> Result<()> {
let notification = UserNotification::AgentTurnComplete {
thread_id: "b5f6c1c2-1111-2222-3333-444455556666".to_string(),
turn_id: "12345".to_string(),
input_messages: vec!["Rename `foo` to `bar` and update the callsites.".to_string()],
last_assistant_message: Some(
@@ -76,7 +78,7 @@ mod tests {
let serialized = serde_json::to_string(&notification)?;
assert_eq!(
serialized,
r#"{"type":"agent-turn-complete","turn-id":"12345","input-messages":["Rename `foo` to `bar` and update the callsites."],"last-assistant-message":"Rename complete and verified `cargo build` succeeds."}"#
r#"{"type":"agent-turn-complete","thread-id":"b5f6c1c2-1111-2222-3333-444455556666","turn-id":"12345","input-messages":["Rename `foo` to `bar` and update the callsites."],"last-assistant-message":"Rename complete and verified `cargo build` succeeds."}"#
);
Ok(())
}

View File

@@ -10,8 +10,10 @@ path = "lib.rs"
anyhow = { workspace = true }
assert_cmd = { workspace = true }
codex-core = { workspace = true }
notify = { workspace = true }
regex-lite = { workspace = true }
serde_json = { workspace = true }
tempfile = { workspace = true }
tokio = { workspace = true, features = ["time"] }
walkdir = { workspace = true }
wiremock = { workspace = true }

View File

@@ -164,6 +164,149 @@ pub fn sandbox_network_env_var() -> &'static str {
codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR
}
pub mod fs_wait {
use anyhow::Result;
use anyhow::anyhow;
use notify::RecursiveMode;
use notify::Watcher;
use std::path::Path;
use std::path::PathBuf;
use std::sync::mpsc;
use std::sync::mpsc::RecvTimeoutError;
use std::time::Duration;
use std::time::Instant;
use tokio::task;
use walkdir::WalkDir;
pub async fn wait_for_path_exists(
path: impl Into<PathBuf>,
timeout: Duration,
) -> Result<PathBuf> {
let path = path.into();
task::spawn_blocking(move || wait_for_path_exists_blocking(path, timeout)).await?
}
pub async fn wait_for_matching_file(
root: impl Into<PathBuf>,
timeout: Duration,
predicate: impl FnMut(&Path) -> bool + Send + 'static,
) -> Result<PathBuf> {
let root = root.into();
task::spawn_blocking(move || {
let mut predicate = predicate;
blocking_find_matching_file(root, timeout, &mut predicate)
})
.await?
}
fn wait_for_path_exists_blocking(path: PathBuf, timeout: Duration) -> Result<PathBuf> {
if path.exists() {
return Ok(path);
}
let watch_root = nearest_existing_ancestor(&path);
let (tx, rx) = mpsc::channel();
let mut watcher = notify::recommended_watcher(move |res| {
let _ = tx.send(res);
})?;
watcher.watch(&watch_root, RecursiveMode::Recursive)?;
let deadline = Instant::now() + timeout;
loop {
if path.exists() {
return Ok(path.clone());
}
let now = Instant::now();
if now >= deadline {
break;
}
let remaining = deadline.saturating_duration_since(now);
match rx.recv_timeout(remaining) {
Ok(Ok(_event)) => {
if path.exists() {
return Ok(path.clone());
}
}
Ok(Err(err)) => return Err(err.into()),
Err(RecvTimeoutError::Timeout) => break,
Err(RecvTimeoutError::Disconnected) => break,
}
}
if path.exists() {
Ok(path)
} else {
Err(anyhow!("timed out waiting for {:?}", path))
}
}
fn blocking_find_matching_file(
root: PathBuf,
timeout: Duration,
predicate: &mut impl FnMut(&Path) -> bool,
) -> Result<PathBuf> {
let root = wait_for_path_exists_blocking(root, timeout)?;
if let Some(found) = scan_for_match(&root, predicate) {
return Ok(found);
}
let (tx, rx) = mpsc::channel();
let mut watcher = notify::recommended_watcher(move |res| {
let _ = tx.send(res);
})?;
watcher.watch(&root, RecursiveMode::Recursive)?;
let deadline = Instant::now() + timeout;
while Instant::now() < deadline {
let remaining = deadline.saturating_duration_since(Instant::now());
match rx.recv_timeout(remaining) {
Ok(Ok(_event)) => {
if let Some(found) = scan_for_match(&root, predicate) {
return Ok(found);
}
}
Ok(Err(err)) => return Err(err.into()),
Err(RecvTimeoutError::Timeout) => break,
Err(RecvTimeoutError::Disconnected) => break,
}
}
if let Some(found) = scan_for_match(&root, predicate) {
Ok(found)
} else {
Err(anyhow!("timed out waiting for matching file in {:?}", root))
}
}
fn scan_for_match(root: &Path, predicate: &mut impl FnMut(&Path) -> bool) -> Option<PathBuf> {
for entry in WalkDir::new(root).into_iter().filter_map(Result::ok) {
let path = entry.path();
if !entry.file_type().is_file() {
continue;
}
if predicate(path) {
return Some(path.to_path_buf());
}
}
None
}
fn nearest_existing_ancestor(path: &Path) -> PathBuf {
let mut current = path;
loop {
if current.exists() {
return current.to_path_buf();
}
match current.parent() {
Some(parent) => current = parent,
None => return PathBuf::from("."),
}
}
}
}
#[macro_export]
macro_rules! skip_if_sandbox {
() => {{

View File

@@ -239,6 +239,20 @@ pub fn ev_apply_patch_function_call(call_id: &str, patch: &str) -> Value {
})
}
pub fn ev_function_call_output(call_id: &str, content: &str) -> Value {
serde_json::json!({
"type": "response.output_item.done",
"item": {
"type": "function_call_output",
"call_id": call_id,
"output": {
"content": content,
"success": true
}
}
})
}
pub fn sse_failed(id: &str, code: &str, message: &str) -> String {
sse(vec![serde_json::json!({
"type": "response.failed",

View File

@@ -1,4 +1,5 @@
use std::mem::swap;
use std::path::PathBuf;
use std::sync::Arc;
use codex_core::CodexAuth;
@@ -39,6 +40,12 @@ impl TestCodexBuilder {
let mut config = load_default_config_for_test(&home);
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.codex_linux_sandbox_exe = Some(PathBuf::from(
assert_cmd::Command::cargo_bin("codex")?
.get_program()
.to_os_string(),
));
let mut mutators = vec![];
swap(&mut self.config_mutators, &mut mutators);

View File

@@ -0,0 +1,102 @@
use std::sync::Arc;
use codex_app_server_protocol::AuthMode;
use codex_core::ContentItem;
use codex_core::ModelClient;
use codex_core::ModelProviderInfo;
use codex_core::Prompt;
use codex_core::ResponseEvent;
use codex_core::ResponseItem;
use codex_core::WireApi;
use codex_otel::otel_event_manager::OtelEventManager;
use codex_protocol::ConversationId;
use core_test_support::load_default_config_for_test;
use core_test_support::responses;
use futures::StreamExt;
use tempfile::TempDir;
use wiremock::matchers::header;
#[tokio::test]
async fn responses_stream_includes_task_type_header() {
core_test_support::skip_if_no_network!();
let server = responses::start_mock_server().await;
let response_body = responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_completed("resp-1"),
]);
let request_recorder = responses::mount_sse_once_match(
&server,
header("Codex-Task-Type", "standard"),
response_body,
)
.await;
let provider = ModelProviderInfo {
name: "mock".into(),
base_url: Some(format!("{}/v1", server.uri())),
env_key: None,
env_key_instructions: None,
wire_api: WireApi::Responses,
query_params: None,
http_headers: None,
env_http_headers: None,
request_max_retries: Some(0),
stream_max_retries: Some(0),
stream_idle_timeout_ms: Some(5_000),
requires_openai_auth: false,
};
let codex_home = TempDir::new().expect("failed to create TempDir");
let mut config = load_default_config_for_test(&codex_home);
config.model_provider_id = provider.name.clone();
config.model_provider = provider.clone();
let effort = config.model_reasoning_effort;
let summary = config.model_reasoning_summary;
let config = Arc::new(config);
let conversation_id = ConversationId::new();
let otel_event_manager = OtelEventManager::new(
conversation_id,
config.model.as_str(),
config.model_family.slug.as_str(),
None,
Some(AuthMode::ChatGPT),
false,
"test".to_string(),
);
let client = ModelClient::new(
Arc::clone(&config),
None,
otel_event_manager,
provider,
effort,
summary,
conversation_id,
);
let mut prompt = Prompt::default();
prompt.input = vec![ResponseItem::Message {
id: None,
role: "user".into(),
content: vec![ContentItem::InputText {
text: "hello".into(),
}],
}];
let mut stream = client.stream(&prompt).await.expect("stream failed");
while let Some(event) = stream.next().await {
if matches!(event, Ok(ResponseEvent::Completed { .. })) {
break;
}
}
let request = request_recorder.single_request();
assert_eq!(
request.header("Codex-Task-Type").as_deref(),
Some("standard")
);
}

View File

@@ -1,12 +1,11 @@
use assert_cmd::Command as AssertCommand;
use codex_core::RolloutRecorder;
use codex_core::protocol::GitInfo;
use core_test_support::fs_wait;
use core_test_support::skip_if_no_network;
use std::time::Duration;
use std::time::Instant;
use tempfile::TempDir;
use uuid::Uuid;
use walkdir::WalkDir;
use wiremock::Mock;
use wiremock::MockServer;
use wiremock::ResponseTemplate;
@@ -211,12 +210,12 @@ async fn responses_api_stream_cli() {
/// End-to-end: create a session (writes rollout), verify the file, then resume and confirm append.
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn integration_creates_and_checks_session_file() {
async fn integration_creates_and_checks_session_file() -> anyhow::Result<()> {
// Honor sandbox network restrictions for CI parity with the other tests.
skip_if_no_network!();
skip_if_no_network!(Ok(()));
// 1. Temp home so we read/write isolated session files.
let home = TempDir::new().unwrap();
let home = TempDir::new()?;
// 2. Unique marker we'll look for in the session log.
let marker = format!("integration-test-{}", Uuid::new_v4());
@@ -254,63 +253,20 @@ async fn integration_creates_and_checks_session_file() {
// Wait for sessions dir to appear.
let sessions_dir = home.path().join("sessions");
let dir_deadline = Instant::now() + Duration::from_secs(5);
while !sessions_dir.exists() && Instant::now() < dir_deadline {
std::thread::sleep(Duration::from_millis(50));
}
assert!(sessions_dir.exists(), "sessions directory never appeared");
fs_wait::wait_for_path_exists(&sessions_dir, Duration::from_secs(5)).await?;
// Find the session file that contains `marker`.
let deadline = Instant::now() + Duration::from_secs(10);
let mut matching_path: Option<std::path::PathBuf> = None;
while Instant::now() < deadline && matching_path.is_none() {
for entry in WalkDir::new(&sessions_dir) {
let entry = match entry {
Ok(e) => e,
Err(_) => continue,
};
if !entry.file_type().is_file() {
continue;
}
if !entry.file_name().to_string_lossy().ends_with(".jsonl") {
continue;
}
let path = entry.path();
let Ok(content) = std::fs::read_to_string(path) else {
continue;
};
let mut lines = content.lines();
if lines.next().is_none() {
continue;
}
for line in lines {
if line.trim().is_empty() {
continue;
}
let item: serde_json::Value = match serde_json::from_str(line) {
Ok(v) => v,
Err(_) => continue,
};
if item.get("type").and_then(|t| t.as_str()) == Some("response_item")
&& let Some(payload) = item.get("payload")
&& payload.get("type").and_then(|t| t.as_str()) == Some("message")
&& let Some(c) = payload.get("content")
&& c.to_string().contains(&marker)
{
matching_path = Some(path.to_path_buf());
break;
}
}
let marker_clone = marker.clone();
let path = fs_wait::wait_for_matching_file(&sessions_dir, Duration::from_secs(10), move |p| {
if p.extension().and_then(|ext| ext.to_str()) != Some("jsonl") {
return false;
}
if matching_path.is_none() {
std::thread::sleep(Duration::from_millis(50));
}
}
let path = match matching_path {
Some(p) => p,
None => panic!("No session file containing the marker was found"),
};
let Ok(content) = std::fs::read_to_string(p) else {
return false;
};
content.contains(&marker_clone)
})
.await?;
// Basic sanity checks on location and metadata.
let rel = match path.strip_prefix(&sessions_dir) {
@@ -418,42 +374,25 @@ async fn integration_creates_and_checks_session_file() {
assert!(output2.status.success(), "resume codex-cli run failed");
// Find the new session file containing the resumed marker.
let deadline = Instant::now() + Duration::from_secs(10);
let mut resumed_path: Option<std::path::PathBuf> = None;
while Instant::now() < deadline && resumed_path.is_none() {
for entry in WalkDir::new(&sessions_dir) {
let entry = match entry {
Ok(e) => e,
Err(_) => continue,
};
if !entry.file_type().is_file() {
continue;
let marker2_clone = marker2.clone();
let resumed_path =
fs_wait::wait_for_matching_file(&sessions_dir, Duration::from_secs(10), move |p| {
if p.extension().and_then(|ext| ext.to_str()) != Some("jsonl") {
return false;
}
if !entry.file_name().to_string_lossy().ends_with(".jsonl") {
continue;
}
let p = entry.path();
let Ok(c) = std::fs::read_to_string(p) else {
continue;
};
if c.contains(&marker2) {
resumed_path = Some(p.to_path_buf());
break;
}
}
if resumed_path.is_none() {
std::thread::sleep(Duration::from_millis(50));
}
}
std::fs::read_to_string(p)
.map(|content| content.contains(&marker2_clone))
.unwrap_or(false)
})
.await?;
let resumed_path = resumed_path.expect("No resumed session file found containing the marker2");
// Resume should write to the existing log file.
assert_eq!(
resumed_path, path,
"resume should create a new session file"
);
let resumed_content = std::fs::read_to_string(&resumed_path).unwrap();
let resumed_content = std::fs::read_to_string(&resumed_path)?;
assert!(
resumed_content.contains(&marker),
"resumed file missing original marker"
@@ -462,6 +401,7 @@ async fn integration_creates_and_checks_session_file() {
resumed_content.contains(&marker2),
"resumed file missing resumed marker"
);
Ok(())
}
/// Integration test to verify git info is collected and recorded in session files.

View File

@@ -19,17 +19,20 @@ use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_completed_with_tokens;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_function_call_output;
use core_test_support::responses::mount_sse_once_match;
use core_test_support::responses::mount_sse_sequence;
use core_test_support::responses::sse;
use core_test_support::responses::sse_failed;
use core_test_support::responses::start_mock_server;
use pretty_assertions::assert_eq;
use serde_json::Value;
// --- Test helpers -----------------------------------------------------------
pub(super) const FIRST_REPLY: &str = "FIRST_REPLY";
pub(super) const SUMMARY_TEXT: &str = "SUMMARY_ONLY_CONTEXT";
const THIRD_USER_MSG: &str = "next turn";
const THIRD_ASSISTANT_MSG: &str = "post compact assistant";
const AUTO_SUMMARY_TEXT: &str = "AUTO_SUMMARY";
const FIRST_AUTO_MSG: &str = "token limit start";
const SECOND_AUTO_MSG: &str = "token limit push";
@@ -644,6 +647,10 @@ async fn manual_compact_retries_after_context_window_error() {
ev_assistant_message("m2", SUMMARY_TEXT),
ev_completed("r2"),
]);
let third_turn = sse(vec![
ev_assistant_message("m3", THIRD_ASSISTANT_MSG),
ev_completed("r3"),
]);
let request_log = mount_sse_sequence(
&server,
@@ -651,6 +658,7 @@ async fn manual_compact_retries_after_context_window_error() {
user_turn.clone(),
compact_failed.clone(),
compact_succeeds.clone(),
third_turn,
],
)
.await;
@@ -688,17 +696,29 @@ async fn manual_compact_retries_after_context_window_error() {
panic!("expected background event after compact retry");
};
assert!(
event.message.contains("Trimmed 1 older conversation item"),
event
.message
.contains("Trimmed 2 older conversation item(s)"),
"background event should mention trimmed item count: {}",
event.message
);
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
.submit(Op::UserInput {
items: vec![InputItem::Text {
text: THIRD_USER_MSG.into(),
}],
})
.await
.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = request_log.requests();
assert_eq!(
requests.len(),
3,
"expected user turn and two compact attempts"
4,
"expected user turn, two compact attempts, and one follow-up turn"
);
let compact_attempt = requests[1].body_json();
@@ -710,42 +730,415 @@ async fn manual_compact_retries_after_context_window_error() {
let retry_input = retry_attempt["input"]
.as_array()
.unwrap_or_else(|| panic!("retry attempt missing input array: {retry_attempt}"));
assert_eq!(
compact_input
.last()
.and_then(|item| item.get("content"))
.and_then(|v| v.as_array())
fn extract_text(item: &Value) -> Option<String> {
item.get("content")
.and_then(Value::as_array)
.and_then(|items| items.first())
.and_then(|entry| entry.get("text"))
.and_then(|text| text.as_str()),
.and_then(Value::as_str)
.map(str::to_string)
}
assert_eq!(
extract_text(compact_input.last().expect("compact input empty")).as_deref(),
Some(SUMMARIZATION_PROMPT),
"compact attempt should include summarization prompt"
"compact attempt should include summarization prompt",
);
assert_eq!(
retry_input
.last()
.and_then(|item| item.get("content"))
.and_then(|v| v.as_array())
.and_then(|items| items.first())
.and_then(|entry| entry.get("text"))
.and_then(|text| text.as_str()),
extract_text(retry_input.last().expect("retry input empty")).as_deref(),
Some(SUMMARIZATION_PROMPT),
"retry attempt should include summarization prompt"
"retry attempt should include summarization prompt",
);
let contains_text = |items: &[Value], needle: &str| {
items
.iter()
.any(|item| extract_text(item).is_some_and(|text| text == needle))
};
assert!(
contains_text(compact_input, "first turn"),
"compact attempt should include original user message",
);
assert!(
contains_text(compact_input, FIRST_REPLY),
"compact attempt should include original assistant reply",
);
assert!(
!contains_text(retry_input, "first turn"),
"retry should drop original user message",
);
assert!(
!contains_text(retry_input, FIRST_REPLY),
"retry should drop assistant reply tied to original user message",
);
assert_eq!(
retry_input.len(),
compact_input.len().saturating_sub(1),
"retry should drop exactly one history item (before {} vs after {})",
compact_input.len().saturating_sub(retry_input.len()),
2,
"retry should drop the most recent user turn (before {} vs after {})",
compact_input.len(),
retry_input.len()
);
if let (Some(first_before), Some(first_after)) = (compact_input.first(), retry_input.first()) {
assert_ne!(
first_before, first_after,
"retry should drop the oldest conversation item"
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn manual_compact_trims_last_user_turn_with_function_calls_on_context_error() {
skip_if_no_network!();
// Scenario 1: ensure the retry trims the most recent turn when function calls are involved.
const FIRST_USER_MSG: &str = "first user turn";
const SECOND_USER_MSG: &str = "second user turn";
const FIRST_CALL_A: &str = "call-first-a";
const FIRST_CALL_B: &str = "call-first-b";
const SECOND_CALL_A: &str = "call-second-a";
const SECOND_CALL_B: &str = "call-second-b";
{
let server = start_mock_server().await;
let first_turn_initial = sse(vec![ev_function_call(FIRST_CALL_A, "tool.first.a", "{}")]);
let first_turn_second_call = sse(vec![
ev_function_call_output(FIRST_CALL_A, "first-call-a output"),
ev_function_call(FIRST_CALL_B, "tool.first.b", "{}"),
]);
let first_turn_complete = sse(vec![
ev_function_call_output(FIRST_CALL_B, "first-call-b output"),
ev_assistant_message("assistant-first", "first turn complete"),
ev_completed("resp-first"),
]);
let second_turn_initial = sse(vec![ev_function_call(SECOND_CALL_A, "tool.second.a", "{}")]);
let second_turn_second_call = sse(vec![
ev_function_call_output(SECOND_CALL_A, "second-call-a output"),
ev_function_call(SECOND_CALL_B, "tool.second.b", "{}"),
]);
let second_turn_complete = sse(vec![
ev_function_call_output(SECOND_CALL_B, "second-call-b output"),
ev_assistant_message("assistant-second", "second turn complete"),
ev_completed("resp-second"),
]);
let compact_failed = sse_failed(
"resp-fail",
"context_length_exceeded",
CONTEXT_LIMIT_MESSAGE,
);
let compact_retry = sse(vec![
ev_assistant_message("assistant-summary", SUMMARY_TEXT),
ev_completed("resp-summary"),
]);
let request_log = mount_sse_sequence(
&server,
vec![
first_turn_initial,
first_turn_second_call,
first_turn_complete,
second_turn_initial,
second_turn_second_call,
second_turn_complete,
compact_failed,
compact_retry,
],
)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&home);
config.model_provider = model_provider;
config.model_auto_compact_token_limit = Some(200_000);
let codex = ConversationManager::with_auth(CodexAuth::from_api_key("dummy"))
.new_conversation(config)
.await
.unwrap()
.conversation;
codex
.submit(Op::UserInput {
items: vec![InputItem::Text {
text: FIRST_USER_MSG.into(),
}],
})
.await
.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
.submit(Op::UserInput {
items: vec![InputItem::Text {
text: SECOND_USER_MSG.into(),
}],
})
.await
.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex.submit(Op::Compact).await.unwrap();
let EventMsg::BackgroundEvent(event) =
wait_for_event(&codex, |ev| matches!(ev, EventMsg::BackgroundEvent(_))).await
else {
panic!("expected background event after compact retry");
};
assert!(
event
.message
.contains("Trimmed 2 older conversation item(s)"),
"background event should report trimming chunked user turn: {}",
event.message
);
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = request_log.requests();
assert_eq!(
requests.len(),
8,
"expected two user turns (with tool call round-trips) followed by compact attempt + retry"
);
let compact_attempt = requests[6].body_json();
let retry_attempt = requests[7].body_json();
fn extract_text(item: &Value) -> Option<String> {
item.get("content")
.and_then(Value::as_array)
.and_then(|items| items.first())
.and_then(|entry| entry.get("text"))
.and_then(Value::as_str)
.map(str::to_string)
}
let contains_text = |items: &[Value], needle: &str| {
items
.iter()
.any(|item| extract_text(item).is_some_and(|text| text == needle))
};
assert!(
contains_text(
compact_attempt["input"].as_array().unwrap(),
SECOND_USER_MSG
),
"initial compact attempt should include most recent user message",
);
assert!(
!contains_text(retry_attempt["input"].as_array().unwrap(), SECOND_USER_MSG),
"retry should drop the most recent user message",
);
assert!(
contains_text(
compact_attempt["input"].as_array().unwrap(),
"second turn complete"
),
"initial compact attempt should include assistant reply for most recent turn",
);
assert!(
!contains_text(
retry_attempt["input"].as_array().unwrap(),
"second turn complete"
),
"retry should drop assistant reply for most recent turn",
);
assert_eq!(
compact_attempt["input"]
.as_array()
.unwrap()
.len()
.saturating_sub(retry_attempt["input"].as_array().unwrap().len()),
2,
"retry should drop the most recent user turn from the prompt",
);
let retry_call_ids: std::collections::HashSet<_> = retry_attempt["input"]
.as_array()
.unwrap()
.iter()
.filter_map(|item| item.get("call_id").and_then(|v| v.as_str()))
.collect();
assert!(
!retry_call_ids.contains(SECOND_CALL_A),
"retry should remove function call {SECOND_CALL_A}"
);
assert!(
!retry_call_ids.contains(SECOND_CALL_B),
"retry should remove function call {SECOND_CALL_B}"
);
}
// Scenario 2: after a retry succeeds, the trimmed turn is restored to history for the next user input.
{
const SIMPLE_FIRST_USER_MSG: &str = "first user turn";
const SIMPLE_FIRST_ASSISTANT_MSG: &str = "first assistant reply";
const SIMPLE_SECOND_USER_MSG: &str = "second user turn";
const SIMPLE_SECOND_ASSISTANT_MSG: &str = "second assistant reply";
const SIMPLE_THIRD_USER_MSG: &str = "post compact user";
const SIMPLE_THIRD_ASSISTANT_MSG: &str = "post compact assistant";
let server = start_mock_server().await;
let first_turn = sse(vec![
ev_assistant_message("assistant-first", SIMPLE_FIRST_ASSISTANT_MSG),
ev_completed("resp-first"),
]);
let second_turn = sse(vec![
ev_assistant_message("assistant-second", SIMPLE_SECOND_ASSISTANT_MSG),
ev_completed("resp-second"),
]);
let compact_failed = sse_failed(
"resp-fail",
"context_length_exceeded",
CONTEXT_LIMIT_MESSAGE,
);
let compact_retry = sse(vec![
ev_assistant_message("assistant-summary", SUMMARY_TEXT),
ev_completed("resp-summary"),
]);
let third_turn = sse(vec![
ev_assistant_message("assistant-third", SIMPLE_THIRD_ASSISTANT_MSG),
ev_completed("resp-third"),
]);
let request_log = mount_sse_sequence(
&server,
vec![
first_turn,
second_turn,
compact_failed,
compact_retry,
third_turn,
],
)
.await;
let model_provider = ModelProviderInfo {
base_url: Some(format!("{}/v1", server.uri())),
..built_in_model_providers()["openai"].clone()
};
let home = TempDir::new().unwrap();
let mut config = load_default_config_for_test(&home);
config.model_provider = model_provider;
config.model_auto_compact_token_limit = Some(200_000);
let codex = ConversationManager::with_auth(CodexAuth::from_api_key("dummy"))
.new_conversation(config)
.await
.unwrap()
.conversation;
codex
.submit(Op::UserInput {
items: vec![InputItem::Text {
text: SIMPLE_FIRST_USER_MSG.into(),
}],
})
.await
.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
.submit(Op::UserInput {
items: vec![InputItem::Text {
text: SIMPLE_SECOND_USER_MSG.into(),
}],
})
.await
.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex.submit(Op::Compact).await.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
codex
.submit(Op::UserInput {
items: vec![InputItem::Text {
text: SIMPLE_THIRD_USER_MSG.into(),
}],
})
.await
.unwrap();
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
let requests = request_log.requests();
assert_eq!(
requests.len(),
5,
"expected two user turns, two compact attempts, and a post-compact turn",
);
let retry_request = &requests[3];
let retry_body = retry_request.body_json();
let retry_input = retry_body
.get("input")
.and_then(Value::as_array)
.expect("retry request missing input array");
assert!(
retry_input.iter().all(|item| {
item.get("content")
.and_then(Value::as_array)
.and_then(|entries| entries.first())
.and_then(|entry| entry.get("text"))
.and_then(Value::as_str)
.map(|text| {
text != SIMPLE_SECOND_USER_MSG && text != SIMPLE_SECOND_ASSISTANT_MSG
})
.unwrap_or(true)
}),
"retry compact input should omit trimmed second turn",
);
let final_request = &requests[4];
let body = final_request.body_json();
let input_items = body
.get("input")
.and_then(Value::as_array)
.expect("final request missing input array");
fn message_index(items: &[Value], needle: &str) -> Option<usize> {
items.iter().position(|item| {
item.get("type").and_then(Value::as_str) == Some("message")
&& item
.get("content")
.and_then(Value::as_array)
.and_then(|entries| entries.first())
.and_then(|entry| entry.get("text"))
.and_then(Value::as_str)
.is_some_and(|text| text == needle)
})
}
let summary_index = input_items
.iter()
.position(|item| {
item.get("content")
.and_then(Value::as_array)
.and_then(|entries| entries.first())
.and_then(|entry| entry.get("text"))
.and_then(Value::as_str)
.map(|text| text.contains(SUMMARY_TEXT))
.unwrap_or(false)
})
.expect("final request should include summary bridge");
let second_user_index = message_index(input_items, SIMPLE_SECOND_USER_MSG)
.expect("trimmed second user message should remain in history");
let second_assistant_index = message_index(input_items, SIMPLE_SECOND_ASSISTANT_MSG)
.expect("trimmed assistant reply should remain in history");
let third_user_index = message_index(input_items, SIMPLE_THIRD_USER_MSG)
.expect("post-compact user turn should be present");
assert!(
summary_index < second_user_index,
"summary bridge should precede restored user message"
);
assert!(
second_user_index < second_assistant_index,
"restored user message should precede assistant reply"
);
assert!(
second_assistant_index < third_user_index,
"restored assistant reply should precede new user turn"
);
} else {
panic!("expected non-empty compact inputs");
}
}

View File

@@ -4,6 +4,7 @@ use codex_core::CodexAuth;
use codex_core::ConversationManager;
use codex_core::ModelProviderInfo;
use codex_core::built_in_model_providers;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::EventMsg;
use codex_core::protocol::InputItem;
@@ -56,12 +57,12 @@ async fn collect_tool_identifiers_for_model(model: &str) -> Vec<String> {
config.model = model.to_string();
config.model_family =
find_family_for_model(model).unwrap_or_else(|| panic!("unknown model family for {model}"));
config.include_plan_tool = false;
config.include_apply_patch_tool = false;
config.include_view_image_tool = false;
config.tools_web_search_request = false;
config.use_experimental_streamable_shell_tool = false;
config.use_experimental_unified_exec_tool = false;
config.features.disable(Feature::PlanTool);
config.features.disable(Feature::ApplyPatchFreeform);
config.features.disable(Feature::ViewImageTool);
config.features.disable(Feature::WebSearchRequest);
config.features.disable(Feature::StreamableShell);
config.features.disable(Feature::UnifiedExec);
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));

View File

@@ -5,6 +5,7 @@ use codex_core::ConversationManager;
use codex_core::ModelProviderInfo;
use codex_core::built_in_model_providers;
use codex_core::config::OPENAI_DEFAULT_MODEL;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
@@ -99,10 +100,10 @@ async fn codex_mini_latest_tools() {
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
config.features.disable(Feature::ApplyPatchFreeform);
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));
config.include_apply_patch_tool = false;
config.model = "codex-mini-latest".to_string();
config.model_family = find_family_for_model("codex-mini-latest").unwrap();
@@ -185,7 +186,7 @@ async fn prompt_tools_are_consistent_across_requests() {
config.cwd = cwd.path().to_path_buf();
config.model_provider = model_provider;
config.user_instructions = Some("be consistent and helpful".to_string());
config.include_plan_tool = true;
config.features.enable(Feature::PlanTool);
let conversation_manager =
ConversationManager::with_auth(CodexAuth::from_api_key("Test API Key"));

View File

@@ -9,6 +9,7 @@ use std::time::UNIX_EPOCH;
use codex_core::config_types::McpServerConfig;
use codex_core::config_types::McpServerTransportConfig;
use codex_core::features::Feature;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
@@ -74,7 +75,7 @@ async fn stdio_server_round_trip() -> anyhow::Result<()> {
let fixture = test_codex()
.with_config(move |config| {
config.use_experimental_use_rmcp_client = true;
config.features.enable(Feature::RmcpClient);
config.mcp_servers.insert(
server_name.to_string(),
McpServerConfig {
@@ -227,7 +228,7 @@ async fn streamable_http_tool_call_round_trip() -> anyhow::Result<()> {
let fixture = test_codex()
.with_config(move |config| {
config.use_experimental_use_rmcp_client = true;
config.features.enable(Feature::RmcpClient);
config.mcp_servers.insert(
server_name.to_string(),
McpServerConfig {
@@ -408,7 +409,7 @@ async fn streamable_http_with_oauth_round_trip() -> anyhow::Result<()> {
let fixture = test_codex()
.with_config(move |config| {
config.use_experimental_use_rmcp_client = true;
config.features.enable(Feature::RmcpClient);
config.mcp_servers.insert(
server_name.to_string(),
McpServerConfig {

View File

@@ -1,6 +1,7 @@
#![cfg(not(target_os = "windows"))]
use anyhow::Result;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
@@ -9,9 +10,12 @@ use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::config_types::ReasoningSummary;
use core_test_support::assert_regex_match;
use core_test_support::responses::ev_apply_patch_function_call;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_custom_tool_call;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_local_shell_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::mount_sse_sequence;
use core_test_support::responses::sse;
@@ -20,8 +24,11 @@ use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
use core_test_support::test_codex::test_codex;
use core_test_support::wait_for_event;
use pretty_assertions::assert_eq;
use regex_lite::Regex;
use serde_json::Value;
use serde_json::json;
use std::fs;
async fn submit_turn(test: &TestCodex, prompt: &str, sandbox_policy: SandboxPolicy) -> Result<()> {
let session_model = test.session_configured.model.clone();
@@ -71,13 +78,28 @@ fn find_function_call_output<'a>(bodies: &'a [Value], call_id: &str) -> Option<&
None
}
fn find_custom_tool_call_output<'a>(bodies: &'a [Value], call_id: &str) -> Option<&'a Value> {
for body in bodies {
if let Some(items) = body.get("input").and_then(Value::as_array) {
for item in items {
if item.get("type").and_then(Value::as_str) == Some("custom_tool_call_output")
&& item.get("call_id").and_then(Value::as_str) == Some(call_id)
{
return Some(item);
}
}
}
}
None
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn shell_output_stays_json_without_freeform_apply_patch() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = false;
config.features.disable(Feature::ApplyPatchFreeform);
config.model = "gpt-5".to_string();
config.model_family = find_family_for_model("gpt-5").expect("gpt-5 is a model family");
});
@@ -119,7 +141,12 @@ async fn shell_output_stays_json_without_freeform_apply_patch() -> Result<()> {
.and_then(Value::as_str)
.expect("shell output string");
let parsed: Value = serde_json::from_str(output)?;
let mut parsed: Value = serde_json::from_str(output)?;
if let Some(metadata) = parsed.get_mut("metadata").and_then(Value::as_object_mut) {
// duration_seconds is non-deterministic; remove it for deep equality
let _ = metadata.remove("duration_seconds");
}
assert_eq!(
parsed
.get("metadata")
@@ -143,7 +170,7 @@ async fn shell_output_is_structured_with_freeform_apply_patch() -> Result<()> {
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
config.features.enable(Feature::ApplyPatchFreeform);
});
let test = builder.build(&server).await?;
@@ -198,6 +225,83 @@ freeform shell
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn shell_output_for_freeform_tool_records_duration() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
#[cfg(target_os = "linux")]
let sleep_cmd = vec!["/bin/bash", "-c", "sleep 1"];
#[cfg(target_os = "macos")]
let sleep_cmd = vec!["/bin/bash", "-c", "sleep 1"];
#[cfg(windows)]
let sleep_cmd = "timeout 1";
let call_id = "shell-structured";
let args = json!({
"command": sleep_cmd,
"timeout_ms": 2_000,
});
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_function_call(call_id, "shell", &serde_json::to_string(&args)?),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"run the structured shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("structured output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("structured output string");
let expected_pattern = r#"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
$"#;
assert_regex_match(expected_pattern, output);
let wall_time_regex = Regex::new(r"(?m)^Wall (?:time|Clock): ([0-9]+(?:\.[0-9]+)?) seconds$")
.expect("compile wall time regex");
let wall_time_seconds = wall_time_regex
.captures(output)
.and_then(|caps| caps.get(1))
.and_then(|value| value.as_str().parse::<f32>().ok())
.expect("expected structured shell output to contain wall time seconds");
assert!(
wall_time_seconds > 0.5,
"expected wall time to be greater than zero seconds, got {wall_time_seconds}"
);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn shell_output_reserializes_truncated_content() -> Result<()> {
skip_if_no_network!(Ok(()));
@@ -213,7 +317,7 @@ async fn shell_output_reserializes_truncated_content() -> Result<()> {
let call_id = "shell-truncated";
let args = json!({
"command": ["/bin/sh", "-c", "seq 1 400"],
"timeout_ms": 1_000,
"timeout_ms": 5_000,
});
let responses = vec![
sse(vec![
@@ -275,3 +379,428 @@ $"#;
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_custom_tool_output_is_structured() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "apply-patch-structured";
let file_name = "structured.txt";
let patch = format!(
r#"*** Begin Patch
*** Add File: {file_name}
+from custom tool
*** End Patch
"#
);
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_custom_tool_call(call_id, "apply_patch", &patch),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"apply the patch via custom tool",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("apply_patch output string");
let expected_pattern = format!(
r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
Success. Updated the following files:
A {file_name}
?$"
);
assert_regex_match(&expected_pattern, output);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_custom_tool_call_creates_file() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "apply-patch-add-file";
let file_name = "custom_tool_apply_patch.txt";
let patch = format!(
"*** Begin Patch\n*** Add File: {file_name}\n+custom tool content\n*** End Patch\n"
);
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_custom_tool_call(call_id, "apply_patch", &patch),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "apply_patch done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"apply the patch via custom tool to create a file",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("apply_patch output string");
let expected_pattern = format!(
r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
Success. Updated the following files:
A {file_name}
?$"
);
assert_regex_match(&expected_pattern, output);
let new_file_path = test.cwd.path().join(file_name);
let created_contents = fs::read_to_string(&new_file_path)?;
assert_eq!(
created_contents, "custom tool content\n",
"expected file contents for {file_name}"
);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_custom_tool_call_updates_existing_file() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "apply-patch-update-file";
let file_name = "custom_tool_apply_patch_existing.txt";
let file_path = test.cwd.path().join(file_name);
fs::write(&file_path, "before\n")?;
let patch = format!(
"*** Begin Patch\n*** Update File: {file_name}\n@@\n-before\n+after\n*** End Patch\n"
);
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_custom_tool_call(call_id, "apply_patch", &patch),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "apply_patch update done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"apply the patch via custom tool to update a file",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("apply_patch output string");
let expected_pattern = format!(
r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
Success. Updated the following files:
M {file_name}
?$"
);
assert_regex_match(&expected_pattern, output);
let updated_contents = fs::read_to_string(file_path)?;
assert_eq!(updated_contents, "after\n", "expected updated file content");
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_custom_tool_call_reports_failure_output() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "apply-patch-failure";
let missing_file = "missing_custom_tool_apply_patch.txt";
let patch = format!(
"*** Begin Patch\n*** Update File: {missing_file}\n@@\n-before\n+after\n*** End Patch\n"
);
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_custom_tool_call(call_id, "apply_patch", &patch),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "apply_patch failure done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"attempt a failing apply_patch via custom tool",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_custom_tool_call_output(&bodies, call_id).expect("apply_patch output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("apply_patch output string");
let expected_output = format!(
"apply_patch verification failed: Failed to read file to update {}/{missing_file}: No such file or directory (os error 2)",
test.cwd.path().to_string_lossy()
);
assert_eq!(output, expected_output);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn apply_patch_function_call_output_is_structured() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "apply-patch-function";
let file_name = "function_apply_patch.txt";
let patch =
format!("*** Begin Patch\n*** Add File: {file_name}\n+via function call\n*** End Patch\n");
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_apply_patch_function_call(call_id, &patch),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "apply_patch function done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"apply the patch via function-call apply_patch",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("apply_patch function output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("apply_patch output string");
let expected_pattern = format!(
r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
Success. Updated the following files:
A {file_name}
?$"
);
assert_regex_match(&expected_pattern, output);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn shell_output_is_structured_for_nonzero_exit() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.model = "gpt-5-codex".to_string();
config.model_family =
find_family_for_model("gpt-5-codex").expect("gpt-5-codex is a model family");
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "shell-nonzero-exit";
let args = json!({
"command": ["/bin/sh", "-c", "exit 42"],
"timeout_ms": 1_000,
});
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_function_call(call_id, "shell", &serde_json::to_string(&args)?),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "shell failure handled"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"run the failing shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item = find_function_call_output(&bodies, call_id).expect("shell output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("shell output string");
let expected_pattern = r"(?s)^Exit code: 42
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
?$";
assert_regex_match(expected_pattern, output);
Ok(())
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn local_shell_call_output_is_structured() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.model = "gpt-5-codex".to_string();
config.model_family =
find_family_for_model("gpt-5-codex").expect("gpt-5-codex is a model family");
config.include_apply_patch_tool = true;
});
let test = builder.build(&server).await?;
let call_id = "local-shell-call";
let responses = vec![
sse(vec![
json!({"type": "response.created", "response": {"id": "resp-1"}}),
ev_local_shell_call(call_id, "completed", vec!["/bin/echo", "local shell"]),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "local shell done"),
ev_completed("resp-2"),
]),
];
mount_sse_sequence(&server, responses).await;
submit_turn(
&test,
"run the local shell command",
SandboxPolicy::DangerFullAccess,
)
.await?;
let requests = server
.received_requests()
.await
.expect("recorded requests present");
let bodies = request_bodies(&requests)?;
let output_item =
find_function_call_output(&bodies, call_id).expect("local shell output present");
let output = output_item
.get("output")
.and_then(Value::as_str)
.expect("local shell output string");
let expected_pattern = r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
local shell
?$";
assert_regex_match(expected_pattern, output);
Ok(())
}

View File

@@ -1,6 +1,9 @@
#![cfg(not(target_os = "windows"))]
use std::fs;
use assert_matches::assert_matches;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
@@ -104,7 +107,7 @@ async fn update_plan_tool_emits_plan_update_event() -> anyhow::Result<()> {
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_plan_tool = true;
config.features.enable(Feature::PlanTool);
});
let TestCodex {
codex,
@@ -191,7 +194,7 @@ async fn update_plan_tool_rejects_malformed_payload() -> anyhow::Result<()> {
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_plan_tool = true;
config.features.enable(Feature::PlanTool);
});
let TestCodex {
codex,
@@ -285,7 +288,7 @@ async fn apply_patch_tool_executes_and_emits_patch_events() -> anyhow::Result<()
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
config.features.enable(Feature::ApplyPatchFreeform);
});
let TestCodex {
codex,
@@ -294,15 +297,19 @@ async fn apply_patch_tool_executes_and_emits_patch_events() -> anyhow::Result<()
..
} = builder.build(&server).await?;
let file_name = "notes.txt";
let file_path = cwd.path().join(file_name);
let call_id = "apply-patch-call";
let patch_content = r#"*** Begin Patch
*** Add File: notes.txt
let patch_content = format!(
r#"*** Begin Patch
*** Add File: {file_name}
+Tool harness apply patch
*** End Patch"#;
*** End Patch"#
);
let first_response = sse(vec![
ev_response_created("resp-1"),
ev_apply_patch_function_call(call_id, patch_content),
ev_apply_patch_function_call(call_id, &patch_content),
ev_completed("resp-1"),
]);
responses::mount_sse_once_match(&server, any(), first_response).await;
@@ -351,6 +358,7 @@ async fn apply_patch_tool_executes_and_emits_patch_events() -> anyhow::Result<()
assert!(saw_patch_begin, "expected PatchApplyBegin event");
let patch_end_success =
patch_end_success.expect("expected PatchApplyEnd event to capture success flag");
assert!(patch_end_success);
let req = second_mock.single_request();
let output_item = req.function_call_output(call_id);
@@ -360,38 +368,21 @@ async fn apply_patch_tool_executes_and_emits_patch_events() -> anyhow::Result<()
);
let output_text = extract_output_text(&output_item).expect("output text present");
if let Ok(exec_output) = serde_json::from_str::<Value>(output_text) {
let exit_code = exec_output["metadata"]["exit_code"]
.as_i64()
.expect("exit_code present");
let summary = exec_output["output"].as_str().expect("output field");
assert_eq!(
exit_code, 0,
"expected apply_patch exit_code=0, got {exit_code}, summary: {summary:?}"
);
assert!(
patch_end_success,
"expected PatchApplyEnd success flag, summary: {summary:?}"
);
assert!(
summary.contains("Success."),
"expected apply_patch summary to note success, got {summary:?}"
);
let expected_pattern = format!(
r"(?s)^Exit code: 0
Wall time: [0-9]+(?:\.[0-9]+)? seconds
Output:
Success. Updated the following files:
A {file_name}
?$"
);
assert_regex_match(&expected_pattern, output_text);
let patched_path = cwd.path().join("notes.txt");
let contents = std::fs::read_to_string(&patched_path)
.unwrap_or_else(|e| panic!("failed reading {}: {e}", patched_path.display()));
assert_eq!(contents, "Tool harness apply patch\n");
} else {
assert!(
output_text.contains("codex-run-as-apply-patch"),
"expected apply_patch failure message to mention codex-run-as-apply-patch, got {output_text:?}"
);
assert!(
!patch_end_success,
"expected PatchApplyEnd to report success=false when apply_patch invocation fails"
);
}
let updated_contents = fs::read_to_string(file_path)?;
assert_eq!(
updated_contents, "Tool harness apply patch\n",
"expected updated file content"
);
Ok(())
}
@@ -403,7 +394,7 @@ async fn apply_patch_reports_parse_diagnostics() -> anyhow::Result<()> {
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.include_apply_patch_tool = true;
config.features.enable(Feature::ApplyPatchFreeform);
});
let TestCodex {
codex,

View File

@@ -2,6 +2,7 @@
#![allow(clippy::unwrap_used, clippy::expect_used)]
use anyhow::Result;
use codex_core::features::Feature;
use codex_core::model_family::find_family_for_model;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
@@ -293,7 +294,11 @@ async fn collect_tools(use_unified_exec: bool) -> Result<Vec<String>> {
let mock = mount_sse_sequence(&server, responses).await;
let mut builder = test_codex().with_config(move |config| {
config.use_experimental_unified_exec_tool = use_unified_exec;
if use_unified_exec {
config.features.enable(Feature::UnifiedExec);
} else {
config.features.disable(Feature::UnifiedExec);
}
});
let test = builder.build(&server).await?;

View File

@@ -3,6 +3,7 @@
use std::collections::HashMap;
use anyhow::Result;
use codex_core::features::Feature;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::EventMsg;
use codex_core::protocol::InputItem;
@@ -42,7 +43,13 @@ fn collect_tool_outputs(bodies: &[Value]) -> Result<HashMap<String, Value>> {
if let Some(call_id) = item.get("call_id").and_then(Value::as_str) {
let content = extract_output_text(item)
.ok_or_else(|| anyhow::anyhow!("missing tool output content"))?;
let parsed: Value = serde_json::from_str(content)?;
let trimmed = content.trim();
if trimmed.is_empty() {
continue;
}
let parsed: Value = serde_json::from_str(trimmed).map_err(|err| {
anyhow::anyhow!("failed to parse tool output content {trimmed:?}: {err}")
})?;
outputs.insert(call_id.to_string(), parsed);
}
}
@@ -59,7 +66,7 @@ async fn unified_exec_reuses_session_via_stdin() -> Result<()> {
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.use_experimental_unified_exec_tool = true;
config.features.enable(Feature::UnifiedExec);
});
let TestCodex {
codex,
@@ -176,6 +183,7 @@ async fn unified_exec_streams_after_lagged_output() -> Result<()> {
let mut builder = test_codex().with_config(|config| {
config.use_experimental_unified_exec_tool = true;
config.features.enable(Feature::UnifiedExec);
});
let TestCodex {
codex,
@@ -213,7 +221,7 @@ PY
let second_args = serde_json::json!({
"input": Vec::<String>::new(),
"session_id": "0",
"timeout_ms": 800,
"timeout_ms": 2_000,
});
let responses = vec![
@@ -300,7 +308,7 @@ async fn unified_exec_timeout_and_followup_poll() -> Result<()> {
let server = start_mock_server().await;
let mut builder = test_codex().with_config(|config| {
config.use_experimental_unified_exec_tool = true;
config.features.enable(Feature::UnifiedExec);
});
let TestCodex {
codex,

View File

@@ -5,6 +5,7 @@ use std::os::unix::fs::PermissionsExt;
use codex_core::protocol::EventMsg;
use codex_core::protocol::InputItem;
use codex_core::protocol::Op;
use core_test_support::fs_wait;
use core_test_support::responses;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex::TestCodex;
@@ -17,8 +18,7 @@ use responses::ev_assistant_message;
use responses::ev_completed;
use responses::sse;
use responses::start_mock_server;
use tokio::time::Duration;
use tokio::time::sleep;
use std::time::Duration;
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn summarize_context_three_requests_and_instructions() -> anyhow::Result<()> {
@@ -60,14 +60,7 @@ echo -n "${@: -1}" > $(dirname "${0}")/notify.txt"#,
wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
// We fork the notify script, so we need to wait for it to write to the file.
for _ in 0..100u32 {
if notify_file.exists() {
break;
}
sleep(Duration::from_millis(100)).await;
}
assert!(notify_file.exists());
fs_wait::wait_for_path_exists(&notify_file, Duration::from_secs(5)).await?;
Ok(())
}

View File

@@ -17,6 +17,7 @@ use codex_core::ConversationManager;
use codex_core::NewConversation;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
use codex_core::features::Feature;
use codex_core::git_info::get_git_repo_root;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::Event;
@@ -168,8 +169,7 @@ pub async fn run_main(cli: Cli, codex_linux_sandbox_exe: Option<PathBuf>) -> any
model,
review_model: None,
config_profile,
// This CLI is intended to be headless and has no affordances for asking
// the user for approval.
// Default to never ask for approvals in headless mode. Feature flags can override.
approval_policy: Some(AskForApproval::Never),
sandbox_mode,
cwd: cwd.map(|p| p.canonicalize().unwrap_or(p)),
@@ -192,6 +192,7 @@ pub async fn run_main(cli: Cli, codex_linux_sandbox_exe: Option<PathBuf>) -> any
};
let config = Config::load_with_cli_overrides(cli_kv_overrides, overrides).await?;
let approve_all_enabled = config.features.enabled(Feature::ApproveAll);
let otel = codex_core::otel_init::build_provider(&config, env!("CARGO_PKG_VERSION"));
@@ -360,6 +361,34 @@ pub async fn run_main(cli: Cli, codex_linux_sandbox_exe: Option<PathBuf>) -> any
if matches!(event.msg, EventMsg::Error(_)) {
error_seen = true;
}
// Auto-approve requests when the approve_all feature is enabled.
if approve_all_enabled {
match &event.msg {
EventMsg::ExecApprovalRequest(_) => {
if let Err(e) = conversation
.submit(Op::ExecApproval {
id: event.id.clone(),
decision: codex_core::protocol::ReviewDecision::Approved,
})
.await
{
error!("failed to auto-approve exec: {e}");
}
}
EventMsg::ApplyPatchApprovalRequest(_) => {
if let Err(e) = conversation
.submit(Op::PatchApproval {
id: event.id.clone(),
decision: codex_core::protocol::ReviewDecision::Approved,
})
.await
{
error!("failed to auto-approve patch: {e}");
}
}
_ => {}
}
}
let shutdown: CodexStatus = event_processor.process_event(event);
match shutdown {
CodexStatus::Running => continue,

View File

@@ -0,0 +1,81 @@
#![cfg(not(target_os = "windows"))]
#![allow(clippy::expect_used, clippy::unwrap_used)]
use anyhow::Result;
use core_test_support::responses;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
use core_test_support::responses::ev_function_call;
use core_test_support::responses::ev_response_created;
use core_test_support::responses::mount_sse_sequence;
use core_test_support::responses::sse;
use core_test_support::skip_if_no_network;
use core_test_support::test_codex_exec::test_codex_exec;
use serde_json::Value;
use serde_json::json;
async fn run_exec_with_args(args: &[&str]) -> Result<String> {
let test = test_codex_exec();
let call_id = "exec-approve";
let exec_args = json!({
"command": [
if cfg!(windows) { "cmd.exe" } else { "/bin/sh" },
if cfg!(windows) { "/C" } else { "-lc" },
"echo approve-all-ok",
],
"timeout_ms": 1500,
"with_escalated_permissions": true
});
let response_streams = vec![
sse(vec![
ev_response_created("resp-1"),
ev_function_call(call_id, "shell", &serde_json::to_string(&exec_args)?),
ev_completed("resp-1"),
]),
sse(vec![
ev_assistant_message("msg-1", "done"),
ev_completed("resp-2"),
]),
];
let server = responses::start_mock_server().await;
let mock = mount_sse_sequence(&server, response_streams).await;
test.cmd_with_server(&server).args(args).assert().success();
let requests = mock.requests();
assert!(requests.len() >= 2, "expected at least two responses POSTs");
let item = requests[1].function_call_output(call_id);
let output_str = item
.get("output")
.and_then(Value::as_str)
.expect("function_call_output.output should be a string");
Ok(output_str.to_string())
}
/// Setting `features.approve_all=true` should switch to auto-approvals.
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn approve_all_auto_accepts_exec() -> Result<()> {
skip_if_no_network!(Ok(()));
let output = run_exec_with_args(&[
"--skip-git-repo-check",
"-c",
"features.approve_all=true",
"train",
])
.await?;
assert!(
output.contains("Exit code: 0"),
"expected Exit code: 0 in output: {output}"
);
assert!(
output.contains("approve-all-ok"),
"expected command output in response: {output}"
);
Ok(())
}

View File

@@ -1,5 +1,6 @@
// Aggregates all former standalone integration tests as modules.
mod apply_patch;
mod approve_all;
mod auth_env;
mod originator;
mod output_schema;

View File

@@ -178,6 +178,7 @@ async fn run_codex_tool_session_inner(
cwd,
call_id,
reason: _,
parsed_cmd,
}) => {
handle_exec_approval_request(
command,
@@ -188,6 +189,7 @@ async fn run_codex_tool_session_inner(
request_id_str.clone(),
event.id.clone(),
call_id,
parsed_cmd,
)
.await;
continue;

View File

@@ -4,6 +4,7 @@ use std::sync::Arc;
use codex_core::CodexConversation;
use codex_core::protocol::Op;
use codex_core::protocol::ReviewDecision;
use codex_protocol::parse_command::ParsedCommand;
use mcp_types::ElicitRequest;
use mcp_types::ElicitRequestParamsRequestedSchema;
use mcp_types::JSONRPCErrorError;
@@ -35,6 +36,7 @@ pub struct ExecApprovalElicitRequestParams {
pub codex_call_id: String,
pub codex_command: Vec<String>,
pub codex_cwd: PathBuf,
pub codex_parsed_cmd: Vec<ParsedCommand>,
}
// TODO(mbolin): ExecApprovalResponse does not conform to ElicitResult. See:
@@ -56,6 +58,7 @@ pub(crate) async fn handle_exec_approval_request(
tool_call_id: String,
event_id: String,
call_id: String,
codex_parsed_cmd: Vec<ParsedCommand>,
) {
let escaped_command =
shlex::try_join(command.iter().map(String::as_str)).unwrap_or_else(|_| command.join(" "));
@@ -77,6 +80,7 @@ pub(crate) async fn handle_exec_approval_request(
codex_call_id: call_id,
codex_command: command,
codex_cwd: cwd,
codex_parsed_cmd,
};
let params_json = match serde_json::to_value(&params) {
Ok(value) => value,

View File

@@ -3,6 +3,7 @@ use std::env;
use std::path::Path;
use std::path::PathBuf;
use codex_core::parse_command;
use codex_core::protocol::FileChange;
use codex_core::protocol::ReviewDecision;
use codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR;
@@ -176,6 +177,7 @@ fn create_expected_elicitation_request(
shlex::try_join(command.iter().map(std::convert::AsRef::as_ref))?,
workdir.to_string_lossy()
);
let codex_parsed_cmd = parse_command::parse_command(&command);
Ok(JSONRPCRequest {
jsonrpc: JSONRPC_VERSION.into(),
id: elicitation_request_id,
@@ -193,6 +195,7 @@ fn create_expected_elicitation_request(
codex_command: command,
codex_cwd: workdir.to_path_buf(),
codex_call_id: "call1234".to_string(),
codex_parsed_cmd,
})?),
})
}

View File

@@ -1178,6 +1178,7 @@ pub struct ExecApprovalRequestEvent {
/// Optional human-readable reason for the approval (e.g. retry without sandbox).
#[serde(skip_serializing_if = "Option::is_none")]
pub reason: Option<String>,
pub parsed_cmd: Vec<ParsedCommand>,
}
#[derive(Debug, Clone, Deserialize, Serialize, TS)]
@@ -1203,6 +1204,11 @@ pub struct StreamErrorEvent {
pub message: String,
}
#[derive(Debug, Clone, Deserialize, Serialize, TS)]
pub struct StreamInfoEvent {
pub message: String,
}
#[derive(Debug, Clone, Deserialize, Serialize, TS)]
pub struct PatchApplyBeginEvent {
/// Identifier so this can be paired with the PatchApplyEnd event.

View File

@@ -44,7 +44,7 @@ pub async fn determine_streamable_http_auth_status(
}
/// Attempt to determine whether a streamable HTTP MCP server advertises OAuth login.
async fn supports_oauth_login(url: &str) -> Result<bool> {
pub async fn supports_oauth_login(url: &str) -> Result<bool> {
let base_url = Url::parse(url)?;
let client = Client::builder().timeout(DISCOVERY_TIMEOUT).build()?;

View File

@@ -7,6 +7,7 @@ mod rmcp_client;
mod utils;
pub use auth_status::determine_streamable_http_auth_status;
pub use auth_status::supports_oauth_login;
pub use codex_protocol::protocol::McpAuthStatus;
pub use oauth::OAuthCredentialsStoreMode;
pub use oauth::StoredOAuthTokens;

View File

@@ -427,8 +427,9 @@ impl App {
tui.frame_requester().schedule_frame();
}
// Esc primes/advances backtracking only in normal (not working) mode
// with an empty composer. In any other state, forward Esc so the
// active UI (e.g. status indicator, modals, popups) handles it.
// with the composer focused and empty. In any other state, forward
// Esc so the active UI (e.g. status indicator, modals, popups)
// handles it.
KeyEvent {
code: KeyCode::Esc,
kind: KeyEventKind::Press | KeyEventKind::Repeat,

View File

@@ -82,15 +82,16 @@ impl App {
/// Handle global Esc presses for backtracking when no overlay is present.
pub(crate) fn handle_backtrack_esc_key(&mut self, tui: &mut tui::Tui) {
// Only handle backtracking when composer is empty to avoid clobbering edits.
if self.chat_widget.composer_is_empty() {
if !self.backtrack.primed {
self.prime_backtrack();
} else if self.overlay.is_none() {
self.open_backtrack_preview(tui);
} else if self.backtrack.overlay_preview_active {
self.step_backtrack_and_highlight(tui);
}
if !self.chat_widget.composer_is_empty() {
return;
}
if !self.backtrack.primed {
self.prime_backtrack();
} else if self.overlay.is_none() {
self.open_backtrack_preview(tui);
} else if self.backtrack.overlay_preview_active {
self.step_backtrack_and_highlight(tui);
}
}

View File

@@ -165,8 +165,9 @@ impl ChatComposer {
.unwrap_or_else(|| footer_height(footer_props));
let footer_spacing = Self::footer_spacing(footer_hint_height);
let footer_total_height = footer_hint_height + footer_spacing;
const COLS_WITH_MARGIN: u16 = LIVE_PREFIX_COLS + 1;
self.textarea
.desired_height(width.saturating_sub(LIVE_PREFIX_COLS))
.desired_height(width.saturating_sub(COLS_WITH_MARGIN))
+ 2
+ match &self.active_popup {
ActivePopup::None => footer_total_height,
@@ -197,7 +198,9 @@ impl ChatComposer {
let [composer_rect, popup_rect] =
Layout::vertical([Constraint::Min(1), popup_constraint]).areas(area);
let mut textarea_rect = composer_rect;
textarea_rect.width = textarea_rect.width.saturating_sub(LIVE_PREFIX_COLS);
textarea_rect.width = textarea_rect.width.saturating_sub(
LIVE_PREFIX_COLS + 1, /* keep a one-column right margin for wrapping */
);
textarea_rect.x = textarea_rect.x.saturating_add(LIVE_PREFIX_COLS);
[composer_rect, textarea_rect, popup_rect]
}
@@ -313,14 +316,23 @@ impl ChatComposer {
self.sync_file_search_popup();
}
pub(crate) fn clear_for_ctrl_c(&mut self) {
self.set_text_content(String::new());
self.history.reset_navigation();
}
/// Get the current composer text.
pub(crate) fn current_text(&self) -> String {
self.textarea.text().to_string()
}
/// Attempt to start a burst by retro-capturing recent chars before the cursor.
pub fn attach_image(&mut self, path: PathBuf, width: u32, height: u32, format_label: &str) {
let placeholder = format!("[image {width}x{height} {format_label}]");
pub fn attach_image(&mut self, path: PathBuf, width: u32, height: u32, _format_label: &str) {
let file_label = path
.file_name()
.map(|name| name.to_string_lossy().into_owned())
.unwrap_or_else(|| "image".to_string());
let placeholder = format!("[{file_label} {width}x{height}]");
// Insert as an element to match large paste placeholder behavior:
// styled distinctly and treated atomically for cursor/mutations.
self.textarea.insert_element(&placeholder);
@@ -845,10 +857,12 @@ impl ChatComposer {
return (InputResult::None, true);
}
if key_event.code == KeyCode::Esc {
let next_mode = esc_hint_mode(self.footer_mode, self.is_task_running);
if next_mode != self.footer_mode {
self.footer_mode = next_mode;
return (InputResult::None, true);
if self.is_empty() {
let next_mode = esc_hint_mode(self.footer_mode, self.is_task_running);
if next_mode != self.footer_mode {
self.footer_mode = next_mode;
return (InputResult::None, true);
}
}
} else {
self.footer_mode = reset_mode_after_activity(self.footer_mode);
@@ -958,6 +972,7 @@ impl ChatComposer {
}
let mut text = self.textarea.text().to_string();
let original_input = text.clone();
let input_starts_with_space = original_input.starts_with(' ');
self.textarea.set_text("");
// Replace all pending pastes in the text
@@ -971,6 +986,35 @@ impl ChatComposer {
// If there is neither text nor attachments, suppress submission entirely.
let has_attachments = !self.attached_images.is_empty();
text = text.trim().to_string();
if let Some((name, _rest)) = parse_slash_name(&text) {
let treat_as_plain_text = input_starts_with_space || name.contains('/');
if !treat_as_plain_text {
let is_builtin = built_in_slash_commands()
.into_iter()
.any(|(command_name, _)| command_name == name);
let prompt_prefix = format!("{PROMPTS_CMD_PREFIX}:");
let is_known_prompt = name
.strip_prefix(&prompt_prefix)
.map(|prompt_name| {
self.custom_prompts
.iter()
.any(|prompt| prompt.name == prompt_name)
})
.unwrap_or(false);
if !is_builtin && !is_known_prompt {
let message = format!(
r#"Unrecognized command '/{name}'. Type "/" for a list of supported commands."#
);
self.app_event_tx.send(AppEvent::InsertHistoryCell(Box::new(
history_cell::new_info_event(message, None),
)));
self.textarea.set_text(&original_input);
self.textarea.set_cursor(original_input.len());
return (InputResult::None, true);
}
}
}
let expanded_prompt = match expand_custom_prompt(&text, &self.custom_prompts) {
Ok(expanded) => expanded,
Err(err) => {
@@ -1755,6 +1799,35 @@ mod tests {
});
}
#[test]
fn esc_hint_stays_hidden_with_draft_content() {
use crossterm::event::KeyCode;
use crossterm::event::KeyEvent;
use crossterm::event::KeyModifiers;
let (tx, _rx) = unbounded_channel::<AppEvent>();
let sender = AppEventSender::new(tx);
let mut composer = ChatComposer::new(
true,
sender,
true,
"Ask Codex to do anything".to_string(),
false,
);
type_chars_humanlike(&mut composer, &['d']);
assert!(!composer.is_empty());
assert_eq!(composer.current_text(), "d");
assert_eq!(composer.footer_mode, FooterMode::ShortcutSummary);
assert!(matches!(composer.active_popup, ActivePopup::None));
let _ = composer.handle_key_event(KeyEvent::new(KeyCode::Esc, KeyModifiers::NONE));
assert_eq!(composer.footer_mode, FooterMode::ShortcutSummary);
assert!(!composer.esc_backtrack_hint);
}
#[test]
fn question_mark_only_toggles_on_first_char() {
use crossterm::event::KeyCode;
@@ -2581,7 +2654,7 @@ mod tests {
let (result, _) =
composer.handle_key_event(KeyEvent::new(KeyCode::Enter, KeyModifiers::NONE));
match result {
InputResult::Submitted(text) => assert_eq!(text, "[image 32x16 PNG] hi"),
InputResult::Submitted(text) => assert_eq!(text, "[image1.png 32x16] hi"),
_ => panic!("expected Submitted"),
}
let imgs = composer.take_recent_submission_images();
@@ -2604,7 +2677,7 @@ mod tests {
let (result, _) =
composer.handle_key_event(KeyEvent::new(KeyCode::Enter, KeyModifiers::NONE));
match result {
InputResult::Submitted(text) => assert_eq!(text, "[image 10x5 PNG]"),
InputResult::Submitted(text) => assert_eq!(text, "[image2.png 10x5]"),
_ => panic!("expected Submitted"),
}
let imgs = composer.take_recent_submission_images();
@@ -2677,7 +2750,12 @@ mod tests {
composer.handle_key_event(KeyEvent::new(KeyCode::Backspace, KeyModifiers::NONE));
assert_eq!(composer.attached_images.len(), 1);
assert!(composer.textarea.text().starts_with("[image 10x5 PNG]"));
assert!(
composer
.textarea
.text()
.starts_with("[image_multibyte.png 10x5]")
);
}
#[test]
@@ -2700,21 +2778,31 @@ mod tests {
composer.handle_paste(" ".into());
composer.attach_image(path2.clone(), 10, 5, "PNG");
let ph = composer.attached_images[0].placeholder.clone();
let placeholder1 = composer.attached_images[0].placeholder.clone();
let placeholder2 = composer.attached_images[1].placeholder.clone();
let text = composer.textarea.text().to_string();
let start1 = text.find(&ph).expect("first placeholder present");
let end1 = start1 + ph.len();
let start1 = text.find(&placeholder1).expect("first placeholder present");
let end1 = start1 + placeholder1.len();
composer.textarea.set_cursor(end1);
// Backspace should delete the first placeholder and its mapping.
composer.handle_key_event(KeyEvent::new(KeyCode::Backspace, KeyModifiers::NONE));
let new_text = composer.textarea.text().to_string();
assert_eq!(1, new_text.matches(&ph).count(), "one placeholder remains");
assert_eq!(
0,
new_text.matches(&placeholder1).count(),
"first placeholder removed"
);
assert_eq!(
1,
new_text.matches(&placeholder2).count(),
"second placeholder remains"
);
assert_eq!(
vec![AttachedImage {
path: path2,
placeholder: "[image 10x5 PNG]".to_string()
placeholder: "[image_dup2.png 10x5]".to_string()
}],
composer.attached_images,
"one image mapping remains"
@@ -2741,7 +2829,12 @@ mod tests {
let needs_redraw = composer.handle_paste(tmp_path.to_string_lossy().to_string());
assert!(needs_redraw);
assert!(composer.textarea.text().starts_with("[image 3x2 PNG] "));
assert!(
composer
.textarea
.text()
.starts_with("[codex_tui_test_paste_image.png 3x2] ")
);
let imgs = composer.take_recent_submission_images();
assert_eq!(imgs, vec![tmp_path]);
@@ -2853,6 +2946,76 @@ mod tests {
assert!(composer.textarea.is_empty());
}
#[test]
fn slash_path_input_submits_without_command_error() {
use crossterm::event::KeyCode;
use crossterm::event::KeyEvent;
use crossterm::event::KeyModifiers;
let (tx, mut rx) = unbounded_channel::<AppEvent>();
let sender = AppEventSender::new(tx);
let mut composer = ChatComposer::new(
true,
sender,
false,
"Ask Codex to do anything".to_string(),
false,
);
composer
.textarea
.set_text("/Users/example/project/src/main.rs");
let (result, _needs_redraw) =
composer.handle_key_event(KeyEvent::new(KeyCode::Enter, KeyModifiers::NONE));
if let InputResult::Submitted(text) = result {
assert_eq!(text, "/Users/example/project/src/main.rs");
} else {
panic!("expected Submitted");
}
assert!(composer.textarea.is_empty());
match rx.try_recv() {
Ok(event) => panic!("unexpected event: {event:?}"),
Err(tokio::sync::mpsc::error::TryRecvError::Empty) => {}
Err(err) => panic!("unexpected channel state: {err:?}"),
}
}
#[test]
fn slash_with_leading_space_submits_as_text() {
use crossterm::event::KeyCode;
use crossterm::event::KeyEvent;
use crossterm::event::KeyModifiers;
let (tx, mut rx) = unbounded_channel::<AppEvent>();
let sender = AppEventSender::new(tx);
let mut composer = ChatComposer::new(
true,
sender,
false,
"Ask Codex to do anything".to_string(),
false,
);
composer.textarea.set_text(" /this-looks-like-a-command");
let (result, _needs_redraw) =
composer.handle_key_event(KeyEvent::new(KeyCode::Enter, KeyModifiers::NONE));
if let InputResult::Submitted(text) = result {
assert_eq!(text, "/this-looks-like-a-command");
} else {
panic!("expected Submitted");
}
assert!(composer.textarea.is_empty());
match rx.try_recv() {
Ok(event) => panic!("unexpected event: {event:?}"),
Err(tokio::sync::mpsc::error::TryRecvError::Empty) => {}
Err(err) => panic!("unexpected channel state: {err:?}"),
}
}
#[test]
fn custom_prompt_invalid_args_reports_error() {
let (tx, mut rx) = unbounded_channel::<AppEvent>();

View File

@@ -70,6 +70,12 @@ impl ChatComposerHistory {
self.local_history.push(text.to_string());
}
/// Reset navigation tracking so the next Up key resumes from the latest entry.
pub fn reset_navigation(&mut self) {
self.history_cursor = None;
self.last_history_text = None;
}
/// Should Up/Down key presses be interpreted as history navigation given
/// the current content and cursor position of `textarea`?
pub fn should_handle_navigation(&self, text: &str, cursor: usize) -> bool {
@@ -271,4 +277,24 @@ mod tests {
history.on_entry_response(1, 1, Some("older".into()))
);
}
#[test]
fn reset_navigation_resets_cursor() {
let (tx, _rx) = unbounded_channel::<AppEvent>();
let tx = AppEventSender::new(tx);
let mut history = ChatComposerHistory::new();
history.set_metadata(1, 3);
history.fetched_history.insert(1, "command2".into());
history.fetched_history.insert(2, "command3".into());
assert_eq!(Some("command3".into()), history.navigate_up(&tx));
assert_eq!(Some("command2".into()), history.navigate_up(&tx));
history.reset_navigation();
assert!(history.history_cursor.is_none());
assert!(history.last_history_text.is_none());
assert_eq!(Some("command3".into()), history.navigate_up(&tx));
}
}

View File

@@ -37,6 +37,7 @@ pub(crate) struct SelectionItem {
pub name: String,
pub display_shortcut: Option<KeyBinding>,
pub description: Option<String>,
pub selected_description: Option<String>,
pub is_current: bool,
pub actions: Vec<SelectionAction>,
pub dismiss_on_select: bool,
@@ -193,12 +194,16 @@ impl ListSelectionView {
} else {
format!("{prefix} {n}. {name_with_marker}")
};
let description = is_selected
.then(|| item.selected_description.clone())
.flatten()
.or_else(|| item.description.clone());
GenericDisplayRow {
name: display_name,
display_shortcut: item.display_shortcut,
match_indices: None,
is_current: item.is_current,
description: item.description.clone(),
description,
}
})
})

View File

@@ -236,7 +236,7 @@ impl BottomPane {
CancellationEvent::NotHandled
} else {
self.view_stack.pop();
self.set_composer_text(String::new());
self.clear_composer_for_ctrl_c();
self.show_ctrl_c_quit_hint();
CancellationEvent::Handled
}
@@ -270,6 +270,11 @@ impl BottomPane {
self.request_redraw();
}
pub(crate) fn clear_composer_for_ctrl_c(&mut self) {
self.composer.clear_for_ctrl_c();
self.request_redraw();
}
/// Get the current composer text (for tests and programmatic checks).
pub(crate) fn composer_text(&self) -> String {
self.composer.current_text()

View File

@@ -2,4 +2,4 @@
source: tui/src/bottom_pane/mod.rs
expression: "render_snapshot(&pane, area1)"
---
Ask Codex to do an
Ask Codex to do a

View File

@@ -3,4 +3,4 @@ source: tui/src/bottom_pane/mod.rs
expression: "render_snapshot(&pane, area2)"
---
Ask Codex to do an
Ask Codex to do a

View File

@@ -26,6 +26,7 @@ pub(crate) struct TextArea {
wrap_cache: RefCell<Option<WrapCache>>,
preferred_col: Option<usize>,
elements: Vec<TextElement>,
kill_buffer: String,
}
#[derive(Debug, Clone)]
@@ -48,6 +49,7 @@ impl TextArea {
wrap_cache: RefCell::new(None),
preferred_col: None,
elements: Vec::new(),
kill_buffer: String::new(),
}
}
@@ -57,6 +59,7 @@ impl TextArea {
self.wrap_cache.replace(None);
self.preferred_col = None;
self.elements.clear();
self.kill_buffer.clear();
}
pub fn text(&self) -> &str {
@@ -305,6 +308,13 @@ impl TextArea {
} => {
self.kill_to_end_of_line();
}
KeyEvent {
code: KeyCode::Char('y'),
modifiers: KeyModifiers::CONTROL,
..
} => {
self.yank();
}
// Cursor movement
KeyEvent {
@@ -437,7 +447,7 @@ impl TextArea {
pub fn delete_backward_word(&mut self) {
let start = self.beginning_of_previous_word();
self.replace_range(start..self.cursor_pos, "");
self.kill_range(start..self.cursor_pos);
}
/// Delete text to the right of the cursor using "word" semantics.
@@ -448,32 +458,63 @@ impl TextArea {
pub fn delete_forward_word(&mut self) {
let end = self.end_of_next_word();
if end > self.cursor_pos {
self.replace_range(self.cursor_pos..end, "");
self.kill_range(self.cursor_pos..end);
}
}
pub fn kill_to_end_of_line(&mut self) {
let eol = self.end_of_current_line();
if self.cursor_pos == eol {
let range = if self.cursor_pos == eol {
if eol < self.text.len() {
self.replace_range(self.cursor_pos..eol + 1, "");
Some(self.cursor_pos..eol + 1)
} else {
None
}
} else {
self.replace_range(self.cursor_pos..eol, "");
Some(self.cursor_pos..eol)
};
if let Some(range) = range {
self.kill_range(range);
}
}
pub fn kill_to_beginning_of_line(&mut self) {
let bol = self.beginning_of_current_line();
if self.cursor_pos == bol {
if bol > 0 {
self.replace_range(bol - 1..bol, "");
}
let range = if self.cursor_pos == bol {
if bol > 0 { Some(bol - 1..bol) } else { None }
} else {
self.replace_range(bol..self.cursor_pos, "");
Some(bol..self.cursor_pos)
};
if let Some(range) = range {
self.kill_range(range);
}
}
pub fn yank(&mut self) {
if self.kill_buffer.is_empty() {
return;
}
let text = self.kill_buffer.clone();
self.insert_str(&text);
}
fn kill_range(&mut self, range: Range<usize>) {
let range = self.expand_range_to_element_boundaries(range);
if range.start >= range.end {
return;
}
let removed = self.text[range.clone()].to_string();
if removed.is_empty() {
return;
}
self.kill_buffer = removed;
self.replace_range_raw(range, "");
}
/// Move the cursor left by a single grapheme cluster.
pub fn move_cursor_left(&mut self) {
self.cursor_pos = self.prev_atomic_boundary(self.cursor_pos);
@@ -799,16 +840,22 @@ impl TextArea {
}
pub(crate) fn beginning_of_previous_word(&self) -> usize {
if let Some(first_non_ws) = self.text[..self.cursor_pos].rfind(|c: char| !c.is_whitespace())
{
let candidate = self.text[..first_non_ws]
.rfind(|c: char| c.is_whitespace())
.map(|i| i + 1)
.unwrap_or(0);
self.adjust_pos_out_of_elements(candidate, true)
} else {
0
}
let prefix = &self.text[..self.cursor_pos];
let Some((first_non_ws_idx, _)) = prefix
.char_indices()
.rev()
.find(|&(_, ch)| !ch.is_whitespace())
else {
return 0;
};
let before = &prefix[..first_non_ws_idx];
let candidate = before
.char_indices()
.rev()
.find(|&(_, ch)| ch.is_whitespace())
.map(|(idx, ch)| idx + ch.len_utf8())
.unwrap_or(0);
self.adjust_pos_out_of_elements(candidate, true)
}
pub(crate) fn end_of_next_word(&self) -> usize {
@@ -1192,6 +1239,39 @@ mod tests {
assert_eq!(t.cursor(), elem_range.start);
}
#[test]
fn yank_restores_last_kill() {
let mut t = ta_with("hello");
t.set_cursor(0);
t.kill_to_end_of_line();
assert_eq!(t.text(), "");
assert_eq!(t.cursor(), 0);
t.yank();
assert_eq!(t.text(), "hello");
assert_eq!(t.cursor(), 5);
let mut t = ta_with("hello world");
t.set_cursor(t.text().len());
t.delete_backward_word();
assert_eq!(t.text(), "hello ");
assert_eq!(t.cursor(), 6);
t.yank();
assert_eq!(t.text(), "hello world");
assert_eq!(t.cursor(), 11);
let mut t = ta_with("hello");
t.set_cursor(5);
t.kill_to_beginning_of_line();
assert_eq!(t.text(), "");
assert_eq!(t.cursor(), 0);
t.yank();
assert_eq!(t.text(), "hello");
assert_eq!(t.cursor(), 5);
}
#[test]
fn cursor_left_and_right_handle_graphemes() {
let mut t = ta_with("a👍b");
@@ -1262,6 +1342,15 @@ mod tests {
assert_eq!(t.cursor(), 6);
}
#[test]
fn delete_backward_word_handles_narrow_no_break_space() {
let mut t = ta_with("32\u{202F}AM");
t.set_cursor(t.text().len());
t.input(KeyEvent::new(KeyCode::Backspace, KeyModifiers::ALT));
pretty_assertions::assert_eq!(t.text(), "32\u{202F}");
pretty_assertions::assert_eq!(t.cursor(), t.text().len());
}
#[test]
fn delete_forward_word_with_without_alt_modifier() {
let mut t = ta_with("hello world");

View File

@@ -53,6 +53,8 @@ use ratatui::buffer::Buffer;
use ratatui::layout::Constraint;
use ratatui::layout::Layout;
use ratatui::layout::Rect;
use ratatui::style::Stylize;
use ratatui::text::Line;
use ratatui::widgets::Widget;
use ratatui::widgets::WidgetRef;
use tokio::sync::mpsc::UnboundedSender;
@@ -81,6 +83,7 @@ use crate::history_cell::AgentMessageCell;
use crate::history_cell::HistoryCell;
use crate::history_cell::McpToolCallCell;
use crate::markdown::append_markdown;
use crate::render::renderable::ColumnRenderable;
use crate::slash_command::SlashCommand;
use crate::status::RateLimitSnapshotDisplay;
use crate::text_formatting::truncate_text;
@@ -240,6 +243,10 @@ pub(crate) struct ChatWidget {
reasoning_buffer: String,
// Accumulates full reasoning content for transcript-only recording
full_reasoning_buffer: String,
// Current status header shown in the status indicator.
current_status_header: String,
// Previous status header to restore after a transient stream retry.
retry_status_header: Option<String>,
conversation_id: Option<ConversationId>,
frame_requester: FrameRequester,
// Whether to include the initial welcome banner on session configured
@@ -303,6 +310,14 @@ impl ChatWidget {
}
}
fn set_status_header(&mut self, header: String) {
if self.current_status_header == header {
return;
}
self.current_status_header = header.clone();
self.bottom_pane.update_status_header(header);
}
// --- Small event handlers ---
fn on_session_configured(&mut self, event: codex_core::protocol::SessionConfiguredEvent) {
self.bottom_pane
@@ -352,7 +367,7 @@ impl ChatWidget {
if let Some(header) = extract_first_bold(&self.reasoning_buffer) {
// Update the shimmer header to the extracted reasoning chunk header.
self.bottom_pane.update_status_header(header);
self.set_status_header(header);
} else {
// Fallback while we don't yet have a bold header: leave existing header as-is.
}
@@ -386,6 +401,8 @@ impl ChatWidget {
fn on_task_started(&mut self) {
self.bottom_pane.clear_ctrl_c_quit_hint();
self.bottom_pane.set_task_running(true);
self.retry_status_header = None;
self.set_status_header(String::from("Working"));
self.full_reasoning_buffer.clear();
self.reasoning_buffer.clear();
self.request_redraw();
@@ -621,9 +638,10 @@ impl ChatWidget {
}
fn on_stream_error(&mut self, message: String) {
// Show stream errors in the transcript so users see retry/backoff info.
self.add_to_history(history_cell::new_stream_error_event(message));
self.request_redraw();
if self.retry_status_header.is_none() {
self.retry_status_header = Some(self.current_status_header.clone());
}
self.set_status_header(message);
}
/// Periodic tick to commit at most one queued line to history with a small delay,
@@ -928,6 +946,8 @@ impl ChatWidget {
interrupts: InterruptManager::new(),
reasoning_buffer: String::new(),
full_reasoning_buffer: String::new(),
current_status_header: String::from("Working"),
retry_status_header: None,
conversation_id: None,
queued_user_messages: VecDeque::new(),
show_welcome_banner: true,
@@ -991,6 +1011,8 @@ impl ChatWidget {
interrupts: InterruptManager::new(),
reasoning_buffer: String::new(),
full_reasoning_buffer: String::new(),
current_status_header: String::from("Working"),
retry_status_header: None,
conversation_id: None,
queued_user_messages: VecDeque::new(),
show_welcome_banner: true,
@@ -1015,20 +1037,20 @@ impl ChatWidget {
pub(crate) fn handle_key_event(&mut self, key_event: KeyEvent) {
match key_event {
KeyEvent {
code: KeyCode::Char('c'),
modifiers: crossterm::event::KeyModifiers::CONTROL,
code: KeyCode::Char(c),
modifiers,
kind: KeyEventKind::Press,
..
} => {
} if modifiers.contains(KeyModifiers::CONTROL) && c.eq_ignore_ascii_case(&'c') => {
self.on_ctrl_c();
return;
}
KeyEvent {
code: KeyCode::Char('v'),
modifiers: KeyModifiers::CONTROL,
code: KeyCode::Char(c),
modifiers,
kind: KeyEventKind::Press,
..
} => {
} if modifiers.contains(KeyModifiers::CONTROL) && c.eq_ignore_ascii_case(&'v') => {
if let Ok((path, info)) = paste_image_to_temp_png() {
self.attach_image(path, info.width, info.height, info.encoded_format.label());
}
@@ -1699,7 +1721,6 @@ impl ChatWidget {
} else {
default_choice
};
let mut items: Vec<SelectionItem> = Vec::new();
for choice in choices.iter() {
let effort = choice.display;
@@ -1722,6 +1743,14 @@ impl ChatWidget {
.map(|preset| preset.description.to_string())
});
let warning = "⚠ High reasoning effort can quickly consume Plus plan rate limits.";
let show_warning = model_slug == "gpt-5-codex" && effort == ReasoningEffortConfig::High;
let selected_description = show_warning.then(|| {
description
.as_ref()
.map_or(warning.to_string(), |d| format!("{d}\n{warning}"))
});
let model_for_action = model_slug.clone();
let effort_for_action = choice.stored;
let actions: Vec<SelectionAction> = vec![Box::new(move |tx| {
@@ -1751,6 +1780,7 @@ impl ChatWidget {
items.push(SelectionItem {
name: effort_label,
description,
selected_description,
is_current: is_current_model && choice.stored == highlight_choice,
actions,
dismiss_on_select: true,
@@ -1758,9 +1788,13 @@ impl ChatWidget {
});
}
let mut header = ColumnRenderable::new();
header.push(Line::from(
format!("Select Reasoning Level for {model_slug}").bold(),
));
self.bottom_pane.show_selection_view(SelectionViewParams {
title: Some("Select Reasoning Level".to_string()),
subtitle: Some(format!("Reasoning for model {model_slug}")),
header: Box::new(header),
footer_hint: Some(standard_popup_hint_line()),
items,
..Default::default()

View File

@@ -1,6 +1,5 @@
---
source: tui/src/chatwidget/tests.rs
assertion_line: 1152
expression: "lines[start_idx..].join(\"\\n\")"
---
• I need to check the codex-rs repository to explain why the project's binaries
@@ -33,7 +32,7 @@ expression: "lines[start_idx..].join(\"\\n\")"
│ … +1 lines
└ --- ansi-escape/Cargo.toml
[package]
… +7 lines
… +243 lines
] }
tracing = { version

View File

@@ -2,12 +2,13 @@
source: tui/src/chatwidget/tests.rs
expression: popup
---
Select Reasoning Level
Reasoning for model gpt-5-codex
Select Reasoning Level for gpt-5-codex
1. Low Fastest responses with limited reasoning
2. Medium (default) Dynamically adjusts reasoning based on the task
3. High (current) Maximizes reasoning depth for complex or ambiguous
problems
⚠ High reasoning effort can quickly consume Plus plan
rate limits.
Press enter to confirm or esc to go back

View File

@@ -276,6 +276,8 @@ fn make_chatwidget_manual() -> (
interrupts: InterruptManager::new(),
reasoning_buffer: String::new(),
full_reasoning_buffer: String::new(),
current_status_header: String::from("Working"),
retry_status_header: None,
conversation_id: None,
frame_requester: FrameRequester::test_dummy(),
show_welcome_banner: true,
@@ -390,6 +392,7 @@ fn exec_approval_emits_proposed_command_and_decision_history() {
reason: Some(
"this is a test reason such as one that would be produced by the model".into(),
),
parsed_cmd: vec![],
};
chat.handle_codex_event(Event {
id: "sub-short".into(),
@@ -431,6 +434,7 @@ fn exec_approval_decision_truncates_multiline_and_long_commands() {
reason: Some(
"this is a test reason such as one that would be produced by the model".into(),
),
parsed_cmd: vec![],
};
chat.handle_codex_event(Event {
id: "sub-multi".into(),
@@ -478,6 +482,7 @@ fn exec_approval_decision_truncates_multiline_and_long_commands() {
command: vec!["bash".into(), "-lc".into(), long],
cwd: std::env::current_dir().unwrap_or_else(|_| PathBuf::from(".")),
reason: None,
parsed_cmd: vec![],
};
chat.handle_codex_event(Event {
id: "sub-long".into(),
@@ -503,10 +508,7 @@ fn begin_exec(chat: &mut ChatWidget, call_id: &str, raw_cmd: &str) {
// Build the full command vec and parse it using core's parser,
// then convert to protocol variants for the event payload.
let command = vec!["bash".to_string(), "-lc".to_string(), raw_cmd.to_string()];
let parsed_cmd: Vec<ParsedCommand> = codex_core::parse_command::parse_command(&command)
.into_iter()
.map(Into::into)
.collect();
let parsed_cmd: Vec<ParsedCommand> = codex_core::parse_command::parse_command(&command);
chat.handle_codex_event(Event {
id: call_id.to_string(),
msg: EventMsg::ExecCommandBegin(ExecCommandBeginEvent {
@@ -674,6 +676,18 @@ fn streaming_final_answer_keeps_task_running_state() {
assert!(chat.bottom_pane.ctrl_c_quit_hint_visible());
}
#[test]
fn ctrl_c_shutdown_ignores_caps_lock() {
let (mut chat, _rx, mut op_rx) = make_chatwidget_manual();
chat.handle_key_event(KeyEvent::new(KeyCode::Char('C'), KeyModifiers::CONTROL));
match op_rx.try_recv() {
Ok(Op::Shutdown) => {}
other => panic!("expected Op::Shutdown, got {other:?}"),
}
}
#[test]
fn exec_history_cell_shows_working_then_completed() {
let (mut chat, mut rx, _op_rx) = make_chatwidget_manual();
@@ -1191,10 +1205,7 @@ async fn binary_size_transcript_snapshot() {
call_id: e.call_id.clone(),
command: e.command,
cwd: e.cwd,
parsed_cmd: parsed_cmd
.into_iter()
.map(std::convert::Into::into)
.collect(),
parsed_cmd,
}),
}
}
@@ -1309,6 +1320,7 @@ fn approval_modal_exec_snapshot() {
reason: Some(
"this is a test reason such as one that would be produced by the model".into(),
),
parsed_cmd: vec![],
};
chat.handle_codex_event(Event {
id: "sub-approve".into(),
@@ -1352,6 +1364,7 @@ fn approval_modal_exec_without_reason_snapshot() {
command: vec!["bash".into(), "-lc".into(), "echo hello world".into()],
cwd: std::env::current_dir().unwrap_or_else(|_| PathBuf::from(".")),
reason: None,
parsed_cmd: vec![],
};
chat.handle_codex_event(Event {
id: "sub-approve-noreason".into(),
@@ -1561,6 +1574,7 @@ fn status_widget_and_approval_modal_snapshot() {
reason: Some(
"this is a test reason such as one that would be produced by the model".into(),
),
parsed_cmd: vec![],
};
chat.handle_codex_event(Event {
id: "sub-approve-exec".into(),
@@ -2044,9 +2058,10 @@ fn plan_update_renders_history_cell() {
}
#[test]
fn stream_error_is_rendered_to_history() {
fn stream_error_updates_status_indicator() {
let (mut chat, mut rx, _op_rx) = make_chatwidget_manual();
let msg = "stream error: stream disconnected before completion: idle timeout waiting for SSE; retrying 1/5 in 211ms…";
chat.bottom_pane.set_task_running(true);
let msg = "Re-connecting... 2/5";
chat.handle_codex_event(Event {
id: "sub-1".into(),
msg: EventMsg::StreamError(StreamErrorEvent {
@@ -2055,11 +2070,15 @@ fn stream_error_is_rendered_to_history() {
});
let cells = drain_insert_history(&mut rx);
assert!(!cells.is_empty(), "expected a history cell for StreamError");
let blob = lines_to_single_string(cells.last().unwrap());
assert!(blob.contains('⚠'));
assert!(blob.contains("stream error:"));
assert!(blob.contains("idle timeout waiting for SSE"));
assert!(
cells.is_empty(),
"expected no history cell for StreamError event"
);
let status = chat
.bottom_pane
.status_widget()
.expect("status indicator should be visible");
assert_eq!(status.header(), msg);
}
#[test]
@@ -2222,17 +2241,15 @@ fn chatwidget_exec_and_status_layout_vt100_snapshot() {
command: vec!["bash".into(), "-lc".into(), "rg \"Change Approved\"".into()],
cwd: std::env::current_dir().unwrap_or_else(|_| PathBuf::from(".")),
parsed_cmd: vec![
codex_core::parse_command::ParsedCommand::Search {
ParsedCommand::Search {
query: Some("Change Approved".into()),
path: None,
cmd: "rg \"Change Approved\"".into(),
}
.into(),
codex_core::parse_command::ParsedCommand::Read {
},
ParsedCommand::Read {
name: "diff_render.rs".into(),
cmd: "cat diff_render.rs".into(),
}
.into(),
},
],
}),
});

View File

@@ -48,10 +48,16 @@ pub(crate) fn new_active_exec_command(
})
}
#[derive(Clone)]
pub(crate) struct OutputLines {
pub(crate) lines: Vec<Line<'static>>,
pub(crate) omitted: Option<usize>,
}
pub(crate) fn output_lines(
output: Option<&CommandOutput>,
params: OutputLinesParams,
) -> Vec<Line<'static>> {
) -> OutputLines {
let OutputLinesParams {
only_err,
include_angle_pipe,
@@ -63,9 +69,19 @@ pub(crate) fn output_lines(
stderr,
..
} = match output {
Some(output) if only_err && output.exit_code == 0 => return vec![],
Some(output) if only_err && output.exit_code == 0 => {
return OutputLines {
lines: Vec::new(),
omitted: None,
};
}
Some(output) => output,
None => return vec![],
None => {
return OutputLines {
lines: Vec::new(),
omitted: None,
};
}
};
let src = if *exit_code == 0 { stdout } else { stderr };
@@ -73,7 +89,7 @@ pub(crate) fn output_lines(
let total = lines.len();
let limit = TOOL_CALL_MAX_LINES;
let mut out = Vec::new();
let mut out: Vec<Line<'static>> = Vec::new();
let head_end = total.min(limit);
for (i, raw) in lines[..head_end].iter().enumerate() {
@@ -93,6 +109,11 @@ pub(crate) fn output_lines(
}
let show_ellipsis = total > 2 * limit;
let omitted = if show_ellipsis {
Some(total - 2 * limit)
} else {
None
};
if show_ellipsis {
let omitted = total - 2 * limit;
out.push(format!("… +{omitted} lines").into());
@@ -114,7 +135,10 @@ pub(crate) fn output_lines(
out.push(line);
}
out
OutputLines {
lines: out,
omitted,
}
}
pub(crate) fn spinner(start_time: Option<Instant>) -> Span<'static> {
@@ -371,7 +395,7 @@ impl ExecCell {
}
if let Some(output) = call.output.as_ref() {
let raw_output_lines = output_lines(
let raw_output = output_lines(
Some(output),
OutputLinesParams {
only_err: false,
@@ -380,15 +404,18 @@ impl ExecCell {
},
);
if raw_output_lines.is_empty() {
if raw_output.lines.is_empty() {
lines.extend(prefix_lines(
vec![Line::from("(no output)".dim())],
Span::from(layout.output_block.initial_prefix).dim(),
Span::from(layout.output_block.subsequent_prefix),
));
} else {
let trimmed_output =
Self::truncate_lines_middle(&raw_output_lines, layout.output_max_lines);
let trimmed_output = Self::truncate_lines_middle(
&raw_output.lines,
layout.output_max_lines,
raw_output.omitted,
);
let mut wrapped_output: Vec<Line<'static>> = Vec::new();
let output_wrap_width = layout.output_block.wrap_width(width);
@@ -427,7 +454,11 @@ impl ExecCell {
out
}
fn truncate_lines_middle(lines: &[Line<'static>], max: usize) -> Vec<Line<'static>> {
fn truncate_lines_middle(
lines: &[Line<'static>],
max: usize,
omitted_hint: Option<usize>,
) -> Vec<Line<'static>> {
if max == 0 {
return Vec::new();
}
@@ -435,7 +466,17 @@ impl ExecCell {
return lines.to_vec();
}
if max == 1 {
return vec![Self::ellipsis_line(lines.len())];
// Carry forward any previously omitted count and add any
// additionally hidden content lines from this truncation.
let base = omitted_hint.unwrap_or(0);
// When an existing ellipsis is present, `lines` already includes
// that single representation line; exclude it from the count of
// additionally omitted content lines.
let extra = lines
.len()
.saturating_sub(usize::from(omitted_hint.is_some()));
let omitted = base + extra;
return vec![Self::ellipsis_line(omitted)];
}
let head = (max - 1) / 2;
@@ -446,8 +487,12 @@ impl ExecCell {
out.extend(lines[..head].iter().cloned());
}
let omitted = lines.len().saturating_sub(head + tail);
out.push(Self::ellipsis_line(omitted));
let base = omitted_hint.unwrap_or(0);
let additional = lines
.len()
.saturating_sub(head + tail)
.saturating_sub(usize::from(omitted_hint.is_some()));
out.push(Self::ellipsis_line(base + additional));
if tail > 0 {
out.extend(lines[lines.len() - tail..].iter().cloned());

View File

@@ -114,7 +114,11 @@ impl HistoryCell for UserHistoryCell {
fn display_lines(&self, width: u16) -> Vec<Line<'static>> {
let mut lines: Vec<Line<'static>> = Vec::new();
let wrap_width = width.saturating_sub(LIVE_PREFIX_COLS);
let wrap_width = width
.saturating_sub(
LIVE_PREFIX_COLS + 1, /* keep a one-column right margin for wrapping */
)
.max(1);
let style = user_message_style();
@@ -125,7 +129,8 @@ impl HistoryCell for UserHistoryCell {
.map(|l| Line::from(l).style(style))
.collect::<Vec<_>>(),
// Wrap algorithm matches textarea.rs.
RtOptions::new(wrap_width as usize).wrap_algorithm(textwrap::WrapAlgorithm::FirstFit),
RtOptions::new(usize::from(wrap_width))
.wrap_algorithm(textwrap::WrapAlgorithm::FirstFit),
);
lines.push(Line::from("").style(style));
@@ -1062,11 +1067,6 @@ pub(crate) fn new_error_event(message: String) -> PlainHistoryCell {
PlainHistoryCell { lines }
}
pub(crate) fn new_stream_error_event(message: String) -> PlainHistoryCell {
let lines: Vec<Line<'static>> = vec![vec![padded_emoji("⚠️").into(), message.dim()].into()];
PlainHistoryCell { lines }
}
/// Render a userfriendly plan update styled like a checkbox todo list.
pub(crate) fn new_plan_update(update: UpdatePlanArgs) -> PlanUpdateCell {
let UpdatePlanArgs { explanation, plan } = update;
@@ -1153,7 +1153,7 @@ pub(crate) fn new_patch_apply_failure(stderr: String) -> PlainHistoryCell {
lines.push(Line::from("✘ Failed to apply patch".magenta().bold()));
if !stderr.trim().is_empty() {
lines.extend(output_lines(
let output = output_lines(
Some(&CommandOutput {
exit_code: 1,
stdout: String::new(),
@@ -1165,7 +1165,8 @@ pub(crate) fn new_patch_apply_failure(stderr: String) -> PlainHistoryCell {
include_angle_pipe: true,
include_prefix: true,
},
));
);
lines.extend(output.lines);
}
PlainHistoryCell { lines }

View File

@@ -310,6 +310,7 @@ async fn run_ratatui_app(
let current_version = env!("CARGO_PKG_VERSION");
let exe = std::env::current_exe()?;
let managed_by_bun = std::env::var_os("CODEX_MANAGED_BY_BUN").is_some();
let managed_by_npm = std::env::var_os("CODEX_MANAGED_BY_NPM").is_some();
let mut content_lines: Vec<Line<'static>> = vec![
@@ -330,7 +331,14 @@ async fn run_ratatui_app(
Line::from(""),
];
if managed_by_npm {
if managed_by_bun {
let bun_cmd = "bun install -g @openai/codex@latest";
content_lines.push(Line::from(vec![
"Run ".into(),
bun_cmd.cyan(),
" to update.".into(),
]));
} else if managed_by_npm {
let npm_cmd = "npm install -g @openai/codex@latest";
content_lines.push(Line::from(vec![
"Run ".into(),

View File

@@ -26,6 +26,7 @@ use ratatui::widgets::Clear;
use ratatui::widgets::Paragraph;
use ratatui::widgets::Widget;
use ratatui::widgets::WidgetRef;
use ratatui::widgets::Wrap;
pub(crate) enum Overlay {
Transcript(TranscriptOverlay),
@@ -329,9 +330,9 @@ struct CachedRenderable {
}
impl CachedRenderable {
fn new(renderable: Box<dyn Renderable>) -> Self {
fn new(renderable: impl Into<Box<dyn Renderable>>) -> Self {
Self {
renderable,
renderable: renderable.into(),
height: std::cell::Cell::new(None),
last_width: std::cell::Cell::new(None),
}
@@ -400,19 +401,19 @@ impl TranscriptOverlay {
.flat_map(|(i, c)| {
let mut v: Vec<Box<dyn Renderable>> = Vec::new();
let mut cell_renderable = if c.as_any().is::<UserHistoryCell>() {
Box::new(CachedRenderable::new(Box::new(CellRenderable {
Box::new(CachedRenderable::new(CellRenderable {
cell: c.clone(),
style: if highlight_cell == Some(i) {
user_message_style().reversed()
} else {
user_message_style()
},
}))) as Box<dyn Renderable>
})) as Box<dyn Renderable>
} else {
Box::new(CachedRenderable::new(Box::new(CellRenderable {
Box::new(CachedRenderable::new(CellRenderable {
cell: c.clone(),
style: Style::default(),
}))) as Box<dyn Renderable>
})) as Box<dyn Renderable>
};
if !c.is_stream_continuation() && i > 0 {
cell_renderable = Box::new(InsetRenderable::new(
@@ -496,12 +497,8 @@ pub(crate) struct StaticOverlay {
impl StaticOverlay {
pub(crate) fn with_title(lines: Vec<Line<'static>>, title: String) -> Self {
Self::with_renderables(
vec![Box::new(CachedRenderable::new(Box::new(Paragraph::new(
Text::from(lines),
))))],
title,
)
let paragraph = Paragraph::new(Text::from(lines)).wrap(Wrap { trim: false });
Self::with_renderables(vec![Box::new(CachedRenderable::new(paragraph))], title)
}
pub(crate) fn with_renderables(renderables: Vec<Box<dyn Renderable>>, title: String) -> Self {
@@ -812,6 +809,18 @@ mod tests {
assert_snapshot!(term.backend());
}
#[test]
fn static_overlay_wraps_long_lines() {
let mut overlay = StaticOverlay::with_title(
vec!["a very long line that should wrap when rendered within a narrow pager overlay width".into()],
"S T A T I C".to_string(),
);
let mut term = Terminal::new(TestBackend::new(24, 8)).expect("term");
term.draw(|f| overlay.render(f.area(), f.buffer_mut()))
.expect("draw");
assert_snapshot!(term.backend());
}
#[test]
fn pager_view_content_height_counts_renderables() {
let pv = PagerView::new(

View File

@@ -3,6 +3,6 @@ source: tui/src/history_cell.rs
expression: rendered
---
one two
three four
five six
seven
three
four five
six seven

View File

@@ -0,0 +1,13 @@
---
source: tui/src/pager_overlay.rs
assertion_line: 798
expression: term.backend()
---
"/ S T A T I C / / / / / "
"a very long line that "
"should wrap when "
"rendered within a narrow"
"─────────────────── 0% ─"
" ↑/↓ to scroll pgup/pg"
" q to quit "
" "

View File

@@ -103,6 +103,11 @@ impl StatusIndicatorWidget {
}
}
#[cfg(test)]
pub(crate) fn header(&self) -> &str {
&self.header
}
/// Replace the queued messages displayed beneath the header.
pub(crate) fn set_queued_messages(&mut self, queued: Vec<String>) {
self.queued_messages = queued;

View File

@@ -2492,7 +2492,7 @@
{"ts":"2025-08-09T15:51:59.856Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:51:59.858Z","dir":"to_tui","kind":"app_event","variant":"Redraw"}
{"ts":"2025-08-09T15:51:59.939Z","dir":"to_tui","kind":"log_line","line":"[INFO codex_core::codex] FunctionCall: {\"command\":[\"bash\",\"-lc\",\"just fix\"],\"with_escalated_permissions\":true,\"justifica"}
{"ts":"2025-08-09T15:51:59.939Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_KOxVodT3X5ci7LJmudvcovhW","command":["bash","-lc","just fix"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Run clippy with network and system permissions to apply lint fixes across workspace."}}}
{"ts":"2025-08-09T15:51:59.939Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_KOxVodT3X5ci7LJmudvcovhW","command":["bash","-lc","just fix"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Run clippy with network and system permissions to apply lint fixes across workspace.","parsed_cmd":[]}}}
{"ts":"2025-08-09T15:51:59.939Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:51:59.939Z","dir":"to_tui","kind":"insert_history","lines":5}
{"ts":"2025-08-09T15:51:59.939Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
@@ -4172,7 +4172,7 @@
{"ts":"2025-08-09T15:53:09.375Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:53:09.376Z","dir":"to_tui","kind":"app_event","variant":"Redraw"}
{"ts":"2025-08-09T15:53:09.448Z","dir":"to_tui","kind":"log_line","line":"[INFO codex_core::codex] FunctionCall: {\"command\":[\"bash\",\"-lc\",\"just fix\"],\"with_escalated_permissions\":true,\"justifica"}
{"ts":"2025-08-09T15:53:09.448Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_POl3hxI2xeszBLv9IOM7L2ir","command":["bash","-lc","just fix"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Clippy needs broader permissions; allow to run and apply lint fixes."}}}
{"ts":"2025-08-09T15:53:09.448Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_POl3hxI2xeszBLv9IOM7L2ir","command":["bash","-lc","just fix"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Clippy needs broader permissions; allow to run and apply lint fixes.","parsed_cmd":[]}}}
{"ts":"2025-08-09T15:53:09.448Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:53:09.449Z","dir":"to_tui","kind":"insert_history","lines":5}
{"ts":"2025-08-09T15:53:09.449Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
@@ -7776,7 +7776,7 @@
{"ts":"2025-08-09T15:58:28.583Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:58:28.590Z","dir":"to_tui","kind":"app_event","variant":"Redraw"}
{"ts":"2025-08-09T15:58:28.594Z","dir":"to_tui","kind":"log_line","line":"[INFO codex_core::codex] FunctionCall: {\"command\":[\"bash\",\"-lc\",\"cargo test -p codex-core shell::tests::test_current_she"}
{"ts":"2025-08-09T15:58:28.594Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_iMa8Qnw0dYLba4rVysxebmkV","command":["bash","-lc","cargo test -p codex-core shell::tests::test_current_shell_detects_zsh -- --nocapture"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Run the macOS shell detection test without sandbox limits so dscl can read user shell."}}}
{"ts":"2025-08-09T15:58:28.594Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_iMa8Qnw0dYLba4rVysxebmkV","command":["bash","-lc","cargo test -p codex-core shell::tests::test_current_shell_detects_zsh -- --nocapture"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Run the macOS shell detection test without sandbox limits so dscl can read user shell.","parsed_cmd":[]}}}
{"ts":"2025-08-09T15:58:28.594Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:58:28.594Z","dir":"to_tui","kind":"insert_history","lines":5}
{"ts":"2025-08-09T15:58:28.594Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
@@ -8730,7 +8730,7 @@
{"ts":"2025-08-09T15:59:01.983Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:59:01.985Z","dir":"to_tui","kind":"app_event","variant":"Redraw"}
{"ts":"2025-08-09T15:59:02.005Z","dir":"to_tui","kind":"log_line","line":"[INFO codex_core::codex] FunctionCall: {\"command\":[\"bash\",\"-lc\",\"cargo test --all-features\"],\"with_escalated_permissions"}
{"ts":"2025-08-09T15:59:02.005Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_JDFGIuFhYCIiQO1Aq2L9lBO1","command":["bash","-lc","cargo test --all-features"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Run full test suite without sandbox constraints to validate the merge."}}}
{"ts":"2025-08-09T15:59:02.005Z","dir":"to_tui","kind":"codex_event","payload":{"id":"1","msg":{"type":"exec_approval_request","call_id":"call_JDFGIuFhYCIiQO1Aq2L9lBO1","command":["bash","-lc","cargo test --all-features"],"cwd":"/Users/easong/code/codex/codex-rs","reason":"Run full test suite without sandbox constraints to validate the merge.","parsed_cmd":[]}}}
{"ts":"2025-08-09T15:59:02.006Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}
{"ts":"2025-08-09T15:59:02.006Z","dir":"to_tui","kind":"insert_history","lines":5}
{"ts":"2025-08-09T15:59:02.006Z","dir":"to_tui","kind":"app_event","variant":"RequestRedraw"}

View File

@@ -370,9 +370,8 @@ API_KEY = "value"
experimental_use_rmcp_client = true
[mcp_servers.figma]
url = "https://mcp.linear.app/mcp"
# Optional bearer token to be passed into an `Authorization: Bearer <token>` header
# Use this with caution because the token is in plaintext and can be read by Codex itself.
bearer_token = "<token>"
# Optional environment variable containing a bearer token to use for auth
bearer_token_env_var = "<token>"
```
For oauth login, you must enable `experimental_use_rmcp_client = true` and then run `codex mcp login server_name`
@@ -603,6 +602,7 @@ Specify a program that will be executed to get notified about events generated b
```json
{
"type": "agent-turn-complete",
"thread-id": "b5f6c1c2-1111-2222-3333-444455556666",
"turn-id": "12345",
"input-messages": ["Rename `foo` to `bar` and update the callsites."],
"last-assistant-message": "Rename complete and verified `cargo build` succeeds."
@@ -611,6 +611,8 @@ Specify a program that will be executed to get notified about events generated b
The `"type"` property will always be set. Currently, `"agent-turn-complete"` is the only notification type that is supported.
`"thread-id"` contains a string that identifies the Codex session that produced the notification; you can use it to correlate multiple turns that belong to the same task.
As an example, here is a Python script that parses the JSON and decides whether to show a desktop push notification using [terminal-notifier](https://github.com/julienXX/terminal-notifier) on macOS:
```python
@@ -645,6 +647,8 @@ def main() -> int:
print(f"not sending a push notification for: {notification_type}")
return 0
thread_id = notification.get("thread-id", "")
subprocess.check_output(
[
"terminal-notifier",
@@ -653,7 +657,7 @@ def main() -> int:
"-message",
message,
"-group",
"codex",
"codex-" + thread_id,
"-ignoreDnD",
"-activate",
"com.googlecode.iterm2",