(cleanup) add link to https://cookbook.openai.com for apply_patch and minor tidy up by tibo-openai · Pull Request #2 · openai/codex

tibo-openai · 2025-04-16T17:08:44Z

No description provided.

Fix Dockerfile

…vider ✨ Add GitHub Models provider

Critical bugfix - auto-update detection now working Fixes: - Version parser now handles -termux suffix correctly - Tag parser accepts both rust-v* and v*-termux formats - Added test coverage for Termux tag validation Technical details: - v0.58.0 upstream merge overwrote -termux suffix support - New upstream code rejected v*-termux tag format - Prevented version.json creation and update notifications Affected versions: v0.58.0 through v0.58.3 Termux patches (4 total): - Patch #1: Browser login fix (termux-open-url) - Patch #2: RAM optimizations (lto=false, codegen-units=16) - Patch #3: Auto-update URL (@mmmbuto/codex-cli-termux) - Patch #4: Auto-update detection (this release)

Upstream changes (40+ commits): - New: codex-shell-tool-mcp MCP server - New: execpolicycheck CLI command - New: TUI animations toggle feature - New: Shell timeout increased to 1 hour - Refactor: execpolicy migration (execpolicy2 -> execpolicy) - Removed: tiktoken-rs, shell_command feature - Improved: FreeBSD portability, fuzzy search (8->20 results) Termux patches verified compatible: - #1 Browser login (termux-open-url) ✅ - #2 RAM optimizations (lto=false, codegen-units=16) ✅ - #8 Bash execution (sandbox, LD_*, shell detection) ✅ - #9 Auto-update execution ✅ Stats: 195 files changed, +5915 insertions, -2293 deletions 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>

- 8 new tests to validate Termux patches after upstream merge - Critical tests for Patch #1, #2, #4, #9 (conflicting patches) - Test auto-update execution, browser login, RAM optimization - Essential for post-merge validation before npm publish Test suite v1.2 → v1.3 (90 total tests)

Merged 61 commits from openai/codex upstream (Nov 21-27, 2025). ✅ ALL TERMUX PATCHES PRESERVED: - Patch #1: Browser login (termux-open-url) - VERIFIED ✅ - Patch #2: RAM optimization (lto=false) - VERIFIED ✅ - Patch #4: Auto-update URL (DioNanos/codex-termux) - VERIFIED ✅ - Patch #5: Version parser (-termux suffix) - VERIFIED ✅ - Patch #6: NPM package name (@mmmbuto/codex-cli-termux) - VERIFIED ✅ - Patch #9: Auto-update execution (npm install) - VERIFIED ✅ NEW UPSTREAM FEATURES: - Enterprise skip upgrade config (check_for_update_on_startup) - Unified exec pruning strategy - Custom env for unified exec process - Compaction events - Thread/turn ID in notifications - Config management app-server - Cargo audit workflow + cargo-deny - NetBSD build fixes - Windows sandbox enhancements - Documentation improvements UPSTREAM CHANGES: - 232 files changed - +11247 insertions, -7808 deletions - New codex-api/ and codex-client/ crates - Rollout session error improvements - Apply patch parsing fixes CONFLICTS RESOLVED: - README.md: Merged Testing section (Termux) + Documentation links (upstream) NEXT: Compilation + Category 13 test suite validation

- Updated workspace version to 0.80.0-termux - Termux patch #2 (compilation) preserved: lto=false, codegen-units=16 - Termux patch #9 (auto-update execution) preserved in main.rs - Process hardening removal upstream improves bash execution (patch #8) Upstream changes (77 commits from rust-v0.79.0): - Metrics capabilities (otel/metrics modules) - Requirements/list API - Thread fork endpoints - Elevated sandbox onboarding - Skills explicit invocation via V2 API - Updated dependencies: lru 0.16.3, zstd added - Removed codex-process-hardening from Codex CLI - Updated models.json

- Set version = "0.95.0" - Applied Patch #2: lto = false, codegen-units = 16 - Preserved all upstream v0.95.0 workspace members and dependencies - All Termux patches (#1-#6, #9) verified present - Next step: Build ARM64 binaries

Patches applied: - Patch #1: Browser login fix (termux-open-url) - Patch #2: Compilation optimizations (lto=false, codegen-units=16) - Patch #3: Version alignment (0.104.0) - Patch #4: Auto-update URL (DioNanos/codex-termux) - Patch #5: Version parser (-termux suffix) - Patch #6: NPM package name (@mmmbuto/codex-cli-termux) - Patch #9: Auto-update execution (tui/src/main.rs) Updated files: - codex-rs/Cargo.toml: version 0.104.0 - codex-rs/tui/src/main.rs: add update_action execution - npm-package/package.json: 0.104.0-termux - patches/README.md: update validated version

…i#21190) ## Why We found this while reviewing openai#21091, but confirmed it is not introduced by that PR: the order-sensitive `current_text_with_pending()` replacement loop already existed, and `main` already allowed active same-size large pastes to use prefix-overlapping labels such as `[Pasted Content N chars]` and `[Pasted Content N chars] openai#2`. openai#21091 fixes placeholder numbering after a draft is cleared, so a fresh same-size paste can reuse the base label. This PR fixes a different path: when a draft already contains multiple active same-size large pastes, the placeholders can overlap by prefix, for example `[Pasted Content N chars]` and `[Pasted Content N chars] openai#2`. That overlap breaks `current_text_with_pending()` when the composer materializes the draft text for the external editor. Replacing the base placeholder first can partially rewrite the `openai#2` placeholder, leaving the external editor seeded with corrupted text instead of both paste payloads. | Before | After | |---|---| | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 18 09" src="https://github.com/user-attachments/assets/88a2936c-cf00-4adc-8567-8fd8f398b4a8" /> | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 20 31" src="https://github.com/user-attachments/assets/119cff52-43c8-432a-9367-418d82f4ed82" /> | | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 18 57" src="https://github.com/user-attachments/assets/026031bb-839b-4252-a0fd-9ba9616435fe" /> | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 21 31" src="https://github.com/user-attachments/assets/8cb6f2c8-3a5d-411b-8623-dca666ee3c08" /> | ## What Changed - Changed `current_text_with_pending()` to expand pending pastes through the existing element-range based `expand_pending_pastes()` helper instead of global string replacement. - Added a regression test with two different same-length large pastes to ensure both overlapping placeholders expand to their original payloads. ## How to Test 1. Start Codex TUI. 2. Paste a large string, for example 1004 `A` characters. ```shell perl -e 'print "A" x 1004' | pbcopy ``` 3. Paste a second large string with the same length, for example 1004 `B` characters. ```shell perl -e 'print "B" x 1004' | pbcopy ``` 4. Open the external editor from the composer. 5. Confirm the editor is seeded with the full `A...` payload followed by the full `B...` payload, with no literal `openai#2` left behind. Targeted tests: - `cargo test -p codex-tui current_text_with_pending_expands_overlapping_placeholders` - `just argument-comment-lint-from-source -p codex-tui` I also ran `cargo test -p codex-tui`; it reached the full crate suite but failed two unrelated local status tests because this machine's `/etc/codex/requirements.toml` rejects `DangerFullAccess`.

…ub catalog Closes the test-coverage Blocking findings from QA-002 (rust-qa-agent), test-auditor, rust-code-reviewer Important #2, and simplification F6: - core/src/session/tests.rs: integration test exercising swap_model_provider end-to-end. Asserts both halves of the model_provider_id <-> model_provider invariant move together (services.model_client provider id flips, state.session_configuration.provider flips, original_config_do_not_use.model_provider_id flips, auth_manager Arc unchanged). This is the test that would have caught the Critical bug fixed in 04cc95d. - core/src/config/edit_tests.rs: SetModelAndProvider TOML round-trip at profile scope, mirroring existing SetModel test shape. Plus a None-effort test that asserts a pre-existing effort value is preserved. - tui/src/model_catalog.rs: two unit tests for the try_list_models conditional stub — one with ollama in configured_provider_ids, one without — proving the code-reviewer Important #3 guard works. Drive-by cleanups (within Milestone-1 scope; bigger ones deferred per simplification F2/F3/F5): - tui/src/chatwidget.rs: deleted the model_selection_actions one-line shim (simplification F4 + rust-code-reviewer Important #4) and updated its single call site to pass preset.provider_id.clone() through directly. Eliminates a latent bug where a future local preset satisfying is_auto_model() would silently drop its provider id. - tui/src/app_command.rs: removed spurious #[allow(clippy::too_many_arguments)] from a 4-arg fn (QA-003). - core/src/session/mod.rs: added SAFETY doc on swap_model_provider explaining the available_during_task() == false invariant that makes the two-phase lock pattern safe (QA-004 documentation requirement). - core/src/session/session.rs: TODO(milestone-1-followup) at SessionSettingsUpdate.model_provider_id field to track the simplification F2 dead-field cleanup (high-churn, deferred).

Pulls in three CC improvements ecodex was missing: 1. Praxic phase emoji ⚙ → 🔨 — east-asian-width fix. ⚙ (U+2699) renders narrow on some terminals + ambiguous on others, causing digit overlap or clipping next to the phase indicator. 🔨 (U+1F528) is unambiguously wide, matches CLAUDE_CODE_SETUP.md docs. 2. Threshold info no longer rendered in live statusline — it's Sentinel-scoped (set/dynamically-adjusted at session level, not per-action) and not actionable for the AI mid-tool-call. Keeping it cluttered the line without informing decisions. Helpers (get_dynamic_threshold, format_threshold) remain for `empirica sentinel-status` and debug surfaces. 3. Caller-side threshold fetch removed (paired with #2) — saves a DB round trip per statusline render tick. Verbatim copy of CC's version (1461 lines vs ecodex's prior 1455 — the +6 is the threshold-removal comment block). No ecodex-specific patches to preserve in the diff. Also synced to ~/.codex/plugins/cache/nubaeon/empirica/0.1.0/ so next session picks it up. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

- Remove numeric prefixes for disabled rows in shared list rendering. These numbers are shortcuts, Ex: Pressing "2" selects option `openai#2`. Disabled items can not be selected, so keeping numbers on these items is misleading. - Apply the same behavior in both tui and tui_app_server. - Update affected snapshots for apps/plugins loading and plugin detail rows. _**This is a global change.**_ Before: <img width="1680" height="488" alt="image" src="https://github.com/user-attachments/assets/4bcf94ad-285f-48d3-a235-a85b58ee58e2" /> After: <img width="1706" height="484" alt="image" src="https://github.com/user-attachments/assets/76bb6107-a562-42fe-ae94-29440447ca77" />

…i#21190) ## Why We found this while reviewing openai#21091, but confirmed it is not introduced by that PR: the order-sensitive `current_text_with_pending()` replacement loop already existed, and `main` already allowed active same-size large pastes to use prefix-overlapping labels such as `[Pasted Content N chars]` and `[Pasted Content N chars] openai#2`. openai#21091 fixes placeholder numbering after a draft is cleared, so a fresh same-size paste can reuse the base label. This PR fixes a different path: when a draft already contains multiple active same-size large pastes, the placeholders can overlap by prefix, for example `[Pasted Content N chars]` and `[Pasted Content N chars] openai#2`. That overlap breaks `current_text_with_pending()` when the composer materializes the draft text for the external editor. Replacing the base placeholder first can partially rewrite the `openai#2` placeholder, leaving the external editor seeded with corrupted text instead of both paste payloads. | Before | After | |---|---| | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 18 09" src="https://github.com/user-attachments/assets/88a2936c-cf00-4adc-8567-8fd8f398b4a8" /> | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 20 31" src="https://github.com/user-attachments/assets/119cff52-43c8-432a-9367-418d82f4ed82" /> | | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 18 57" src="https://github.com/user-attachments/assets/026031bb-839b-4252-a0fd-9ba9616435fe" /> | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 21 31" src="https://github.com/user-attachments/assets/8cb6f2c8-3a5d-411b-8623-dca666ee3c08" /> | ## What Changed - Changed `current_text_with_pending()` to expand pending pastes through the existing element-range based `expand_pending_pastes()` helper instead of global string replacement. - Added a regression test with two different same-length large pastes to ensure both overlapping placeholders expand to their original payloads. ## How to Test 1. Start Codex TUI. 2. Paste a large string, for example 1004 `A` characters. ```shell perl -e 'print "A" x 1004' | pbcopy ``` 3. Paste a second large string with the same length, for example 1004 `B` characters. ```shell perl -e 'print "B" x 1004' | pbcopy ``` 4. Open the external editor from the composer. 5. Confirm the editor is seeded with the full `A...` payload followed by the full `B...` payload, with no literal `openai#2` left behind. Targeted tests: - `cargo test -p codex-tui current_text_with_pending_expands_overlapping_placeholders` - `just argument-comment-lint-from-source -p codex-tui` I also ran `cargo test -p codex-tui`; it reached the full crate suite but failed two unrelated local status tests because this machine's `/etc/codex/requirements.toml` rejects `DangerFullAccess`.

…e spec Probes 17/17b/17c/17d (ran keyed, SDK 0.3.178) establish the canUseTool contract for the chat REPL and flip two premises: - auto is NOT a no-prompt mode: it routes mutations (Edit) through the broker (2/2), identical to default; only bypassPermissions is silent (2/2). - canUseTool UI hints (title/displayName) are absent headlessly → dialog reconstructs from toolName+input. deny verified safe (denied Edit left file ORIGINAL, turn completed); allow {updatedInput} works; Read-gating is non-deterministic. Spec decision #2 corrected: REPL default = default mode (broker-live), bypassPermissions = the silence toggle.

- Remove numeric prefixes for disabled rows in shared list rendering. These numbers are shortcuts, Ex: Pressing "2" selects option `openai#2`. Disabled items can not be selected, so keeping numbers on these items is misleading. - Apply the same behavior in both tui and tui_app_server. - Update affected snapshots for apps/plugins loading and plugin detail rows. _**This is a global change.**_ Before: <img width="1680" height="488" alt="image" src="https://github.com/user-attachments/assets/4bcf94ad-285f-48d3-a235-a85b58ee58e2" /> After: <img width="1706" height="484" alt="image" src="https://github.com/user-attachments/assets/76bb6107-a562-42fe-ae94-29440447ca77" />

…i#21190) ## Why We found this while reviewing openai#21091, but confirmed it is not introduced by that PR: the order-sensitive `current_text_with_pending()` replacement loop already existed, and `main` already allowed active same-size large pastes to use prefix-overlapping labels such as `[Pasted Content N chars]` and `[Pasted Content N chars] openai#2`. openai#21091 fixes placeholder numbering after a draft is cleared, so a fresh same-size paste can reuse the base label. This PR fixes a different path: when a draft already contains multiple active same-size large pastes, the placeholders can overlap by prefix, for example `[Pasted Content N chars]` and `[Pasted Content N chars] openai#2`. That overlap breaks `current_text_with_pending()` when the composer materializes the draft text for the external editor. Replacing the base placeholder first can partially rewrite the `openai#2` placeholder, leaving the external editor seeded with corrupted text instead of both paste payloads. | Before | After | |---|---| | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 18 09" src="https://github.com/user-attachments/assets/88a2936c-cf00-4adc-8567-8fd8f398b4a8" /> | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 20 31" src="https://github.com/user-attachments/assets/119cff52-43c8-432a-9367-418d82f4ed82" /> | | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 18 57" src="https://github.com/user-attachments/assets/026031bb-839b-4252-a0fd-9ba9616435fe" /> | <img width="1230" height="1008" alt="CleanShot 2026-05-05 at 10 21 31" src="https://github.com/user-attachments/assets/8cb6f2c8-3a5d-411b-8623-dca666ee3c08" /> | ## What Changed - Changed `current_text_with_pending()` to expand pending pastes through the existing element-range based `expand_pending_pastes()` helper instead of global string replacement. - Added a regression test with two different same-length large pastes to ensure both overlapping placeholders expand to their original payloads. ## How to Test 1. Start Codex TUI. 2. Paste a large string, for example 1004 `A` characters. ```shell perl -e 'print "A" x 1004' | pbcopy ``` 3. Paste a second large string with the same length, for example 1004 `B` characters. ```shell perl -e 'print "B" x 1004' | pbcopy ``` 4. Open the external editor from the composer. 5. Confirm the editor is seeded with the full `A...` payload followed by the full `B...` payload, with no literal `openai#2` left behind. Targeted tests: - `cargo test -p codex-tui current_text_with_pending_expands_overlapping_placeholders` - `just argument-comment-lint-from-source -p codex-tui` I also ran `cargo test -p codex-tui`; it reached the full crate suite but failed two unrelated local status tests because this machine's `/etc/codex/requirements.toml` rejects `DangerFullAccess`.

w

e2e64d9

tibo-openai changed the title ~~(cleanup) add link to https://cookbook.openai.com and minor tidy up~~ (cleanup) add link to https://cookbook.openai.com for apply_patch and minor tidy up Apr 16, 2025

ibigio merged commit 1c26c27 into main Apr 16, 2025

ibigio deleted the tibo/codex/1 branch April 16, 2025 17:15

rumple pushed a commit to rumple/codex that referenced this pull request Apr 24, 2025

Add link to cookbook (openai#2)

dca27c8

avalanchexia pushed a commit to avalanchexia/codex that referenced this pull request Jun 7, 2025

Merge pull request openai#2 from JayDoubleu/main

17f8286

Fix Dockerfile

salah9003 mentioned this pull request Aug 22, 2025

Feature Request: Multimodal Image Support via @ File Search #2589

Closed

mazdak pushed a commit to mazdak/codex that referenced this pull request Sep 3, 2025

Merge pull request openai#2 from a5c-incubator/feat/github-models-pro…

e1fc7f6

…vider ✨ Add GitHub Models provider

shepting pushed a commit to shepting/codex that referenced this pull request Sep 16, 2025

allow configuration of ollama base url (openai#2)

6f8686e

synestiqx mentioned this pull request Sep 28, 2025

30k token limit when using API after ChatGPT exhaustion (IDE) #4216

Closed

gitcnd mentioned this pull request Oct 24, 2025

Does nothing, with no errors, in Cursor #5592

Closed

techjoec mentioned this pull request Nov 2, 2025

Codex Cloud Rate Limit issue #6008

Closed

celia-oai mentioned this pull request Nov 20, 2025

[app-server & core] introduce new codex error code and v2 app-server error events #6938

Merged

charley-oai mentioned this pull request Jan 27, 2026

Fix large paste placeholder numbering after deletion #10032

Closed

locainin mentioned this pull request Jan 28, 2026

Rate limit reached for organization org- on requests per min (RPM): Limit 3000, Used 3000, Requested 1 #10060

Closed

atlascodesai mentioned this pull request Feb 2, 2026

Resume picker "Updated" column not sorted by updated time #10388

Closed

RyderFreeman4Logos mentioned this pull request Feb 8, 2026

feat(hooks): comprehensive hook system with lifecycle events and steering #11067

Closed

4 tasks

pthrasher mentioned this pull request Feb 9, 2026

Codex client output truncated when scrolling in Zellij #2558

Closed

charley-oai mentioned this pull request Feb 11, 2026

Fix pre-turn compaction context handling and remote compact edge cases #11234

Closed

claylong mentioned this pull request Feb 17, 2026

IDE-integrated diff / approval #2998

Open

This was referenced Mar 8, 2026

python-sdk: generated type foundation (all v2 schemas) #13953

Merged

python-sdk: public Codex API, thread/turn controls, and UX examples #13954

Closed

Bayar-Tsogtbaatar mentioned this pull request May 6, 2026

Lifecycle dependency for subagents: keep reviewer agent active until implementation agent completes #21380

Closed

tdinaizii mentioned this pull request May 14, 2026

Codex Desktop (Windows) 26.506.3741.0 — process exits silently ~4s after launch, no window, no logs #22633

Open

Mouriya-Emma mentioned this pull request May 16, 2026

Spike: 验证 codex embedded TUI 模式下通过外置 control socket 注入 turn 的可行性 mouriya-s-lab/codex#1

Closed

Pedro-Fernandes mentioned this pull request May 20, 2026

Codex v0.131.0 terminal corruption after codex_apps MCP startup on Windows #23512

Closed

kalineh mentioned this pull request May 24, 2026

Windows sandbox setup helper triggers UAC installer detection (os error 740) #24050

Closed

jshaofa-ui mentioned this pull request May 25, 2026

multi_agent_v1.close_agent can hang for hours when closing an unresponsive subagent #24389

Open

dryoungdo mentioned this pull request May 25, 2026

codex exec on macOS initializes external MCP and lists tools but never sends tools/call #14115

Open

sulmusic2-star mentioned this pull request May 26, 2026

Codex Desktop 26.519.22136: local project conversation history missing after update, threads still exist in state_5.sqlite #23979

Open

rpc1990 mentioned this pull request May 26, 2026

[VS Code extension] Fast speed preference periodically resets to Standard; default-service-tier is cleared despite user selection #24661

Open

Avishayy mentioned this pull request May 27, 2026

Codex Desktop: gpt-5.5 xhigh turn stalled 30m before first output, then resumed normally #24260

Open

YugantM mentioned this pull request May 30, 2026

Add HVTracker badge to README? #25350

Closed

jparta mentioned this pull request Jun 1, 2026

Add intraline diff highlighting to Codex review. #24575

Open

aditya-oss-prog mentioned this pull request Jun 9, 2026

GPT-5.5 xhigh via Amazon Bedrock stops automatically mid-task in Codex CLI, same as GPT-5.4 xhigh #26860

Open

gktdgktd mentioned this pull request Jun 12, 2026

Windows Computer Use @oai/sky exports mismatch still reproduces on 26.609.30741 / @oai/sky 0.4.10 #27747

Closed

mitre88 mentioned this pull request Jun 14, 2026

thread-store: a single BOM-prefixed rollout .jsonl fails the entire thread read and blocks message creation #28139

Open

bongsu-rapportlabs mentioned this pull request Jun 21, 2026

TypeScript SDK JSONL parser fails on multiline MCP tool results (patch available) #23131

Open

snailwei mentioned this pull request Jun 23, 2026

Bug Report: Codex App — LLM Returns Invalid JSON in Function Call Arguments #29549

Open

Keksuccino mentioned this pull request Jun 26, 2026

[Codex App] Confusing "Enter" behavior setting #30174

Open

renandecarlo mentioned this pull request Jun 26, 2026

Codex Desktop sidebar buttons + scroll + window buttons overlapping #30308

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

(cleanup) add link to https://cookbook.openai.com for apply_patch and minor tidy up#2

(cleanup) add link to https://cookbook.openai.com for apply_patch and minor tidy up#2
ibigio merged 1 commit into
mainfrom
tibo/codex/1

tibo-openai commented Apr 16, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

tibo-openai commented Apr 16, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants