Repository navigation
Conversation
Contributor
Author
|
CI follow-up (2026-07-23): reproduced the red gates against the clean base commit The image fallback patch remains focused and its route suite passes 21/21. I am not copying unrelated base-red repairs into this PR. #8306 overlaps one route hunk and should land first; after the release baseline repairs land, #8307 should be updated from the new base and CI rerun. |
…der maps cheaper-inference#6533 guard: every registered catalog engine must have an aggressivenessOf() rank above "off" and a non-generic expectedReductionFactor(), or floor-mode escalation ranking silently treats it as inert. The `codex-responses` engine (Responses Tool Output, stackPriority 12, added with the 12-engine catalog) was never added to AGGRESSIVENESS / REDUCTION_FACTOR in adaptiveCompression/ladder.ts, so aggressivenessOf() returned 0 (== off) and expectedReductionFactor() fell back to the generic 0.9. Placed it between rtk (20) and ionizer (25) with a conservative, lossless-first reduction factor of 0.85 — it stays a single-mode engine, not part of the automatic DEFAULT_LADDER (like ionizer / relevance / omniglyph). Validated: ladder-engine-maps-6533 now 3/3; the compression/adaptive ladder test files stay green (24/0).
…cheaper-inference#8123 model realignment Base-red slice 3. Two test files referenced antigravity model IDs that cheaper-inference#8013 (align official clients / callable catalog) and cheaper-inference#8123 (discover live AGY models) intentionally renamed/retired — the candidate builder and static catalog are correct; the tests were stale. - auto-combo-credentialed-model-pool: claude-sonnet-5 -> claude-sonnet-4-6 and the gemini-3.5-flash-{low,medium,high} tier -> gemini-3.6-flash-{low,medium,high} (verified against the live createVirtualAutoCombo candidate set); the exclusion wildcard and per-account transparency assertions are unchanged in intent. - T31 static-catalog: gemini-3-pro-preview was retired by cheaper-inference#8013, so assert the current client-visible top flash tier (gemini-3.6-flash-high) plus its absence. Test-only. Validated: auto-combo-credentialed-model-pool 4/4, the T31/T33/T34/T38 model-specs file 8/8.
… agy alias Same cheaper-inference#8013 realignment as this PR: `getModelInfo` resolves the stored `agy/…` alias target to its canonical provider id `antigravity` (ALIAS_TO_PROVIDER_ID). The stored alias STRING stays `agy/gemini-pro-agent`; only the resolved `provider` is now `antigravity`, not `agy`. Test updated to match. 6/0.
…hrowing ReferenceError Base-red slice 4 (single root cause across the whole handleChat cluster). src/sse/handlers/chat.ts:1273 references PROVIDER_BREAKER_FAILURE_STATUSES to decide whether an all-rate-limited provider result should trip the provider breaker, but the constant was only a FILE-LOCAL const in chatPredicates.ts (a refactor extracted it out of chat.ts and never re-exported it). Every request that reached that branch threw `ReferenceError: PROVIDER_BREAKER_FAILURE_STATUSES is not defined`, so the global fallback and breaker-gate paths blew up — surfacing as "All models failed | PROVIDER_BREAKER_FAILURE_STATUSES is not defined" and breaking the handleChat coverage tests (combo-error passthrough, 503 for cooled-down/open-breaker, budget-error, model cooldown, body-derived retry-after, non-JSON rate-limit bodies). Fix: export the const from chatPredicates.ts and import it in chat.ts (one canonical definition, restoring the pre-refactor behavior). Validated: chat-route-coverage 15/0 (was 12/3), chat-cooldown-aware-retry 6/0, chat-rate-limit-body-lock 2/0; breaker guards (7907, combo-breaker-429, openrouter-6842) unchanged; typecheck:core clean.
…tes.ts Same root cause as the chat.ts import fix in this PR: the const was extracted out of chat.ts into chatPredicates.ts, so the nvidia-quota Phase-1 guard (which greps the source for the `= new Set([...])` declaration to prove 429 was not added to the breaker classification) must read chatPredicates.ts, not chat.ts. Now 13/0.
…e was added Base-red slice 5. cheaper-inference#8034/cheaper-inference#8064 (per-model/endpoint compression exclusion filter) appended the new `compression-exclusions` sidebar item to the END of the compression group, after `compression-studio`. That broke the deliberate "Compression Studio is the last item of the compression group" invariant asserted by two tests (sidebar-engine-items "Studio must be last" and sidebar-visibility's exact-order array, whose expected list places compression-studio immediately before cli-code). Rather than weaken the invariant, moved `compression-exclusions` ABOVE `compression-studio` in sections.ts (new pages go above the Studio showcase) and added the new item to sidebar-visibility's expected order. Both tests, plus the other four sidebar suites (reorg / back-compat / tools / engine-items), stay green.
… URL Base-red slice 6, rebased onto the advanced release/v3.8.49 (91fd5f9). The oauth grok-cli cheaper-inference#7610 guard was already fixed on the base by cheaper-inference#8027 (it reads the warning from grokCliAuthJson.ts) — dropped from this slice to avoid a conflicting duplicate. Remaining two, still red on the current base: - i18n cheaper-inference#7258: the "focused repro" asserted zh-TW.json STILL carries raw __MISSING__: placeholders. That backlog was filled (the "no locale has a raw __MISSING__: leaf" invariant is the durable guard); retired the now-inverted repro. - qianfan: Baidu renamed the product page (product/wenxinworkshop -> product-s/ qianfan_home); updated the expected website URL. Validated (clean env): i18n 4/0, qianfan 5/0; oauth-modal-grok 2/0 already green on base.
…s + eslint baseline The release branch accumulated deterministic unit-test failures (fast-path red on every open PR). These are the ones with a clear, surgical root cause: 1. cheaper-inference#6863 combo model-lockout — the cheaper-inference#7940/cheaper-inference#7980 "cap exactCooldownMs against maxCooldownMs" clamp was also clamping an AUTHORITATIVE parsed upstream quota reset (e.g. "Resets in 92h27m28s") down to maxCooldownMs, so an exhausted model was retried far too early. recordModelLockoutFailure now takes exactCooldownIsUpstreamReset — set by the combo callers when the exact cooldown is a real upstream reset — which exempts it from the cap. The cheaper-inference#7980 computed until-midnight cap is unchanged (flag absent → still capped). 2. cheaper-inference#5786 streaming claude←codex — stripInternalReasoningPlaceholder (cheaper-inference#8081/cheaper-inference#8162) unconditionally .trim()'d every value. On the per-delta streaming path this ate the meaningful edge spaces of each delta ("Hello, " + "world." + " Bye." glued to "Hello,world.Bye."). It now only collapses to "" when whitespace is all that remains after removing the placeholder, preserving real content verbatim. 3. SPAWN_CAPABLE_PREFIXES test — cheaper-inference#7892 added /api/vnc-session (11th spawn-capable prefix, spawns Docker) but the client-safe guard test still expected 10 and did not list it. Aligned to 11 + added the entry to the checklist. 4. ESLint baseline — cheaper-inference#8008/cheaper-inference#8062 merged new test files with no-explicit-any without refreshing the frozen suppressions, so "No new ESLint warnings" went red for the whole branch. Regenerated the two affected entries (combo-routing-engine.test.ts 269→271, oauth-refresh-connection-dedup-8059.test.ts +1). Validated: the three failing tests now pass; the sibling guards they interact with stay green (cheaper-inference#7980 exact-cooldown-cap 4/4, cheaper-inference#8162 placeholder suites 17+12+41, account-fallback 77); typecheck:core clean; lint:json --max-warnings 0 exits 0. NOTE: the release branch has ~20 further real base-red failures (compression-engine catalog, handleChat fallback, provider candidate transparency, i18n, misc). Those are tracked separately, one focused PR per root-cause cluster; this PR is the first slice.
fenix007
force-pushed
the
fix/codex-image-account-fallback
branch
from
July 23, 2026 17:30
dfd35ea to
41582b9
Compare
6 tasks done
diegosouzapw
added a commit
to herjarsa/OmniRoute
that referenced
this pull request
Jul 27, 2026
…ests into dedicated file The Codex account-fallback tests from PR cheaper-inference#8477 (rebase of cheaper-inference#8307) push the shared image-generation-route test file above the 800-line testCap from config/quality/file-size-baseline.json. Extract the 5 new tests (skips-excluded-model, retries-one-sibling, no-retry-ordinary-400, stops-after-one-fallback, attributes-call-logs) plus their shared fixtures into tests/unit/image-generation-route-codex-account-fallback.test.ts to keep both files well under the cap. No production code touched; all 21 tests still pass. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
diegosouzapw
changed the title
diegosouzapw
added
the
deferred-v3.8.50
added 6 commits
July 28, 2026 14:47
Collaborator
|
Re-homed to |
diegosouzapw
added
merge-train-deferred
Collaborator
|
Thanks for tracking down this production incident and writing the original repro/test — the account-fallback diagnosis and the exact upstream error signature are all yours. Since this PR was opened, a general credential-fallback mechanism landed for image generation ( Closing this one in favor of #10838, which carries your co-authorship on the commit. Appreciate the original catch! |
fenix007
pushed a commit
to fenix007/OmniRoute
that referenced
this pull request
Aug 20, 2026
check:complexity-ratchets has been red on stable since fork.3, which rebaselined check:file-size for its ported fixes but left the two complexity ratchets behind. Measured per tag with one ESLint walk each, so the drift is attributed rather than assumed: - cyclomatic + max-lines: 2056 baseline, but pristine v3.8.48 (fork.1, which carries only docker/CI commits) already measures 2058 — inherited from upstream, which does not run this gate on the fast-path PR->release. fork.2 adds +1 (peekCodexSseTransientError, cheaper-inference#7570/cheaper-inference#8043) and fork.3 adds +1 (recordProviderFailure, cheaper-inference#10116). - cognitive: 890 baseline matches fork.1 exactly. fork.2 adds +3 (two functions in codex.ts plus imageCredentialRetry.ts, cheaper-inference#9231/cheaper-inference#8307) and fork.3 adds +2 (accountFallback.ts cheaper-inference#10116, providerLimits.ts cheaper-inference#10534). fork.4 measures the same 2060 / 895: the error-status patch set added zero violations, verified by diffing the per-violation lists between the two trees. No fork-owned code sits over either threshold — every flagged function comes from an upstream commit upstream rebaselined on its own branch, so the ratchets are moved to the measured values instead of refactoring ported code away from its upstream shape. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
400once on an eligible sibling account, with a short model-scoped lockout and per-attempt proxy resolution.Related Issues
Validation
npm run typecheck:corenpm run check:complexity-ratchetsnpm run check:file-sizenpm run check:docs-allactionlint .github/workflows/quality.ymlgit diff --checkTests Added
tests/unit/image-generation-codex-account-fallback.test.ts400responsesCurrent Release Integration
The branch merges the current
release/v3.8.49tip (3b515d90b), including merged #8306. The application diff remains limited to six files: the image route, image handler, extracted call-log result helper, focused fallback tests, changelog fragment, and one-line Stryker test registration. Historical release-base CI repair commits no longer appear in the PR diff.Four additional CI-maintenance files make the new advisory production build reliable for fork PRs:
quality.yml, its regression test, and the two release-quality docs. The final GitHub-hosted webpack build passed in 11 minutes after two unrelated Turbopack runs were terminated by hosted-runner shutdown signals.Reviewer Notes
400, and the exact upstream ChatGPT-account unsupported-model message for the requested model.response_formatbehavior is unchanged.omni-releaserunner.