James Peter
beee09fcc0
tweak(tui/spinner): set spinner generation reasoning effort to Medium\n\n- Use ReasoningEffort::Medium for the background spinner JSON-schema turn\n- Keeps quality while avoiding earlier Minimal/Low issues\n- Scoped only to spinner generation; main turns unchanged
2025-09-13 00:57:15 +10:00
James Peter
4041693f32
Merge origin/main: adopt remote changes (spinner auth pref)
2025-09-13 00:33:39 +10:00
James Peter
e3f313b796
fix(tui/spinner): use active auth preference for custom spinner generation\n\n- Prefer ChatGPT auth when session uses ChatGPT; otherwise use API key\n- Avoids 401 retry loops when background thread created a fresh AuthManager\n- Matches request shape that succeeds in harness (responses=v1, session handling)
2025-09-13 00:33:37 +10:00
GitHub Action
c2910bc85c
chore: remove stray file [!TIP] [skip ci]
2025-09-12 14:13:01 +00:00
GitHub Action
9fcc15a5d4
docs(changelog): update for v0.2.137 [skip ci]
2025-09-12 14:12:15 +00:00
GitHub Action
df4c9314a8
chore(release): 0.2.137 [skip ci]
2025-09-12 14:10:47 +00:00
James Peter
79c69f9642
feat(dev): add scripts/test-responses.js to probe Responses API with ChatGPT/API-key auth\n\n- Reads /auth.json; supports --auth chatgpt|api-key|auto\n- Tests: base, json-schema, tools-web-search, tools-web-search-preview, store\n- Uses Responses v1; non-streaming by default for clear error capture\n- Prints status + x-request-id and error excerpts
2025-09-13 00:02:33 +10:00
James Peter
92c2ecb8d9
Merge origin/main: adopt remote changes
2025-09-12 23:57:19 +10:00
James Peter
acfaeb7dd5
ci(proxy): set Responses v1 and fail-fast on 5xx\n\n- scripts/openai-proxy.js default RESPONSES_BETA=responses=v1\n- Workflows (issue-code, issue-comment, release, upstream-merge): EXIT_ON_5XX=1 and RESPONSES_BETA=responses=v1\n\nFix: API-key 2nd-call 500s addressed in client; proxy now aligned with production headers
2025-09-12 23:57:18 +10:00
James Peter
1ddedb8b48
chore(proxy): add STRICT_HEADERS and RESPONSES_BETA override; minimal allowlist mode
2025-09-12 23:24:59 +10:00
GitHub Action
8016c03cd9
chore(release): 0.2.136 [skip ci]
2025-09-12 13:21:10 +00:00
James Peter
ee7012e4fd
Merge origin/main: keep our proxy debug toggles
2025-09-12 23:11:43 +10:00
James Peter
a88f97c681
chore(proxy): add debug toggles and error-body logging\n\n- Options: DISABLE_KEEPALIVE, FORCE_CLOSE, NO_GZIP, STRIP_SESSION_ID, IDEMPOTENCY_KEY=auto, LOG_ERROR_BODY(+_BYTES)\n- Helps isolate Responses API 500s on subsequent requests
2025-09-12 23:11:31 +10:00
GitHub Action
ca64da78fa
chore(release): 0.2.135 [skip ci]
2025-09-12 12:53:09 +00:00
James Peter
20aae39f88
Merge origin/main: adopt remote version bumps; keep ours elsewhere (centralized proxy + triage cleanup)
2025-09-12 22:43:19 +10:00
James Peter
7304efeb2b
ci(proxy): centralize OpenAI proxy to scripts/openai-proxy.js and use across workflows\n\n- Replace inline JS proxies in upstream-merge, issue-code, issue-comment, release\n- Shared behavior: SSE defaults, allowlist, header scrubbing, structured logs\n- Supports PORT/OPENAI_BASE_URL/LOG_DEST/EXIT_ON_5XX via env\n- Keeps existing port assignments (5055/5056/5057/5058)\n
2025-09-12 22:43:05 +10:00
GitHub Action
77a6d1f732
chore(release): 0.2.134 [skip ci]
2025-09-12 12:28:58 +00:00
James Peter
7e9203c225
chore(dev): add local OpenAI proxy script replicating upstream-merge settings\n\n- scripts/openai-proxy.js mirrors CI proxy (allowlist, SSE defaults, header scrubbing)\n- Usage docs in file header; supports PORT and OPENAI_BASE_URL\n- For local runs: set proxy with real OPENAI_API_KEY; run code with OPENAI_BASE_URL= http://127.0.0.1 :<port>/v1 and dummy key
2025-09-12 22:18:08 +10:00
James Peter
7a28af8130
fix(ci/triage): remove OpenAI proxy and Rust/Code caches; call API directly in safety screen\n\n- Drop local proxy step and OPENAI_BASE_URL override\n- Strip unused Rust toolchain/cache and Code binary cache\n- Keep Node setup only; triage doesn’t build or run Rust\n- Keeps safety screen using Responses API (gpt-5→gpt-4o-mini fallback)
2025-09-12 21:54:27 +10:00
James Peter
df332867f8
Merge origin/main: keep local release script resilience; adopt upstream changes [skip ci]
2025-09-12 21:51:23 +10:00
James Peter
fd38d777bc
fix(release/homebrew): use local artifacts for sha256; add retry/backoff on remote fetch; avoid failing on CDN propagation
...
- Prefer release-assets/* when present to compute sha256 for Homebrew formula.\n- Retry curl with backoff if local file missing; print warnings instead of failing.\n- Keep formula URLs pointing to GitHub Releases; add best-effort HEAD checks.\n- Eliminates 404-induced failures right after uploading assets.
2025-09-12 21:51:05 +10:00
GitHub Action
3d99c3f8a6
chore(release): 0.2.133 [skip ci]
2025-09-12 11:28:27 +00:00
James Peter and GitHub
0942434ba2
Merge pull request #139 from just-every/upstream-merge
...
Upstream merge: openai/codex@main into main
2025-09-12 21:18:28 +10:00
just-every-code
991df9051a
merge(upstream): selective bucketed merge from openai/codex@main
...
- Adopt upstream in common/file-search and add TUI popup/types
- Preserve fork invariants in core (browser/agent/web_fetch gating, UA/version)
- Keep our TUI and CLI flows; resolve conflicts per policy
- Protocol: align mcp_protocol to upstream; add TS for Verbosity
- file-search: provide run_streaming shim to satisfy TUI usage
- common: make sandbox summary robust to new fields
Verification:
- scripts/upstream-merge/verify.sh: PASS (build-fast + API checks)
2025-09-12 10:45:24 +00:00
GitHub Action
a4aa78020f
chore(release): 0.2.132 [skip ci]
2025-09-12 10:34:49 +00:00
James Peter
5f10d0595b
ci: remove chat-completions fallback; use Responses API only via hardened proxies; fix release proxy mkdir; keep failure assertions
2025-09-12 20:25:39 +10:00
just-every-code
3855ca9e4c
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 10:13:42 +00:00
James Peter
3d4687f1b6
ci(proxy): harden OpenAI proxy and assert failures across workflows (issue-code, issue-comment, issue-triage, release); add chat-completions fallback provider; and disable upstream-merge concurrency cancellation
2025-09-12 20:11:55 +10:00
just-every-code
1b68cea1a1
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 09:02:16 +00:00
James Peter
11508f5917
Merge origin/main: adopt remote trivial changes; keep ours for upstream-merge workflow updates
2025-09-12 18:59:25 +10:00
James Peter
e27f320e65
ci(upstream-merge): add chat-completions fallback provider and force it via -c model_provider=openai_chat to avoid Responses API 500s via proxy; keep proxy for both paths
2025-09-12 18:59:12 +10:00
GitHub Action
33f97c80a6
Merge origin/main: adopt remote workflow changes; keep release changes [skip ci]
2025-09-12 08:58:57 +00:00
GitHub Action
c76b39d74e
chore(release): 0.2.131 [skip ci]
2025-09-12 08:57:38 +00:00
just-every-code
7177ef19dd
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 08:55:58 +00:00
James Peter
62695b1e5b
ci(upstream-merge): fail job if Code agent logs contain streaming/server errors or proxy logs show 5xx/upstream_error; ensures failures are visible when agent exits 0
2025-09-12 18:54:13 +10:00
just-every-code
bf6bf94089
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 08:50:03 +00:00
James Peter
fec6aa0f0c
fix(proxy+exec): harden proxy TLS forwarding (set servername, reset Host, normalize hop-by-hop headers) and make exec exit non-zero on agent Error events so workflows fail fast when streaming fails
2025-09-12 18:48:18 +10:00
just-every-code
6fc3e0e090
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 08:27:14 +00:00
James Peter
6ad9cb283c
fix(core/http): set explicit Host header from target URL to avoid proxy TLS SNI handshake failures when using HTTP(S)_PROXY with /v1/responses streaming
2025-09-12 18:25:06 +10:00
just-every-code
9ac636ed83
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 07:30:12 +00:00
James Peter
6e52a0fe84
Merge origin/main: adopt remote version bumps; keep ours elsewhere (proxy logging, server error diagnostics)
2025-09-12 17:27:49 +10:00
James Peter
43e6afe2d7
ci(upstream-merge): add verbose OpenAI proxy with streaming-safe pass-through and rich JSON logs; upload/tail logs\n\n- Forward exact client signature (Accept: text/event-stream, OpenAI-Beta: responses=experimental, originator, UA)\n- Replace only Authorization with secret API key; preserve all other headers\n- Stream SSE verbatim; avoid buffering; generous timeouts\n- Structured JSON logs: request/response headers (redacted), status, x-request-id, byte counts\n- Print log tail in job summary and upload full log as artifact\n\nContext: diagnose intermittent 5xx seen in Code runs; ensure proxy cannot mask origin responses and provides correlation data.
2025-09-12 17:26:56 +10:00
just-every-code
924f9e1818
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 06:46:59 +00:00
Michael Bolin and GitHub
c172e8e997
feat: added SetDefaultModel to JSON-RPC server ( #3512 )
...
This adds `SetDefaultModel`, which takes `model` and `reasoning_effort`
as optional fields. If set, the field will overwrite what is in the
user's `config.toml`.
This reuses logic that was added to support the `/model` command in the
TUI: https://github.com/openai/codex/pull/2799 .
2025-09-11 23:44:17 -07:00
GitHub Action
35bc0cd432
chore(release): 0.2.130 [skip ci]
2025-09-12 06:33:45 +00:00
James Peter
0e71ef8394
Merge origin/main: integrate upstream changes; keep local TUI Left/Right navigation enhancements
2025-09-12 16:21:57 +10:00
James Peter
2994466b77
feat(tui/theme): allow Left/Right to mirror Up/Down; enable Save/Retry navigation via arrows
...
- Treat Left/Right as Up/Down across Theme/Spinner views, including Create flows.\n- In review screens, Left focuses preview toggle (themes) and cycles buttons; Right moves to Save/Retry.\n- Update title hint to display '▲ ▼ ◀ ▶ select'.\n\nAlso includes other local TUI/core adjustments present in the working tree. Verified with ./build-fast.sh.
2025-09-12 16:21:45 +10:00
Michael Bolin and GitHub
9bbeb75361
feat: include reasoning_effort in NewConversationResponse ( #3506 )
...
`ClientRequest::NewConversation` picks up the reasoning level from the user's defaults in `config.toml`, so it should be reported in `NewConversationResponse`.
2025-09-11 21:04:40 -07:00
just-every-code
5289714bfa
Merge remote-tracking branch 'origin/main' into upstream-merge
2025-09-12 01:01:23 +00:00
GitHub Action
e9a605cb1b
chore(release): 0.2.129 [skip ci]
2025-09-12 00:57:31 +00:00
James Peter