pjandro/OmniRoute
1
0
Fork 0
mirror of https://github.com/diegosouzapw/OmniRoute.git synced 2026-08-24 08:02:14 +03:00
Code Issues Packages Projects Releases Wiki Activity

fix(tests): drain base-red cluster from 2026-08-23 merges (#9985)

Browse Source 
Drain the unit-shard reds the 2026-08-23 merge wave left on
release/v3.8.50, each discriminated as stale-test (contract moved
intentionally, test aligned) vs real bug (fixed in code/messages):

1. check-deps 6A.8 allowlist — #11224 added @testing-library/dom and
   @testing-library/user-event to package.json without the gate
   allowlist. Both are legitimate: @testing-library/dom is a required
   peer of @testing-library/react v16, and user-event is the official
   companion for UI tests. Fix: allowlist entries with a justification
   note referencing #11224/#9985.

2. search-route 400-vs-fallback — #11097 intentionally changed the
   zero-credential /v1/search contract: instead of returning 400 it
   promotes the fallback-only duckduckgo-free provider so out-of-the-box
   search works. The test pinned the OLD 400 contract. Fix (contract
   alignment): the test now pins the new fallback contract — 200,
   provider duckduckgo-free, DuckDuckGo lite endpoint called, results
   parsed from lite HTML.

3. codex catalog token limits (4 named reds + 2 sibling sweeps) —
   #11179 raised GPT_5_6_CODEX_CAPABILITIES from the 272K pricing tier
   to the real usable 872K window (live evidence: 390K served past
   272K with HTTP 200) and updated codex-gpt56-catalog.test.ts but
   missed the sibling pins. Stale tests aligned: models-catalog-combo-
   metadata (max_input_tokens now clamps to min(872000, 500000 override)
   = 500000), vscode-token-routes x3 (872000), and two more found in the
   sibling sweep: vscode-token-routes-gpt56 and provider-models-route-
   codex (conservative merge semantics unchanged: pinned 872000 < live
   999999 still wins).

4. CLI catalog counts (3 reds) — #11166 added prime-agent (agent
   category) without the cardinality pins: EXPECTED_AGENT_COUNT 8->9,
   total 34->35, D15 agent list + prime-agent, cli-tools-schema id list.
   Also added the missing English/Vietnamese cliTools descriptions for
   prime-agent (cli-catalog-display-contract) and corrected the stale
   CLI-TOOLS.md agent count (8->9).

5. setup-qwen container guard — since #10057 the container guard exits 2
   on ephemeral runtimes; the two tests exercising the merge/write path
   were environment-sensitive (red on container devboxes). Fix: pass
   allowContainerWrite so the tests are hermetic everywhere; the guard
   keeps its own dedicated coverage.

6. i18n health verdict namespace (real bug, fixed in messages) — #11224
   added the verdict/diagnostics strings to the `sidebar` namespace but
   health/page.tsx reads them via useTranslations("health"), so the page
   rendered raw keys in every locale and the "direct translation calls
   have English messages" gate went red. Fix: keys moved sidebar->health
   in en.json + vi.json (the only locales that had them), plus
   health.healthSubtitle added. The sidebar never referenced them
   (verified: no usage), and sidebar.healthSubtitle (its real sidebar
   key) is untouched.

7. i18n pt-BR drift (22 keys) — the 08-23 wave (#11224/#11228/#11215/
   #11204/#11195) added English keys never translated: common.batch*,
   endpoint.*, cliCommon.concept.acp.warning, resilienceConnections.*,
   plus the moved health.* keys and the prime-agent cliTools
   description. Fix: pt-BR translations added — 0 missing keys vs en;
   vi strict parity re-verified (0 missing, 0 extra).

Validation: every touched test file RED->GREEN individually
(node --import tsx/esm --test), typecheck:core clean, docs-counts-sync
soft-pass, cli-i18n gate PASS, 8/8 unit shards re-run on the branch.

Refs #9985
This commit is contained in:
Xiangzhe
2026-08-23 14:45:05 -03:00
parent 855243ab18
commit c2df757610
14 changed files with 106 additions and 49 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
config/quality/dependency-allowlist.json
View File

@@ -1,5 +1,9 @@
{
"_comment": "Allowlist anti-slopsquatting (check-deps.mjs). Toda dep nova exige adicao EXPLICITA aqui apos verificar que e legitima.",
"_justifications": {
"@testing-library/dom": "Peer dep obrigatoria de @testing-library/react v16 (adicionada no PR #11224); Refs #9985.",
"@testing-library/user-event": "Utilitario oficial do ecossistema testing-library para testes de UI (adicionada no PR #11224); Refs #9985."
},
"allowed": [
"@atjsh/llmlingua-2",
"@aws-sdk/client-bedrock-runtime",
@@ -20,8 +24,10 @@
"@stryker-mutator/tap-runner",
"@swc/helpers",
"@tailwindcss/postcss",
"@testing-library/dom",
"@testing-library/jest-dom",
"@testing-library/react",
"@testing-library/user-event",
"@toon-format/toon",
"@types/better-sqlite3",
"@types/bun",

6
docs/reference/CLI-TOOLS.md
View File

@@ -1,19 +1,19 @@
---
title: "CLI Tools — OmniRoute"
version: 3.8.50
lastUpdated: 2026-08-18
lastUpdated: 2026-08-23
---
# CLI Tools — OmniRoute
Last updated: 2026-08-18
Last updated: 2026-08-23
OmniRoute integrates with three categories of CLI tools spread across three dedicated dashboard pages:
| Page | Route | Concept | Count |
| -------------- | ----------------------- | ------------------------------------------------------------------------- | ------------ |
| **CLI Code's** | `/dashboard/cli-code` | Coding tools you point at OmniRoute (Client → CLI → OmniRoute → Provider) | 26 |
| **CLI Agents** | `/dashboard/cli-agents` | Autonomous agents you point at OmniRoute (same flow, broader scope) | 8 |
| **CLI Agents** | `/dashboard/cli-agents` | Autonomous agents you point at OmniRoute (same flow, broader scope) | 9 |
| **ACP Agents** | `/dashboard/acp-agents` | CLIs that OmniRoute spawns as backend via stdio/ACP (reverse flow) | see registry |
Legacy routes redirect via 308: `/dashboard/cli-tools` → `/dashboard/cli-code`, `/dashboard/agents` → `/dashboard/acp-agents`.

14
src/i18n/messages/en.json
View File

@@ -1221,12 +1221,6 @@
"consoleLogsSubtitle": "Console output",
"logsActivitySubtitle": "User activity log",
"healthSubtitle": "System health check",
"healthVerdictReady": "OmniRoute is ready",
"healthVerdictActionRequired": "Action required to restore full operation",
"healthVerdictCoolingDown": "Cooling down after recent changes",
"advancedDiagnosticsTitle": "Advanced diagnostics",
"hide": "Hide",
"show": "Show",
"costsPricingSubtitle": "Per-model pricing rules",
"costsBudgetSubtitle": "Budget limits",
"costsQuotaShareSubtitle": "Share provider quotas across keys",
@@ -2918,6 +2912,7 @@
"interpreter": "Open Interpreter autonomous coding agent CLI",
"omp": "Oh My Pi terminal coding agent",
"letta": "Letta CLI agent with persistent memory and tool use",
"prime-agent": "Prime Agent — self-improving RLM coding harness with OpenAI-compatible provider support",
"warp": "Warp AI terminal with custom provider support",
"agent-deck": "Agent Deck multi-agent orchestrator"
},
@@ -4627,6 +4622,13 @@
"retry": "Retry",
"allOperational": "All systems operational",
"issuesDetected": "System issues detected",
"healthVerdictReady": "OmniRoute is ready",
"healthVerdictActionRequired": "Action required to restore full operation",
"healthVerdictCoolingDown": "Cooling down after recent changes",
"healthSubtitle": "System health check",
"advancedDiagnosticsTitle": "Advanced diagnostics",
"hide": "Hide",
"show": "Show",
"updatedAt": "Updated {time}",
"latency": "Latency",
"latencyP50": "p50",

26
src/i18n/messages/pt-BR.json
View File

@@ -970,6 +970,13 @@
"batchTimelineCancelled": "Cancelado",
"batchTokenUsage": "Uso de Token",
"batchMetadata": "Metadados",
"batchHeaderSubtitle": "Execute muitas requisições como um único job",
"batchStep1": "1 · Enviar JSONL",
"batchStep1Desc": "Adicionar requisições",
"batchStep2": "2 · Criar lote",
"batchStep2Desc": "Executar job",
"batchStep3": "3 · Obter resultados",
"batchStep3Desc": "Baixar saída",
"batchFileContents": "Conteúdo do Arquivo",
"batchFileUsedByCount": "Usado por {count, plural, one {# lote} other {# lotes}}",
"batchFilePreview": "Prévia",
@@ -2905,6 +2912,7 @@
"interpreter": "CLI do agente de codificação autônomo Open Interpreter",
"omp": "Agente de codificação de terminal Oh My Pi",
"letta": "Agente CLI Letta com memória persistente e uso de ferramentas",
"prime-agent": "Prime Agent — harness de codificação RLM autoevolutivo com suporte a API compatível com OpenAI",
"warp": "Terminal de IA Warp com suporte a provedor personalizado",
"agent-deck": "Orquestrador multi-agente Agent Deck"
},
@@ -3831,6 +3839,9 @@
},
"endpoint": {
"title": "Endpoint da API",
"subtitle": "Use o endpoint compatível com OpenAI na maioria dos SDKs e ferramentas.",
"testEndpoint": "Testar endpoint →",
"advancedProtocols": "Protocolos avançados",
"available": "Endpoints Disponíveis",
"cloudProxy": "Proxy na Nuvem",
"disableConfirm": "Tem certeza que deseja desativar o proxy na nuvem?",
@@ -4611,6 +4622,13 @@
"retry": "Tentar Novamente",
"allOperational": "Todos os sistemas operacionais",
"issuesDetected": "Problemas detectados no sistema",
"healthVerdictReady": "O OmniRoute está pronto",
"healthVerdictActionRequired": "Ação necessária para restaurar a operação plena",
"healthVerdictCoolingDown": "Em resfriamento após mudanças recentes",
"healthSubtitle": "Verificação de saúde do sistema",
"advancedDiagnosticsTitle": "Diagnósticos avançados",
"hide": "Ocultar",
"show": "Mostrar",
"updatedAt": "Atualizado {time}",
"latency": "Latência",
"latencyP50": "p50",
@@ -12034,6 +12052,7 @@
"acp": {
"title": "ACP Agents",
"phrase": "CLIs que o OmniRoute spawna como backend de execução (fluxo reverso)",
"warning": "A maioria dos usuários pode ignorar isto — use apenas quando uma integração exigir.",
"flow": "Cliente → OmniRoute → spawn CLI (stdio/ACP) → resposta",
"seeOther": "Ver →"
}
@@ -13341,6 +13360,13 @@
},
"resilienceConnections": {
"title": "Resiliência de Conexão",
"reassuranceTitle": "Suas conexões se recuperam automaticamente",
"reassuranceDetail": "Normalmente nenhuma ação é necessária. O OmniRoute dá uma pausa temporária em uma conexão após falhas e depois a tenta novamente com segurança.",
"plainStates": {
"healthy": "Requisições podem ser enviadas",
"coolingDown": "Tentando novamente em breve",
"lockedOut": "Precisa da sua atenção"
},
"table": {
"status": "Status",
"provider": "Provedor",

16
src/i18n/messages/vi.json
View File

@@ -1300,13 +1300,7 @@
"open": "mở",
"close": "đóng"
},
"noResults": "Không có kết quả",
"healthVerdictReady": "OmniRoute đã sẵn sàng",
"healthVerdictActionRequired": "Cần hành động để khôi phục hoạt động đầy đủ",
"healthVerdictCoolingDown": "Đang nguội sau các thay đổi gần đây",
"advancedDiagnosticsTitle": "Chẩn đoán nâng cao",
"hide": "Ẩn",
"show": "Hiện"
"noResults": "Không có kết quả"
},
"webhooks": {
"title": "Webhook",
@@ -2918,6 +2912,7 @@
"interpreter": "Tác nhân lập trình tự trị Open Interpreter CLI",
"omp": "Tác nhân lập trình Oh My Pi trên terminal",
"letta": "Tác nhân Letta CLI có bộ nhớ lâu dài và khả năng dùng công cụ",
"prime-agent": "Prime Agent — bộ khung lập trình RLM tự cải tiến hỗ trợ API tương thích OpenAI",
"warp": "Terminal Warp AI hỗ trợ nhà cung cấp tùy chỉnh",
"agent-deck": "Trình điều phối đa tác nhân Agent Deck"
},
@@ -4627,6 +4622,13 @@
"retry": "Thử lại",
"allOperational": "Tất cả hệ thống đang hoạt động bình thường",
"issuesDetected": "Phát hiện sự cố hệ thống",
"healthVerdictReady": "OmniRoute đã sẵn sàng",
"healthVerdictActionRequired": "Cần hành động để khôi phục hoạt động đầy đủ",
"healthVerdictCoolingDown": "Đang nguội sau các thay đổi gần đây",
"healthSubtitle": "Kiểm tra tình trạng hệ thống",
"advancedDiagnosticsTitle": "Chẩn đoán nâng cao",
"hide": "Ẩn",
"show": "Hiện",
"updatedAt": "Đã cập nhật {time}",
"latency": "Độ trễ",
"latencyP50": "p50",

3
src/shared/schemas/cliCatalog.ts
View File

@@ -67,4 +67,5 @@ export const EXPECTED_CODE_COUNT = 21;
// +2 (#6318): "omp" (Oh My Pi) and "letta" (Letta CLI) added as agent entries.
// Note: #6318 originally also shipped duplicate "pi"/"jcode"/"codewhale" entries —
// those tools were already delivered by a separate PR, so only omp+letta landed here.
export const EXPECTED_AGENT_COUNT = 8;
// +1 (#11166): "prime-agent" (PrimeIntellect-ai/prime-agent) added as an agent entry.
export const EXPECTED_AGENT_COUNT = 9;

7
tests/unit/cli-catalog-counts.test.ts
View File

@@ -41,8 +41,8 @@ test("CLI_TOOLS total code entries (including none) equals 26 (21 visible + 5 no
assert.equal(codeAll.length, 26, `Expected 26 total code entries, got ${codeAll.length}`);
});
test("CLI_TOOLS total (code + agent) = 34", () => {
assert.equal(all.length, 34, `Expected 34 total entries, got ${all.length}`);
test("CLI_TOOLS total (code + agent) = 35", () => {
assert.equal(all.length, 35, `Expected 35 total entries, got ${all.length}`);
});
test("All code-none entries have configType mitm OR are legacy excluded entries", () => {
@@ -99,7 +99,7 @@ test("The 21 visible code entries include Qwen Code's rebuilt integration", () =
}
});
test("The 8 agent entries match D15 list exactly (+ omp + letta, #6318)", () => {
test("The 9 agent entries match D15 list exactly (+ omp + letta #6318, + prime-agent #11166)", () => {
const d15Agents = new Set([
"hermes-agent",
"openclaw",
@@ -109,6 +109,7 @@ test("The 8 agent entries match D15 list exactly (+ omp + letta, #6318)", () =>
"agent-deck",
"omp",
"letta",
"prime-agent",
]);
const agentIds = new Set(agentAll.map((t) => t.id));
for (const id of d15Agents) {

2
tests/unit/cli-tools-schema.test.ts
View File

@@ -11,6 +11,7 @@ test("CLI_TOOLS registry contains all expected tools including rebuilt Qwen Code
// (CodeWhale is the actively-maintained successor to DeepSeek TUI).
// omp + letta added by #6318 (agent-category CLI integrations).
// grok-build added — xAI Grok Build TUI coding agent (ported from upstream decolua/9router#2571).
// prime-agent added by #11166 (PrimeIntellect-ai/prime-agent, agent category).
const expected = [
"claude",
"codex",
@@ -46,6 +47,7 @@ test("CLI_TOOLS registry contains all expected tools including rebuilt Qwen Code
"grok-build",
"qwen",
"zcode",
"prime-agent",
];
for (const id of expected) {
assert.ok(id in CLI_TOOLS, `Missing tool: ${id}`);

5
tests/unit/cli/setup-qwen.test.ts
View File

@@ -42,6 +42,9 @@ test("setup-qwen writes current V4 settings and only its dedicated env key", asy
configPath: settingsPath,
envPath,
yes: true,
// These tests exercise the merge/write logic, not the container guard
// (#10057) — keep them hermetic on container devboxes/CI.
allowContainerWrite: true,
});
assert.equal(code, 0);
@@ -76,6 +79,8 @@ test("setup-qwen does not overwrite an invalid settings file", async () => {
model: "model-id",
configPath: settingsPath,
yes: true,
// See above — hermetic regardless of container detection (#10057).
allowContainerWrite: true,
});
assert.equal(code, 1);
assert.equal(await fs.readFile(settingsPath, "utf8"), "{ invalid JSON");

6
tests/unit/models-catalog-combo-metadata.test.ts
View File

@@ -98,7 +98,11 @@ test("single-target Codex combo advertises a larger model context override", asy
assert.equal(response.status, 200);
assert.equal(direct?.context_length, contextWindow);
assert.equal(combo?.context_length, contextWindow);
assert.equal(combo?.max_input_tokens, 272000);
// #11179 raised the static codex catalog cap to maxInputTokens=872000 (the real
// usable window; the old 272000 was just the first pricing tier). The input cap
// can never exceed the total window, so with the 500K override it clamps to it:
// min(872000, 500000) = 500000.
assert.equal(combo?.max_input_tokens, 500000);
} finally {
contextOverrides.removeModelContextOverride("codex", modelId);
}

5
tests/unit/provider-models-route-codex.test.ts
View File

@@ -181,10 +181,11 @@ test("provider models route merges live Codex models with the local catalog then
// merge conservatively — the smaller of live vs. pinned wins, never the
// larger, so a stale/inflated live number can never make OmniRoute promise
// more context than the account can actually serve (#7012). Here the pinned
// GPT-5.6 Codex contract (272000/128000, see GPT_5_6_CODEX_CAPABILITIES)
// GPT-5.6 Codex contract (872000/128000, see GPT_5_6_CODEX_CAPABILITIES — raised
// from the old 272K pricing tier to the real usable window by #11179)
// is smaller than the live payload's 999999/999999, so the pinned value wins.
assert.equal(liveModel?.name, "GPT 5.6 Sol Live");
assert.equal(liveModel?.inputTokenLimit, 272000);
assert.equal(liveModel?.inputTokenLimit, 872000);
assert.equal(liveModel?.outputTokenLimit, 128000);
assert.equal(liveModel?.apiFormat, "responses");
assert.deepEqual(liveModel?.supportedEndpoints, ["responses"]);

42
tests/unit/search-route.test.ts
View File

@@ -420,25 +420,24 @@ test("v1 search POST preserves stored SearXNG baseUrl for authless providers", a
}
});
test("v1 search POST returns 400 when auto-select finds no configured provider (searxng-search is now fallbackOnly)", async () => {
test("v1 search POST falls back to duckduckgo-free when no provider is configured (#11097)", async () => {
// Contract changed by PR #11097 ("fix(search): fall back to duckduckgo-free when
// no search provider is configured"): zero-credential /v1/search no longer returns
// 400 — it promotes the fallback-only duckduckgo-free provider so out-of-the-box
// search works. This test pins the NEW contract.
const originalFetch = globalThis.fetch;
let capturedUrl = "";
// DuckDuckGo lite HTML shape: result link + snippet cell (see
// open-sse/services/freeWebSearch.ts parseDuckDuckGoLite).
const liteHtml = `<html><body>
<a href="https://example.com/auto-result" class='result-link'>Auto-selected DuckDuckGo result</a>
<td class='result-snippet'>Fallback free search snippet</td>
</body></html>`;
globalThis.fetch = async (url) => {
capturedUrl = String(url);
return new Response(
JSON.stringify({
results: [
{
title: "Auto-selected SearXNG result",
url: "https://searx.example/auto",
content: "Auto-selected self-hosted response",
engines: ["duckduckgo"],
},
],
}),
{ status: 200, headers: { "content-type": "application/json" } }
);
return new Response(liteHtml, { status: 200, headers: { "content-type": "text/html" } });
};
try {
@@ -454,14 +453,15 @@ test("v1 search POST returns 400 when auto-select finds no configured provider (
);
const body = (await response.json()) as any;
assert.equal(response.status, 400);
assert.equal(capturedUrl, "", "fallback-only SearXNG must not receive an upstream request");
assert.ok(body.error?.message || body.error);
assert.match(
String(body.error?.message ?? body.error),
/provider|configured/i,
"the response must explain that no provider was selected"
assert.equal(response.status, 200);
assert.equal(
capturedUrl,
"https://lite.duckduckgo.com/lite/",
"the fallback must call the DuckDuckGo lite endpoint"
);
assert.equal(body.provider, "duckduckgo-free");
assert.equal(body.results[0].title, "Auto-selected DuckDuckGo result");
assert.equal(body.results[0].url, "https://example.com/auto-result");
} finally {
globalThis.fetch = originalFetch;
}

6
tests/unit/vscode-token-routes-gpt56.test.ts
View File

@@ -128,9 +128,11 @@ test("vscode raw models route exposes native GPT-5.6 IDs and effort tiers", asyn
assert.equal(typeof defaultModel.created, "number");
assert.equal(defaultModel.owned_by, "codex");
assert.equal(defaultModel.name, "Codex GPT 5.6 Sol");
assert.equal(defaultModel.context_length, 272000);
// #11179: codex static catalog advertises the usable 872K window (max_context_window),
// not the old 272K pricing tier.
assert.equal(defaultModel.context_length, 872000);
assert.equal(defaultModel.max_output_tokens, 128000);
assert.equal(defaultModel.max_input_tokens, 272000);
assert.equal(defaultModel.max_input_tokens, 872000);
assert.deepEqual(defaultModel.capabilities, {
vision: true,
tool_calling: true,

11
tests/unit/vscode-token-routes.test.ts
View File

@@ -255,7 +255,9 @@ test("vscode combos route resolves combo names through Ollama api/show", async (
assert.equal(body.model, "show-combo");
assert.equal(body.modelfile, "FROM show-combo");
assert.equal(body.details.family, "show-combo");
assert.equal(body.model_info.context_length, 272000);
// #11179: codex static catalog advertises the usable 872K window (max_context_window),
// not the old 272K pricing tier.
assert.equal(body.model_info.context_length, 872000);
assert.deepEqual(body.supportsReasoningEffort, ["none", "low", "medium", "high", "xhigh"]);
assert.equal(body.model_info.capabilities.reasoning, true);
});
@@ -290,7 +292,8 @@ test("vscode tokenized combos root route exposes importable combo metadata", asy
assert.equal(response.status, 200);
assert.ok(combo, "expected balanced-load in combo root response");
assert.equal(combo.url.includes("/responses#models.ai.azure.com"), true);
assert.equal(combo.maxInputTokens, 272000);
// #11179: codex static catalog maxInputTokens is now the usable 872K window.
assert.equal(combo.maxInputTokens, 872000);
assert.equal(combo.toolCalling, true);
assert.deepEqual(combo.supportsReasoningEffort, ["none", "low", "medium", "high", "xhigh"]);
});
@@ -1073,7 +1076,9 @@ test("vscode tokenized api/show route exposes explicit reasoning effort metadata
assert.equal(body.configurationSchema?.properties?.reasoningEffort?.default, "low");
assert.equal(body.model_info["general.basename"], "Codex GPT 5.6 Sol (Default)");
assert.equal(body.model_info["general.architecture"], "codex");
assert.equal(body.model_info["codex.context_length"], 272000);
// #11179: codex static catalog advertises the usable 872K window (max_context_window),
// not the old 272K pricing tier.
assert.equal(body.model_info["codex.context_length"], 872000);
assert.deepEqual(body.model_info.supports_reasoning_effort, [
"low",
"medium",