smashimo/oh-my-pi
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

merged PR #5408: fix(catalog): widen GLM coding-plan idle timeout to OpenCode gateways

Browse Source
This commit is contained in:
can1357
2026-07-16 03:32:04 +02:00
parent f39c1ce103 7cdecd824a
commit b68c7245f7
3 changed files with 37 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
packages/catalog/CHANGELOG.md
+3
View File
@@ -145,6 +145,9 @@
### Fixed
- Extended the reasoning `streamIdleTimeoutMs` floor (300s) to native Kimi K2.7 Code (`kimi-k2.7-code` / `kimi-k2.7-code-highspeed`), which previously fell through to the 120s default and aborted on long reasoning turns ([#4836](https://github.com/can1357/oh-my-pi/issues/4836)).
### Fixed
- Fixed GLM-5.x coding-plan streams via the OpenCode Go/Zen gateways (`opencode.ai/zen/…`) timing out with `OpenAI completions stream stalled while waiting for the next event` during slow plan-writing/reasoning phases. The 600s idle-timeout floor for GLM coding-plan SKUs was gated to the native Z.AI/Zhipu hosts only, so OpenCode-fronted GLM fell back to the 120s default watchdog. ([#4758](https://github.com/can1357/oh-my-pi/issues/4758))
## [16.3.11] - 2026-07-06
packages/catalog/src/compat/openai.ts
+1 -1
View File
@@ -358,7 +358,7 @@ export function buildOpenAICompat(spec: ModelSpec<"openai-completions">): Resolv
// for minutes while reasoning or cold-loading weights; widen the idle
// timeout so warm-ups stop aborting and retrying.
const streamIdleTimeoutMs =
GLM_CODING_PLAN_MODEL_PATTERN.test(spec.id) && (isZai || isZhipu)
GLM_CODING_PLAN_MODEL_PATTERN.test(spec.id) && (isZai || isZhipu || isOpenCodeHost)
? GLM_CODING_PLAN_STREAM_IDLE_TIMEOUT_MS
: provider === "alibaba-coding-plan"
? ALIBABA_CODING_PLAN_STREAM_IDLE_TIMEOUT_MS
packages/catalog/test/zhipu-compat.test.ts
+33
View File
@@ -120,6 +120,39 @@ describe("openai-completions compat — zhipu-coding-plan branch", () => {
});
});
describe("openai-completions compat — GLM coding-plan stream idle timeout", () => {
function glm52(provider: string, baseUrl: string): ModelSpec<"openai-completions"> {
return { ...baseModel, id: "glm-5.2", name: "GLM-5.2", provider, baseUrl };
}
// GLM coding-plan SKUs idle for minutes mid-reasoning; the 600s watchdog
// floor must apply on every gateway that fronts them, not just the native
// Z.AI/Zhipu hosts (issue #4758: GLM-5.2 via opencode-go stalled with
// "OpenAI completions stream stalled while waiting for the next event").
it("widens the idle timeout to 600s for GLM-5.x on Z.AI, Zhipu, and OpenCode gateways", () => {
expect(buildOpenAICompat(glm52("zai", "https://api.z.ai/api/coding/paas/v4")).streamIdleTimeoutMs).toBe(600_000);
expect(
buildOpenAICompat(glm52("zhipu-coding-plan", "https://open.bigmodel.cn/api/coding/paas/v4"))
.streamIdleTimeoutMs,
).toBe(600_000);
expect(buildOpenAICompat(glm52("opencode-go", "https://opencode.ai/zen/go/v1")).streamIdleTimeoutMs).toBe(
600_000,
);
expect(buildOpenAICompat(glm52("opencode-zen", "https://opencode.ai/zen/v1")).streamIdleTimeoutMs).toBe(600_000);
});
it("does not widen non-GLM models on the OpenCode gateway via the GLM floor", () => {
const kimi = buildOpenAICompat({
...baseModel,
id: "kimi-k2.5",
name: "Kimi K2.5",
provider: "opencode-go",
baseUrl: "https://opencode.ai/zen/go/v1",
});
expect(kimi.streamIdleTimeoutMs).toBeUndefined();
});
});
describe("zhipu-coding-plan model discovery", () => {
it("uses the dedicated Coding Plan endpoint by default", async () => {
let requestedUrl = "";