Stacked features 1: add typecheck lint#412
Conversation
|
Thanks a lot for this contribution! A stacked series covering search performance, build caching, and tooling upgrades is a substantial piece of work, and the benchmark numbers look very promising. We really appreciate the effort. 🙏 Two process asks before we dive into review:
Once the descriptions are in place we'll take the stack through review. Thanks again! |
gotcha |
|
Thanks for the guidance. I have updated all seven PRs with standalone English descriptions covering each PR's motivation, owned changes, verification, and review notes. Each description now also links its position and dependency in the stack. I also checked the code added by the stack for non-English comments. The only non-English additions are Chinese search strings used as CJK/trigram test fixtures; code comments remain in English. |
|
Follow-up on the failed unit job: the clean runner did not have Commit Verification:
|
|
Thanks for the contribution and the quick follow-up — merging. Looking forward to the rest of the stack. |
…rboytm#426) The July 2026 optimization stack (paperboytm#412-paperboytm#418) shipped, which turns its 441-line plan into a completed one — and completed plans rot: the doc went stale twice during the stack's own lifetime (feature-flag removals, find semantics). Git history and the PR descriptions now carry the full rationale and verification notes permanently. Keep what stays useful: - The verification matrix moves to CONTRIBUTING as a living 'Verifying changes' section, including the packaged-app checklist and the never-remove-assets-on-grep-alone rule. - The roadmap shrinks to a ~50-line record: what shipped per PR, decisions of record with their data basis, and known follow-ups. Also refresh two stale CONTRIBUTING sections: native-ABI handling is automatic since paperboytm#415 (manual rebuilds are the fallback), and 'no linter config yet' predates oxlint (paperboytm#413). Co-authored-by: Chen <99816898+donteatfriedrice@users.noreply.github.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Stack position
Part 1 of 7. This is the base of the stack; the next PR is #413.
Why
Before changing the compiler or linter, the repository needs a complete and trustworthy TypeScript baseline. Several workspace packages were outside the root typecheck graph, and enabling them exposed real app and landing-page errors.
What changed
typecheckpolicy to every workspace package.typecheckandchecktasks through Turbo.__SPOOL_E2E__build global.Verification
pnpm typecheck: 9/9 workspace packages passed.pnpm lint: passed.pnpm --filter @spool/landing build: passed.pnpm --filter @spool/app build:electron: main, preload, and renderer bundles passed.Review notes
This PR intentionally stays on the existing TypeScript version. It establishes the green baseline required by #413 and #414. The pre-existing shared
better-sqlite3ABI race is not hidden with longer timeouts; #415 owns that fix.