Skip to content

The all-in-one local AI studio for your desktop: chat, image and video generation and a coding agent in one free, open source app. Windows and Linux. No Docker, no terminal, no cloud required.

ci(e2e): four shards on four machines, 90 minutes each at most

purplebdb9660f60fd
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index be798bb7..c30848e0 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -101,6 +101,14 @@ jobs:
e2e:
runs-on: ubuntu-22.04
+ # Four machines, a quarter of the specs each: one serial run took about
+ # 3 h (1.8k tests, workers 1), and the release gate runs this job too.
+ # Inside a shard it stays one worker, as playwright.config.ts sets it.
+ timeout-minutes: 90
+ strategy:
+ fail-fast: false
+ matrix:
+ shard: [1, 2, 3, 4]
steps:
- name: Checkout repository
@@ -127,8 +135,8 @@ jobs:
# The harness stubs the Tauri IPC in-page (e2e/support/tauri-mock.ts) —
# no Rust sidecar, Ollama or ComfyUI processes needed.
- - name: Playwright e2e (Tauri-mocked)
- run: npm run test:e2e
+ - name: Playwright e2e (Tauri-mocked), shard ${{ matrix.shard }} of 4
+ run: npm run test:e2e -- --shard=${{ matrix.shard }}/4
# The voice sidecar is Python and ships as source (src-tauri/resources/
# voice_loop). Its tests need no voice models: they run against hollow model