End-to-end: run two test files at a time in each shard

Each end-to-end file starts its own `wrangler dev` (own port, database and
directory), and one file's server mostly keeps one of the runner's four
cores busy. Locally, two at a time took the passkeys/registry/email files
from 234 s to 134 s (all 48 tests passing in each of three runs) and the
agents files from 261 s to 136 s; security gains little (one big file).
Shards with one file are unchanged.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Br8vDafY9vkPRfDRkkcVBt

test-bot authored 10 days agosigned1 parent c996868commit 2d4766c6c3974cc754f4aeb63bb112273186ff1fBrowse files

CI / Typecheck and bundleFailed
CI / Unit tests (1/2)Failed
CI / Unit tests (2/2)Failed
End-to-end / End-to-end (agent reviews, tasks, Today)Failed
End-to-end / End-to-end (git, PRs, Actions)Failed
End-to-end / End-to-end (large pushes)Failed
End-to-end / End-to-end (passkeys, registry, email)Failed
End-to-end / End-to-end (real histories)Failed
End-to-end / End-to-end (runner Docker)Failed
End-to-end / End-to-end (security)Failed

1 file changed+3−1

modified.github/workflows/e2e.yml+3−1View file

		@@ -81,7 +81,9 @@
81	81	) files=$(ls test/e2e/.test.mjs \| grep -vE "$core\|$agents\|$security") ;;
82	82	esac
83	83	echo "$files"
84		- node --test --test-concurrency=1 "${opts[@]}" $files
	84	+ # Two files at a time: each starts its own local server on its own port, database and
	85	+ # directory, so they don't share state, and a file's server mostly keeps one core busy.
	86	+ node --test --test-concurrency=2 "${opts[@]}" $files
85	87
86	88	# Docker as the hosted runner image runs it, with this change's docker wrapper, daemon config
87	89	# and agent (its registry relay): this job runs on a hosted runner, so the real dockerd is there.

codician/diffinerypublic

End-to-end: run two test files at a time in each shard