chore(release): v3.5.2 — Qoder DashScope Native Integration & Stability by diegosouzapw · Pull Request #999 · diegosouzapw/OmniRoute

diegosouzapw · 2026-04-05T05:03:32Z

🚀 Release v3.5.2

Changes

Qoder API Native Integration: Completely refactored the Qoder Executor to bypass the legacy COSY AES/RSA encryption algorithm, routing directly into the native DashScope OpenAi-compatible URL. Eliminates complex dependencies on Node crypto modules while improving stream fidelity.
Resilience Engine Overhaul: Integrated context overflow graceful fallbacks, proactive OAuth token detection, and empty-content emission prevention (feat(resilience): context overflow fallback, OAuth token detection, empty content guard & context-optimized combo strategy #990).
Context-Optimized Routing Strategy: Added new intelligent routing capability to natively maximize context windows in automated combo deployments (feat(resilience): context overflow fallback, OAuth token detection, empty content guard & context-optimized combo strategy #990).
Responses API Stream Corruption: Fixed deep-cloning corruption where Anthropic/OpenAI translation boundaries stripped response. specific SSE prefixes from streaming boundaries (fix(sse): preserve Responses API events in stream translation #992).
Claude Cache Passthrough Alignment: Aligned CC-Compatible cache markers consistently with upstream Client Pass-Through mode preserving prompt caching.
Turbopack Memory Leak: Pinned Next.js to strict 16.0.10 preventing memory leaks and build staleness from recent upstream Turbopack hashed module regressions (Fix: Pin Next.js to 16.0.10 to prevent Turbopack hashed module bug #987).

Tests

Build, Linter, Typecheck passing
33 Languages localized

⚠️ After merging: run Phase 2 steps to tag, publish, and deploy.

…I child instances, plus workflow bumps

…mpty content guard & context-optimized combo strategy - Add isContextOverflowError + isContextOverflow detectors (400 + token-limit signals) - Auto-fallback to next family model on context overflow in chatCore - Add isEmptyContentResponse to catch fake-success empty responses, trigger fallback + recursive retry - Add OAUTH_INVALID_TOKEN error type (T11) with isOAuthInvalidToken signal matching; warn instead of deactivating node - Add getModelContextLimit helper in modelsDevSync (reads limit_context from synced capabilities) - Upgrade getTokenLimit in contextManager to check models.dev DB before registry (fixes gemini-2.5-pro: 1000000→1048576) - Add findLargerContextModel in modelFamilyFallback for context-aware model selection - Add sortModelsByContextSize + context-optimized combo strategy in combo.ts - Update context-manager unit test for corrected gemini-2.5-pro limit Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ecursion, dedup signals, findLargerContextModel - Fix isEmptyContentResponse: check message.tool_calls/delta.tool_calls instead of firstChoice.tool_calls (wrong OpenAI API path, caused tool-call responses to be falsely flagged as empty) - Fix empty content fallback: replace recursive handleChatCore call (infinite recursion risk + wrong model due to original body.model) with non-recursive pattern — call executeProviderRequest, parse fallback response body, reassign responseBody and fall through to existing processing - Fix context overflow: use findLargerContextModel over family candidates first, fall back to getNextFamilyFallback — ensures we pick a model with actually larger context window on overflow - Fix signal dedup: export CONTEXT_OVERFLOW_SIGNALS + CONTEXT_OVERFLOW_REGEX from errorClassifier.ts; import shared regex in modelFamilyFallback.ts, removing duplicate signal list and per-call RegExp construction Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

When translating Claude-format responses (e.g. GLM) to Responses API format for Codex CLI, the sanitizer stripped {event, data} structured items to {"object":"chat.completion.chunk"}, losing all content and the critical response.completed event. Only run sanitizeStreamingChunk on OpenAI Chat Completions chunks, skipping items that have the Responses API {event, data} structure.

Verifies that {event,data} structured items from the Responses API translator bypass sanitizeStreamingChunk when translating Claude-format providers (e.g. GLM) to Responses API format for Codex CLI.

…fix check Use explicit `response.` prefix check instead of generic `event && data` presence check, as recommended in PR review.

Remove ^ prefix from next and eslint-config-next to prevent automatic upgrades to 16.1.x+ which introduced content-based hashing for external module references in Turbopack. Also remove duplicate Material Symbols @import from globals.css (font already loaded via <link> in layout.tsx). Fixes #509

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: df24a48e26

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-04-05T05:09:12Z

+    const credentialsApiBase =
+      (credentials as any).customApiBase || (credentials as any).resourceUrl;


Read Qoder custom base URL from providerSpecificData

This endpoint override lookup never checks credentials.providerSpecificData, so resourceUrl/customApiBase values stored with provider metadata are ignored and requests always fall back to the hardcoded DashScope URL. In environments where Qoder OAuth returns a tenant-specific resourceUrl, this causes every request to hit the wrong host and fail despite valid credentials.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-04-05T05:09:12Z

+            const fallbackRaw = await fallbackResult.response.text();
+            try {
+              responseBody = fallbackRaw ? JSON.parse(fallbackRaw) : {};


Parse fallback body with SSE-aware logic

The empty-content fallback path assumes the fallback response is JSON and immediately calls JSON.parse, but earlier in this same handler we explicitly support providers that return SSE even when stream=false. For non-stream requests that still use upstream streaming (for example Claude Code compatible providers), a successful fallback response will be valid SSE and this parse will throw, incorrectly returning 502 instead of using the fallback result.

Useful? React with 👍 / 👎.

gemini-code-assist

Code Review

This pull request updates OmniRoute to version 3.5.2, refactoring the Qoder executor for native DashScope integration and introducing a resilience engine overhaul with context overflow fallbacks and empty-content prevention. It also adds a context-optimized routing strategy, fixes Responses API stream corruption, and pins Next.js to version 16.0.10. Additionally, it updates deployment workflows and build scripts to improve package hygiene. Feedback identifies a bug in the empty-content fallback logic where the providerResponse variable is not updated, which could break streaming by failing to pass the new response stream to the client.

gemini-code-assist · 2026-04-05T05:11:45Z

+          const fallbackResult = await executeProviderRequest(nextModel, false);
+          if (fallbackResult.response.ok) {
+            const fallbackRaw = await fallbackResult.response.text();
+            try {
+              responseBody = fallbackRaw ? JSON.parse(fallbackRaw) : {};
+              providerUrl = fallbackResult.url;
+              providerHeaders = fallbackResult.headers;
+              finalBody = fallbackResult.transformedBody;
+              reqLogger.logTargetRequest(providerUrl, providerHeaders, finalBody);
+              log?.info?.(
+                "EMPTY_CONTENT_FALLBACK",
+                `Serving ${nextModel} as fallback for ${model}`
+              );
+              // Fall through — continue processing with the new responseBody


The fallback logic for empty content responses (lines 1872-1885) is inconsistent with the context overflow fallback logic (lines 1661-1672) and contains a bug.

Specifically, it fails to update the providerResponse variable. While it updates responseBody, providerUrl, and providerHeaders, the rest of the handler (especially for streaming requests) will continue to use the original providerResponse. This means the client will receive the original empty response instead of the fallback result.

Additionally, calling .text() on the fallback response (line 1874) buffers the entire content. If the request is streaming, this defeats the purpose of streaming and introduces latency. You should update providerResponse and avoid consuming the body as text if streaming is enabled, allowing the subsequent logic to handle the new response stream correctly.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 837e64de1d

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-04-05T05:53:01Z

+  if (typeof credentials.refreshToken === "string" && credentials.refreshToken.trim()) {
+    return credentials.refreshToken.trim();


Stop using refresh token as request bearer token

getAuthToken falls back to credentials.refreshToken, and execute always forwards that value in Authorization. In OAuth flows a refresh token is not a resource-server credential, so when accessToken is missing this path both guarantees upstream auth failures and exposes a long-lived refresh secret to the provider endpoint/logging chain. This branch should reject and trigger refresh logic instead of sending the refresh token upstream.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-04-05T05:53:01Z

+  const currentProvider = currentParsed.provider || currentParsed.providerAlias || "unknown";
+  const currentModelId = currentParsed.model || currentModel;
+  const currentLimit = getModelContextLimit(currentProvider, currentModelId) ?? 0;


Use real provider id for context-limit fallback lookup

findLargerContextModel derives provider via parseModel, but in this flow models are plain IDs (e.g. "gpt-5", "claude-sonnet-4-6") so parseModel returns no provider and this code falls back to "unknown". That makes getModelContextLimit miss synced capabilities, so the function usually returns null and context-overflow fallback never picks a truly larger-context sibling model.

Useful? React with 👍 / 👎.

…ty (diegosouzapw#999) * feat(qoder): native cosy integration * feat(qoder): implement native COSY encryption algorithm and remove CLI child instances, plus workflow bumps * feat(resilience): context overflow fallback, OAuth token detection, empty content guard & context-optimized combo strategy - Add isContextOverflowError + isContextOverflow detectors (400 + token-limit signals) - Auto-fallback to next family model on context overflow in chatCore - Add isEmptyContentResponse to catch fake-success empty responses, trigger fallback + recursive retry - Add OAUTH_INVALID_TOKEN error type (T11) with isOAuthInvalidToken signal matching; warn instead of deactivating node - Add getModelContextLimit helper in modelsDevSync (reads limit_context from synced capabilities) - Upgrade getTokenLimit in contextManager to check models.dev DB before registry (fixes gemini-2.5-pro: 1000000→1048576) - Add findLargerContextModel in modelFamilyFallback for context-aware model selection - Add sortModelsByContextSize + context-optimized combo strategy in combo.ts - Update context-manager unit test for corrected gemini-2.5-pro limit Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(review): address Gemini code review — tool_calls path, infinite recursion, dedup signals, findLargerContextModel - Fix isEmptyContentResponse: check message.tool_calls/delta.tool_calls instead of firstChoice.tool_calls (wrong OpenAI API path, caused tool-call responses to be falsely flagged as empty) - Fix empty content fallback: replace recursive handleChatCore call (infinite recursion risk + wrong model due to original body.model) with non-recursive pattern — call executeProviderRequest, parse fallback response body, reassign responseBody and fall through to existing processing - Fix context overflow: use findLargerContextModel over family candidates first, fall back to getNextFamilyFallback — ensures we pick a model with actually larger context window on overflow - Fix signal dedup: export CONTEXT_OVERFLOW_SIGNALS + CONTEXT_OVERFLOW_REGEX from errorClassifier.ts; import shared regex in modelFamilyFallback.ts, removing duplicate signal list and per-call RegExp construction Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(UI): add context-optimized strategy to frontend schema and options * fix(sse): preserve Responses API events in stream translation When translating Claude-format responses (e.g. GLM) to Responses API format for Codex CLI, the sanitizer stripped {event, data} structured items to {"object":"chat.completion.chunk"}, losing all content and the critical response.completed event. Only run sanitizeStreamingChunk on OpenAI Chat Completions chunks, skipping items that have the Responses API {event, data} structure. * test(sse): add regression test for Claude→Responses stream sanitization Verifies that {event,data} structured items from the Responses API translator bypass sanitizeStreamingChunk when translating Claude-format providers (e.g. GLM) to Responses API format for Codex CLI. * fix(sse): strengthen Responses API event detection with response. prefix check Use explicit `response.` prefix check instead of generic `event && data` presence check, as recommended in PR review. * fix: pin Next.js to 16.0.10 to prevent Turbopack hashed module bug Remove ^ prefix from next and eslint-config-next to prevent automatic upgrades to 16.1.x+ which introduced content-based hashing for external module references in Turbopack. Also remove duplicate Material Symbols @import from globals.css (font already loaded via <link> in layout.tsx). Fixes diegosouzapw#509 * align cc-compatible cache handling with client passthrough * chore: integrate resilience and turbopack fixes (PRs diegosouzapw#992, diegosouzapw#990, diegosouzapw#987) * chore(release): bump to v3.5.2 — changelog, docs, version sync * docs(i18n): sync documentation updates to 33 languages * fix(qoder): replace any with unknown to comply with strict any-budget --------- Co-authored-by: diegosouzapw <diegosouzapw@users.noreply.github.com> Co-authored-by: oyi77 <oyi77@users.noreply.github.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> Co-authored-by: Chris Staley <christopher-s@users.noreply.github.com> Co-authored-by: Ivan <shanin-i2011@yandex.ru> Co-authored-by: R.D. <rogerproself@gmail.com>

diegosouzapw and others added 13 commits April 4, 2026 17:42

feat(qoder): native cosy integration

540c718

feat(qoder): implement native COSY encryption algorithm and remove CL…

21e3859

…I child instances, plus workflow bumps

fix(UI): add context-optimized strategy to frontend schema and options

076ae3a

test(sse): add regression test for Claude→Responses stream sanitization

d8528a3

Verifies that {event,data} structured items from the Responses API translator bypass sanitizeStreamingChunk when translating Claude-format providers (e.g. GLM) to Responses API format for Codex CLI.

fix(sse): strengthen Responses API event detection with response. pre…

767e6e6

…fix check Use explicit `response.` prefix check instead of generic `event && data` presence check, as recommended in PR review.

align cc-compatible cache handling with client passthrough

443215b

chore: integrate resilience and turbopack fixes (PRs #992, #990, #987)

abc79b8

chore(release): bump to v3.5.2 — changelog, docs, version sync

127e891

docs(i18n): sync documentation updates to 33 languages

df24a48

chatgpt-codex-connector Bot reviewed Apr 5, 2026

View reviewed changes

gemini-code-assist Bot reviewed Apr 5, 2026

View reviewed changes

fix(qoder): replace any with unknown to comply with strict any-budget

837e64d

chatgpt-codex-connector Bot reviewed Apr 5, 2026

View reviewed changes

diegosouzapw merged commit b100325 into main Apr 5, 2026
1 check passed

diegosouzapw deleted the release/v3.5.2 branch April 13, 2026 22:29

Navigation Menu

Uh oh!

chore(release): v3.5.2 — Qoder DashScope Native Integration & Stability - #999

chore(release): v3.5.2 — Qoder DashScope Native Integration & Stability#999
diegosouzapw merged 14 commits into
mainfrom
release/v3.5.2

diegosouzapw commented Apr 5, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Uh oh!

gemini-code-assist Bot left a comment

Uh oh!

gemini-code-assist Bot Apr 5, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

		const credentialsApiBase =
		(credentials as any).customApiBase \|\| (credentials as any).resourceUrl;

		if (typeof credentials.refreshToken === "string" && credentials.refreshToken.trim()) {
		return credentials.refreshToken.trim();

Uh oh!

Conversation

diegosouzapw commented Apr 5, 2026

🚀 Release v3.5.2

Changes

Tests

⚠️ After merging: run Phase 2 steps to tag, publish, and deploy.

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist Bot Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants