mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-09-15 19:32:20 +03:00
* test(infra): retry recursive temp-dir removal on main (main twin of #11968)
`main` has been red since b342c1a361 on the vitest and integration gates:
✖ tests/unit/autoCombo/provider-family-combos.test.ts > auto/<family>
✖ chat pipeline applies Codex OAuth fingerprint and priority tier inside combos
Both call resetStorage() from beforeEach, which does an fs.rmSync(TEST_DATA_DIR,
{recursive: true, force: true}) with no retry, and intermittently loses the race
with a not-yet-released SQLite handle (ENOTEMPTY).
release/v3.8.51 fixed this in #11968 with a mechanical codemod adding
maxRetries/retryDelay to every recursive rm/rmSync/rmdirSync under tests/, but
that PR landed only on the release branch. Because main only receives work at
the release squash, it stayed broken for the whole cycle — and repo-wide gates
then turn every open PR into main red on checks unrelated to their diff.
This is the --base main twin: re-runs the same codemod that already shipped on
the release branch (scripts/ad-hoc/codemod-rm-maxretries.mjs), so the two
branches converge on identical test-teardown semantics. Test-only; no product
logic is touched.
The remaining three failures reported on #12133 (unit full suite exceeding its
4800s ceiling, package-artifact exceeding 1200s, and the boot-smoke that is
skipped as a consequence) are runner-contention timeouts, not code defects —
validate-release-green.mjs runs those heavy gates concurrently on one shared
hosted runner. There is no fix to port for those.
* chore(scripts): carry the rm-maxretries codemod onto main alongside its output
The codemod that generated the previous commit lives in the repo on
release/v3.8.51 (added by #11968) but was never on main. Bringing it over keeps
the tool next to the change it produced, so the transformation stays
reproducible and auditable from either branch.
91 lines
3.7 KiB
TypeScript
91 lines
3.7 KiB
TypeScript
// tests/unit/combo-context-prefix-resolution.test.ts
|
|
// Regression guard for computeComboContextLength()'s provider-prefix bug.
|
|
//
|
|
// resolveNestedComboTargets() returns target.modelStr in "provider/model" form
|
|
// (e.g. "glm/glm-5.2"), but computeComboContextLength() used to pass that
|
|
// qualified string straight into getCanonicalModelMetadata() without stripping
|
|
// the prefix first — unlike the catalog's own getComboTargetCatalogMetadata(),
|
|
// which strips it via getComboTargetModelId()/getProviderPrefixes() before the
|
|
// lookup. The alias-resolution chain (getResolvedModelCapabilities ->
|
|
// resolveCanonicalProviderModel -> resolveProviderModelAlias) does an exact-match
|
|
// lookup keyed by the BARE registry id, so a "provider/model" string only
|
|
// resolved for the handful of models with a curated MODEL_SPECS alias in that
|
|
// exact qualified form — every other registry-only model (the vast majority)
|
|
// silently fell out of the min() computation, and computed_context_length was
|
|
// dropped from the /api/combos response entirely.
|
|
//
|
|
// This test uses glm-5.2 (open-sse/config/providers/registry/glm — real
|
|
// 1,000,000-token context, no "glm/glm-5.2"-form curated alias) — the exact
|
|
// class of model the bug affected. Confirmed empirically before this fix
|
|
// landed: computeComboContextLength() returned `undefined` for a combo whose
|
|
// only member was "glm/glm-5.2", even though the bare "glm-5.2" resolves fine
|
|
// via getCanonicalModelMetadata() directly.
|
|
import test from "node:test";
|
|
import assert from "node:assert/strict";
|
|
import fs from "node:fs";
|
|
import os from "node:os";
|
|
import path from "node:path";
|
|
|
|
const TEST_DATA_DIR = fs.mkdtempSync(
|
|
path.join(os.tmpdir(), "omniroute-combo-context-prefix-resolution-")
|
|
);
|
|
process.env.DATA_DIR = TEST_DATA_DIR;
|
|
|
|
const core = await import("../../src/lib/db/core.ts");
|
|
const { computeComboContextLength } = await import("../../src/lib/combos/comboContext.ts");
|
|
const { setModelContextOverride, removeModelContextOverride } =
|
|
await import("../../src/lib/db/modelContextOverrides.ts");
|
|
|
|
test.after(() => {
|
|
core.resetDbInstance();
|
|
fs.rmSync(TEST_DATA_DIR, { recursive: true, force: true, maxRetries: 5, retryDelay: 100 });
|
|
});
|
|
|
|
test("computeComboContextLength resolves a registry-known, prefixed member (glm/glm-5.2) to its real context window", () => {
|
|
const combo = {
|
|
name: "prefix-resolution-probe-single",
|
|
models: ["glm/glm-5.2"],
|
|
};
|
|
|
|
const result = computeComboContextLength(combo, []);
|
|
|
|
assert.equal(
|
|
result,
|
|
1000000,
|
|
"glm/glm-5.2 is a real registry model with a 1,000,000-token context window " +
|
|
"(open-sse/config/providers/registry/glm) — the prefix must be stripped " +
|
|
"before the canonical-model lookup so it is not silently excluded"
|
|
);
|
|
});
|
|
|
|
test("computeComboContextLength takes the minimum across multiple prefixed, registry-known members", () => {
|
|
const combo = {
|
|
name: "prefix-resolution-probe-multi",
|
|
// glm-4.5 (128,000) is the smaller of the two known windows.
|
|
models: ["glm/glm-5.2", "glm/glm-4.5"],
|
|
};
|
|
|
|
const result = computeComboContextLength(combo, []);
|
|
|
|
assert.equal(
|
|
result,
|
|
128000,
|
|
"the minimum across all resolved (prefix-stripped) members should win, " +
|
|
"matching the catalog's minKnownNumber semantics"
|
|
);
|
|
});
|
|
|
|
test("computeComboContextLength honors a larger persisted Codex GPT-5.6 window", () => {
|
|
const modelId = "gpt-5.6-terra";
|
|
assert.equal(setModelContextOverride("codex", modelId, 500000, "manual"), true);
|
|
try {
|
|
assert.equal(
|
|
computeComboContextLength({ models: [`codex/${modelId}`] }, []),
|
|
500000,
|
|
"the combo aggregate must use the effective override instead of the Codex registry default"
|
|
);
|
|
} finally {
|
|
removeModelContextOverride("codex", modelId);
|
|
}
|
|
});
|