ci: shard package-local test suites

[jgoux]

Summary

• Move TypeScript unit, integration, and end-to-end execution to package-local Vitest scripts orchestrated by Turbo while keeping live tests separate.
• Run unit and integration suites in independent CI jobs so integration tests can be sharded separately later.
• Apply each end-to-end shard to every package that owns an e2e suite, executing packages sequentially within a runner.
• Encode Go and CLI build prerequisites for local e2e entrypoints, then isolate Vitest passthrough arguments from those build tasks.
• Keep Go unit tests package-local and in their dedicated race-and-coverage workflow while the root Vitest fan-out remains uniformly shardable.
• Right-size the unit job to a 4-vCPU Blacksmith runner without preemptively capping workers.
• Remove the test:core alias and document explicit unit and integration commands.

Context

The first e2e shard previously ran every non-CLI-harness e2e suite unsharded, making it consistently much slower than the other shards. Propagating the same Vitest shard to each package distributes those suites uniformly.

Turbo currently provides only the test task graph and prerequisite ordering. Test caching remains disabled until its inputs and outputs are designed separately.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:9d0417ceaa

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

Supabase CLI preview

npx --yes https://pkg.pr.new/supabase/cli/supabase@58fd79378ca5802df52da51144b6f70e9fd45527

Preview package for commit 58fd793.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:d99065a918

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[jgoux]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. What shall we delve into next?

Reviewed commit:4000d4b12b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[avallete]

LGTM — verified the risky spots and they all check out:

• The 8 @parcel/watcher-* platform packages land in devDependencies only; the published supabase package still ships just eciesjs + jose with files: ["dist/supabase.js"], so no platform-mismatch risk for npm consumers (and this quietly fixes the pnpm build break on missing watcher binaries, including win32 which supportedArchitectures doesn't cover).
• Forwarding --shard=N/3 to every e2e package is safe: all three e2e-owning vitest configs set passWithNoTests: true (packages/stack only has 4 e2e files).
• Dropping the keyring-unlock step from test.yml is correct: cli-go-ci.yml keeps the unlock and its merge_group trigger has no path filter, so Go tests still gate every merge even when a PR doesn't touch apps/cli-go/.
• No test files orphaned by the removed vitest project stanzas (api/config have zero integration/e2e files, process-compose zero e2e), and every package with unit tests declares test:unit:run.
• test-summary preserves the branch-protection check name and mirrors the existing test-e2e-summary skipped-handling pattern with identical if conditions.
• No stale test:core / nx affected / NX_BASE references remain.

Two non-blocking comments inline about leftover Nx test-target inference and the dual build orchestrators.