Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 900+ contributors
TypeScript a2aai-agentsai-gatewayanthropicclaudeclaude-codeclinecodexcopilotcursordeepseekfree-aigeminikimillm-gatewaymcpopenaiopenai-proxyqwentoken-saver
/**
* /api/v1/models catalog cache must survive routine routing/health writes.
*
* The unified models catalog build is expensive (multi-second cold build);
* #13389 introduced `skipModelCatalog` for writes that touch only routing /
* health columns. This test pins the known routing-only write sites so a
* refactor cannot silently reintroduce the cache-bust (the live symptom was
* an ~8 s cold rebuild of /api/v1/models on nearly every dashboard hit,
* because 429 cooldown and recovery writes fire continuously during normal
* routing).
*/
import test from "node:test";
import assert from "node:assert/strict";
import { readFileSync } from "node:fs";
import { join, dirname } from "node:path";
import { fileURLToPath } from "node:url";
const __dirname = dirname(fileURLToPath(import.meta.url));
const REPO_ROOT = join(__dirname, "..", "..");
const SITES: Array<{ file: string; context: string }> = [
{
file: "src/lib/db/providers/rateLimit.ts",
context: "setConnectionRateLimitUntil — rate_limited_until write",
},
{
file: "src/lib/db/providers/rateLimit.ts",
context: "batch recovery reset — error/cooldown columns",
},
{
file: "src/lib/db/providers/rateLimit.ts",
context: "clearConnectionErrorIfUnchanged — conditional error clear",
},
{
file: "src/lib/db/providers/codexAccountRecovery.ts",
context: "codex quota/cooldown observation mutate()",
},
];
test("routing-only connection writes skip the model-catalog cache bust", () => {
const sources = new Map<string, string>();
for (const site of SITES) {
if (!sources.has(site.file)) {
sources.set(site.file, readFileSync(join(REPO_ROOT, site.file), "utf8"));
}
}
for (const [file, source] of sources) {
const busts = source.match(/invalidateDbCache\("connections"[^)]*\)/g) ?? [];
const skipping = busts.filter((call) => call.includes("skipModelCatalog"));
assert.ok(
busts.length > 0,
`${file} no longer calls invalidateDbCache("connections") — update this test if the write moved`
);
assert.deepEqual(
busts.length,
skipping.length,
`${file}: every "connections" invalidation must pass { skipModelCatalog: true } (found ${skipping.length}/${busts.length}). Structural writes (create/update/delete connections) should use a narrower scope or keep the default — see readCache.ts #13389 docs.`
);
}
});
test("OAuth token rotation persists with skipModelCatalog", () => {
const source = readFileSync(join(REPO_ROOT, "src/sse/services/tokenRefresh.ts"), "utf8");
assert.ok(
source.includes("catalogRelevant") && source.includes("skipModelCatalog: true"),
"updateProviderCredentials must pass { skipModelCatalog: true } for credential-only updates (token refreshes fire continuously and otherwise cold-rebuild /v1/models)"
);
assert.ok(
source.includes("updates.providerSpecificData !== undefined"),
"the skip must exclude updates carrying providerSpecificData (can change catalog-relevant state)"
);
});