Repository navigation
UN-4270 [FEAT] v1 scaffold, strict TypeScript and import boundary - #2380
Merged
Merged
Conversation
…N-4270) Adds frontend/src/v1/ as an isolated tree for the v1 frontend (P0-T01): - Folder taxonomy from the spec, a public entry (src/v1/index.ts, imported by v0 as `@v1`) and a cloud overlay placeholder. - src/v1/tsconfig.json: strict, no JS, unused locals/params are errors, `@v1` alias only. v0's tsconfig/jsconfig exclude src/v1. - Biome overrides for the two-way import boundary and v1 bans (no `../`, `@/`, prop-types, antd, component CSS; noExplicitAny). - scripts/check-v1.mjs for what Biome cannot express: TypeScript-only files, the folder taxonomy and import contract, PropTypes, `ant-`. - typecheck:v1, lint:v1 and test:v1 scripts; v0 scripts are unchanged. - test-fixtures/v1-boundary/ proves each rule fails and passes as specified, run with node --test so the v0 vitest suite never sees it. - .gitignore re-includes src/v1/lib/ (the Python `lib/` rule matched it). v0 build output, test counts, typecheck and Biome results are unchanged.
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
…270) check-v1.mjs: - Scan comments with a string/template-aware tokenizer; a `/*` or `//` inside a string or line comment no longer hides the imports after it. - Read template-literal specifiers, import.meta.glob and `import(..., opts)`. - Flag specifiers that leave src/v1 (`/`, `..`, `@/`, `src/`) itself, not only via Biome, which skips template literals. - Check `@import`/`url()` and `ant-` in styles/ CSS; report symlinks; catch built `ant-` names; treat .test/.spec .mts/.cts as test files; ignore .DS_Store; drop the unused --root option. biome.json: also ban `..`, `./../**`, `/**`, `antd/**`, `prop-types/**`. tsconfig: skipLibCheck off so v1's own .d.ts files are checked. .gitignore: re-include lib/ at any depth under src/v1. Fixtures: a generated matrix over every folder-contract pair, every import form, every outside spelling, a symlink case, and cases for each new rule; tsc cases assert error codes; output matching excludes the command text and spawn errors fail loudly. Comment and README corrections.
Contributor
Author
Self-review (draft)Four review passes (code, tests, comments, silent failures) were run against Fixed
Not changed, by design or out of scope
v0 regression (re-run on
|
… (UN-4270) - ci-frontend-lint: the typecheck job also runs `typecheck:v1` and `lint:v1`, so the strict v1 tsconfig and the scripts/check-v1.mjs checks (folder contract, TypeScript-only files, PropTypes, `ant-`) gate PRs now rather than from P0-T10. - Pages may no longer import app/, test/ or the v1 entry, only the shared folders (components, shell, services, stores, hooks, lib, styles, plugins). Test files in a page may still import test/. Contract table, README and the fixture matrix updated together.
…(UN-4270) check-v1.mjs: split the comment scanner into small helpers (cognitive complexity 40 -> under 15), drop the nested ternary and nested template literal, use String.raw for the import regexes, keep CLI flags in a Set. Fixture runner: - Cases and the expected folder contract are parsed text tables instead of repeated object literals (the 10% duplication). - Sorts use an explicit localeCompare comparator. - The contract matrix is built by a helper. - Spawns /bin/sh by absolute path and no longer needs `git init` (Biome only needs the ignore file). The git check-ignore call keeps a PATH lookup, with the reason and a NOSONAR marker. Fixture cases: re-exports rewritten as small functions and read-only props, without changing what each case exercises.
jaseemjaskp
marked this pull request as ready for review
October 27, 2026 05:22
Contributor
 [Medium impact] The PR appears safe to merge, with no new actionable issues found.
|
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
…nnot see (UN-4270) Addresses Greptile review on #2380: - import.meta.glob([...]): every pattern in the array is checked, not only the first, so an allowed first pattern no longer hides a forbidden one. - check-v1 --code also scans v0 (src/ outside v1, including the cloud src/plugins overlay) for template-literal imports, globs, require(), @/v1, src/v1, relative paths and CSS @import that reach into v1 other than through `@v1`. Biome still covers plain imports. Fixtures: Greptile's example as a case, a glob array with a forbidden later pattern, and a generated test of every v0 -> v1 form (denied and allowed). The v0 scan reports nothing on the current tree.
muhammad-ali-e
left a comment
muhammad-ali-e
left a comment
Contributor
There was a problem hiding this comment.
I reviewed this from the Lane I side. It looks solid: typecheck:v1, lint:v1 and test:v1 all pass on the branch (40/40), and the folder contract matches the spec. I have one inline comment, on tsconfig.json, and one question:
Static assets. The taxonomy has no folder for static assets, so src/v1/assets/logo.svg fails --files. Lane I's P1-T22 needs SVG icons and provider logos, and v1 can't reach v0's src/assets/. Today components/ui/icons/x.svg passes, imported as ./icons/x.svg?react. Is that the intended place, or should there be an assets/ folder? A line in the README either way would stop each lane picking its own spot. The spec says only that generated icon maps live under v1.
Generated by Claude Code
muhammad-ali-e
left a comment
muhammad-ali-e
left a comment
Contributor
There was a problem hiding this comment.
This follows up my earlier review with a fuller pass, using the team's standardized 19-lens rubric (unstract:review-focused: one cold reader, then specialists for security, testing, and dependencies/build). Nothing here blocks merge. There's no Critical or High finding, and security, cross-document consistency and the v0 no-regression claims all checked out. The inline comments are mostly fixture gaps, where a guard rule can break without any test failing.
Two notes that don't need a line comment:
- CI doesn't run the fixtures. You've already called that P0-T10 scope. Until P0-T10 lands, the fixtures are the only check on the Biome-override semantics that the "Why some rules live in a script" section depends on.
- The "Notes on Testing" section says
test:v1passes 22/22. It now runs 40 tests.
Generated by Claude Code
This was referenced Oct 28, 2026
athul-rs
reviewed
Oct 28, 2026
athul-rs
left a comment
athul-rs
left a comment
Contributor
There was a problem hiding this comment.
Verdict — APPROVE WITH NITS
Summary — Critical: 0 · High: 0 · Medium: 5 · Low: 6 · Lenses run: 19/19 (7 assessed N/A: 5, 6, 8, 10, 12, 18, 19)
Reviewed at 81e15b159 against merge-base 6e14fa790 (71 files, +1211/-11). CI is green at this head, including Run v1 typecheck and Run v1 lint.
The Medium findings are fix-soon, not blockers. #2, #3 and #4 are silent until v1 holds code, so they should land (or be ticketed) before the first real v1 code arrives (UN-4271 onward). If one is expected to be hit sooner, the verdict moves to REQUEST CHANGES.
Findings are anchored inline: #1 ci-frontend-lint.yaml, #2 and #3 check-v1.mjs, #4 tsconfig.json, #5 the fixtures, #6 vite.config.js, #7 manifest.ts, #8 README, #9 check-v1.mjs:98, #10 check-v1.mjs:440, #11 biome.json.
Unanchored findings
frontend/scripts/check-v1.mjsandfrontend/test-fixtures/**/*.mjs(about 900 lines) sit outside bothbiome ci src/and the pre-commit hook (files: ^frontend/src/), so the guard tooling is not linted in CI. Run locally at Biome 2.3.13,scripts/is clean; thetest-fixturesdiagnostics (1 error, 3 warnings) are the intentional bad cases.
Lens checklist
- Lens 1: See findings #3, #8 (the "reliable parser" claim in the description does not hold for interpolated or glob specifiers; the README enforcement claim)
- Lens 2: Clean
- Lens 3: See findings #2, #3, #10, #11
- Lens 4: Clean (security scan: no findings at confidence >= 8)
- Lens 5: N/A — no migrations or schema
- Lens 6: N/A — no concurrency primitives
- Lens 7: See findings #4, #6, #7, #11
- Lens 8: N/A — build and lint tooling, no runtime or external calls
- Lens 9: Clean (
check-v1scans 24 MB of the realsrc/in about 0.2 s) - Lens 10: N/A — no runtime paths
- Lens 11: See findings #1
- Lens 12: N/A — no prompt, model or agent surface
- Lens 13: See findings #1, #5
- Lens 14: Clean (no dependency added, lockfile untouched)
- Lens 15: Clean (Biome 2.3.13
biome ci src/: 0 errors, 799 warnings and 127 infos, identical to the baseline in the description) - Lens 16: See findings #8, #9
- Lens 17: Clean (merge-result sweep against
origin/mainfound no stale restatement outside this PR's files) - Lens 18: N/A — no chart, values, on-prem build or docs surface
- Lens 19: N/A — no scaling or lifecycle values
Open questions for the author
- Is
bun run test:v1meant to join thetypecheckjob in this PR, or is the deferral to UN-4363 firm? (#1) - Is the interpolated-import form (
import(`./pages/${name}`)) allowed in v1 and in v0's@v1mount? If so the checker needs to read it; if not it needs to reject it. (#3) - Does the cloud
manifest.tsoverlay constrain the shape that UN-4285 and UN-4286 will define? (#7)
Assumptions made
- APPROVE WITH NITS assumes the Medium findings can land as follow-ups, given v1 holds no code yet.
athul-rs
added a commit
that referenced
this pull request
Oct 28, 2026
2 of 3 tasks
vishnuszipstack
left a comment
vishnuszipstack
left a comment
Contributor
There was a problem hiding this comment.
Review with the PR Review Toolkit (code reviewer, silent-failure hunter, type design, test coverage, comment accuracy, simplifier). I left out anything already raised in this PR's threads; where a comment builds on an existing one (#2, #3, #5, #9), it says so. Each [Medium] bypass, and the Biome false positive, was reproduced on a scratch copy of the tree.
…ins in the ignore check (UN-4270) Reported by Athul while building #2394/#2395 on this branch: - src/v1/tsconfig.json: skipLibCheck back on, as in tsconfig.json. Off, tsc also checked third-party typings, and @types/chai 5.2.3 and vitest 3.2.6 both declare `containSubset`, so any v1 test importing vitest failed typecheck:v1. A fixture test now imports vitest to keep this covered. - The "no v1 file is git-ignored" check skips src/v1/plugins/<id>/: cloud plugins copied there are git-ignored on purpose (#2394).
From Ali's, Athul's and Vishnu's reviews on #2380, each reproduced first: - tsconfig/jsconfig map `@v1/*` too: once v0 imports the entry, tsc follows it into src/v1, where cross-folder `@v1/...` imports must resolve (would break `bun run typecheck` at UN-4271). Lint keeps v0 on `@v1`. - check-v1 scanner tracks regex literals, so a `/*` inside one no longer hides the rest of the file; CSS blanks only block comments (`//` is URL). - Imports are read by argument: literals, templates by their fixed prefix, every glob pattern, and new URL(..., import.meta.url). In v1 an import the guard cannot read fails closed; a computed v1 path is allowed only in folders that may import everything (the lazy route registry in app/). - v0 specifiers are normalized before the check (/src/v1, @/./v1, ...). - v1 styles may not @import antd packages. - Biome's v0 rule matches only alias spellings of src/v1, so `uuid/v1` or `@/api/v1` no longer fail; relative spellings are check-v1's (in CI). - CI runs `test:v1` (fixtures) on Node 22; setup-biome pinned to a SHA (SonarCloud S7637). Fixtures assert the exact message and count per form, cover each fix, and check that comments, JSX and look-alike class names stay quiet. Against the previous guard, 9 of the new tests fail.
Static files (icons, logos, images) get one home, src/v1/assets/, imported as @v1/assets/... (e.g. logo.svg?react). v1 cannot reach v0's src/assets/. Raised by Ali on #2380: Lane I's P1-T22 needs SVG icons and provider logos, and a .svg outside the taxonomy failed --files. components/ui, components/patterns, shell, pages and styles may import assets/; assets/ imports nothing but itself (room for a generated icon map). README, contract table and the fixture matrix updated together; v1-allowed-imports imports an SVG as a React component under typecheck:v1.
jaseemjaskp
requested review from
athul-rs,
muhammad-ali-e and
vishnuszipstack
October 29, 2026 10:04
…r (UN-4270) Greptile P1 on #2380, introduced by 3889a34: the code-only view blanked string bodies, and the apostrophe in JSX text (`It's`, `{name}'s`) opened a "string" that ran to the end of the line, hiding a forbidden import after it. A quote now starts a string only in expression position (after an operator, `(`, `,`, `=`, `[`, `{` or a keyword such as return/case/import/from) and only if it closes on its line, as JS strings must; otherwise it is JSX text. The regex-literal check shares the same token test. Fixtures: Greptile's example and the `{name}'s` form each must report the import; a JSX attribute and quoted JSX text stay quiet. A differential run over the 409 real v0 files finds no real import hidden by the code-only view.
athul-rs
reviewed
Oct 29, 2026
athul-rs
left a comment
athul-rs
left a comment
Contributor
There was a problem hiding this comment.
PR review — #2380 (FOLLOWUP round 2, head feb8c5f, previous review at 81e15b1)
Reviewed under unstract v2.1.73 (review-deep, 19-lens rubric) with review-lean round-2 scoping: the agents read only the delta 81e15b159..feb8c5f1d (5 commits, 20 files, +550/-164); prior findings verified by the orchestrator.
Signals: lenses 3, 13, 16, 17 raised (check-v1.mjs scanner rewrite; boundary fixtures; README; biome.json pattern change). Security deep-dive: unavailable — skipped by review-lean, the delta is lint and boundary tooling with no runtime attack surface; the round-1 deep-dive on this PR found nothing.
Verdict — REQUEST CHANGES
Summary — Critical: 0 · High: 1 · Medium: 7 · Low: 5 · Lenses run: 19/19 (delta-scoped; 9 N/A for tooling-only scope). Open = 8 new (#12–#19: 1 High, 5 Medium, 2 Low) + 5 from round 1 not fully resolved (#3, #5 Medium; #7, #9, #10 Low).
Prior context fetched — followup-fetch.sh exited 0 (6 of 6 files). Since-boundary: 81e15b159.
Scope change: NO — the delta stays inside the round-1 surface (check-v1.mjs, fixtures, CI step, tsconfig/biome/vite config, README).
Most round-1 fixes landed and hold. CI is green on the head and runs the fixture suite (46/46). The blocker is new in this round: the scanner's comment/string/regex state can be flipped by ordinary JSX text, after which check-v1 prints ok for the rest of the file (#12). Three rounds of point fixes to the hand-written tokenizer have each opened a new hole of the same class, so the request is the structural fix, or at minimum the end-of-file tripwire, not another patch.
Prior findings status
- #1 [Medium]
test:v1run by no CI job → RESOLVED —.github/workflows/ci-frontend-lint.yaml:147runsbun run test:v1; the typecheck job onfeb8c5f1dlogs# tests 46 # pass 46 # fail 0. - #2 [Medium]
/*inside a regex literal read as a comment → RESOLVED — regex literals are tracked (check-v1.mjs:236, fixtureafterSlashStarInRegexatboundary.fixtures.mjs:346); re-ranconst r = /\/*$/;followed by a v0 import on a scratch copy and it is reported. The wider desync class is #12. - #3 [Medium] import forms the checker cannot read pass silently → PARTIALLY RESOLVED — a variable or fully computed specifier now fails closed (fixture
v1-unanalysable-import; re-ranimport(n), reported). Residual:import("./" + x), a template prefix without a trailing slash, and escaped specifiers still pass silently — see #13. - #4 [Medium] root
tscfails once v0 imports@v1→ RESOLVED —@v1/*mapped intsconfig.json:47-48andjsconfig.json:6-7; fixturev0-tsx-imports-entry; CI typecheck green. - #5 [Medium] five guard branches have no fixture → PARTIALLY RESOLVED — items 3 and 4 covered (
variant-outline tenant-id, commented-out imports). Items 1 (@ant-design/icons), 2 (a v1 package import such asreact) and 5 (@/undertypecheck:v1) still have no fixture; deleting those branches leaves the suite green. Author deferred them; no waiver recorded. - #6 [Low] Vite resolves
@v1/<anything>→ RESOLVED —vite.config.js:188now says only lint keeps v0 to the entry. - #7 [Low]
pluginManifesttypedreadonly []→ NOT RESOLVED —src/v1/plugins/manifest.tsis unchanged; the author defers it to #2394, which replaces the file. Residual failure mode unchanged (no declared contract for the cloud overlay until #2394 lands). - #8 [Low] README claims only enforced rules → RESOLVED —
src/v1/README.md:4-6rewritten. The new wording overclaims in a different way; see #18. - #9 [Low] six comments misstate the code → PARTIALLY RESOLVED — five of six fixed. Residual: a
../that stays inside v1 is still reported as "outside src/v1" (check-v1.mjs:375-380); the author keeps the ban by design and defers the message wording. - #10 [Low] v0 symlinks skipped silently → NOT RESOLVED —
check-v1.mjs:682still skips a symlink under v0; the author agreed and deferred it. No waiver recorded. - #11 [Low] v0 rule flags third-party
v1segments → RESOLVED —biome.json:190-199is anchored to alias spellings. The narrowing dropped relative spellings from Biome; see #15.
New findings
Findings #12–#19 are posted inline below. Replies on the round-1 threads carry each prior finding's status.
Unanchored findings
None.
Lens checklist
- Requirements & intent — Clean (code-reviewer; title follows
UN-XXXX [TYPE] title) - Correctness — See findings #19
- Error handling — See findings #12, #13, #14, #16
- Security — N/A — delta is lint and boundary tooling; security deep-dive skipped (review-lean), round-1 scan found nothing
- Data integrity & migrations — N/A — no data or schema
- Concurrency — N/A — no concurrency primitives
- API compatibility — N/A — no public API; the tool boundary with cloud is under #15
- Reliability & resilience — N/A — build tooling, no runtime calls
- Performance & cost — Clean (code-reviewer)
- Observability — N/A — no runtime paths
- Operational safety — Clean (orchestrator: CI step order and exit codes checked; the two new action SHAs match their tags)
- LLM/agent — N/A — not touched
- Testing — See findings #17
- Dependencies & build — Clean (orchestrator: no dependency change)
- Code quality — Clean (orchestrator: pinned Biome ran in CI on the head,
biome-cipassed and the lint report shows no issues) - Documentation accuracy — See findings #18
- Cross-document consistency — See findings #15
- On-prem deployment experience — N/A — no chart, on-prem values or on-prem build path touched
- Scaling & workload lifecycle — N/A — no workload values
Open questions for the author
- Will cloud CI run
check-v1 --codeoversrc/plugins/, or should Biome keep relative spellings? (#15) - Is #7 still expected to close with #2394, and #10 and #5's remaining items tracked somewhere?
Assumptions made
{done}/{total}and/*in JSX text are treated as likely in real v1 pages. If v1 were never to hold JSX text, #12 would drop to Medium.
cost: agents=4 wall=12:41 lenses-examined=10 of 19 files=20 lines=714
…okenizer (UN-4270)
The tokenizer guessed regex vs division and string vs JSX text from the
previous token. JSX text such as `{done}/{total}`, `files/*.ts`, a stray
backtick, a multi-line attribute or `range.from / x` flipped it into a state
that never closed, and check-v1 then passed the rest of the file.
check-v1 now parses with @babel/parser (TypeScript 7 has no JS parser API)
and walks import/export statements, import(), require(), import.meta.glob,
new URL(..., import.meta.url), `typeof import()` and `import x = require()`.
Escapes are decoded and methods named import/require are not imports.
Also:
- v1 rejects concatenated specifiers, template prefixes with no `/`, a
non-literal new URL and files that do not parse; v0 reports a template or
glob prefix that could grow into src/v1.
- v1 styles: check @reference/@plugin/@config and the folder contract.
- Biome: relative v0 -> v1 spellings are back in the v0 group.
- A `../` that stays inside v1 is reported as such, not as "outside".
- v0 symlinks into src/v1 are reported; file symlinks are scanned.
- README: name what Biome covers; asset placement is convention.
- Fixtures for every reported repro and the remaining #5 branches; drop the
unused src/v1/plugins/<id> ignore filter.
…s (UN-4270)
`import("@v1/x" as const)` wrapped the specifier in a TSAsExpression, so
v0 treated it as opaque and skipped it. Unwrap as/satisfies/<T>x/x! first.
Also share one helper between the outside and climbs tests.
vishnuszipstack
left a comment
vishnuszipstack
left a comment
Contributor
There was a problem hiding this comment.
Follow-up review at daa1a18. Every finding below was reproduced on a scratch tree (check-v1: ok on input that Vite bundles across the boundary), and none duplicates an earlier thread.
- import.meta.glob: resolve relative patterns against the `base` option; a
base that is not a string literal is unanalysable in v1.
- See through TypeScript wrappers on the import markers too
(`import.meta.url as string`, `(import.meta.glob as any)()`,
`(require as ...)()`, `new (URL as ...)()`, a glob array `as const`).
- v0: judge a `+` chain or `.concat()` by the string it starts with, as
Vite globs it, and a conditional by both branches.
- Report a `..` after the first wildcard or `${}`, in v1 and v0.
- Read stylesheets with PostCSS: case-insensitive at-rules, `@import"x"`
with no space, and `/*` inside strings.
- Root tsconfig includes vite-env.d.ts, so v1 assets, CSS and
import.meta.env reached through `@v1` typecheck in the v0 lane.
- Parse .d.ts files in declaration mode.
- Split nodes() and checkV0() into helpers (Sonar S3776).
vishnuszipstack
self-requested a review
November 2, 2026 07:02
vishnuszipstack
left a comment
vishnuszipstack
left a comment
Contributor
There was a problem hiding this comment.
Re-review of 73b9ffa. The six findings from my last pass (glob base, TS wrappers on the import markers, v0 concatenation, .. after a wildcard, @import"x", vite-env.d.ts) are fixed. I re-ran each repro and each is now reported. Two new High findings remain, both in forms check-v1 never reads; the inline comments have repros and fixes. Medium items are left out of this pass.
…file types (UN-4270) - CSS values are read with postcss-value-parser: every url() and every string inside image-set() / -webkit-image-set(), with backslash escapes removed as Vite does. @import targets use the same parser; an unreadable one is unanalysable. - v1 holds only code, styles (in styles/), Vite asset types, JSON, Markdown and .gitkeep; --files reports any other type. STYLE_EXT and the Biome component-style group cover all of Vite's CSS languages. - --code reads every non-asset file in v1 and v0, parsing an unknown type as code (Vite can load any file with ?worker), and checks style files wherever they are, so --code alone catches what --files rejects.
vishnuszipstack
approved these changes
Nov 2, 2026
The cloud build copies its v1 plugins into src/v1/plugins/<id>/, as it does v0 plugins into src/plugins/. Ignore them, keeping only the public placeholder manifest.ts tracked, so cloud code cannot be committed here. The ignore-check test skips a local copy of those plugins, and a new test asserts they are ignored while manifest.ts is not.
Contributor
Frontend Lint Report (Biome)✅ All checks passed! No linting or formatting issues found. |
|
Contributor
Unstract test resultsPer-group results
Critical paths
|
praveen-formido
added a commit
that referenced
this pull request
Nov 2, 2026
…fold has landed The scaffold PR #2380 was squash-merged, so every scaffold file both sides touched conflicted. Main's version is taken for the import checker, the tsconfigs, the CI workflow and the Biome boundary groups; this branch keeps the styling policy script and its lint:v1 and tokens:v1 entries, the tokens.css formatter override, the Styling section of the v1 README and the 27 styling fixture rows. Two fixtures main re-added or changed carried an unscoped :root selector with a raw hex, which the styling policy rejects, so they reference a token under .v1 instead. Co-Authored-By: Claude Fable 5.1 <[email protected]>
athul-rs
added a commit
that referenced
this pull request
Nov 3, 2026
Keep the plugin loader layout (plugins/index.ts + .gitkeep) in .gitignore and point the scaffold's ignore and symlink fixtures at it, since this branch replaced the placeholder plugins/manifest.ts. Co-Authored-By: Claude Opus 5.5 <[email protected]>
athul-rs
added a commit
that referenced
this pull request
Nov 3, 2026
main's copies of the scaffold files equal the scaffold head merged in the previous commit, so the conflicts keep this branch's side. The placeholder plugins/manifest.ts that main reintroduces stays removed: plugins/index.ts replaces it. Co-Authored-By: Claude Opus 5.5 <[email protected]>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
frontend/src/v1/, an isolated tree for the v1 frontend. It has the folder taxonomy from the spec plus anassets/folder for static files, a public entry (src/v1/index.ts, which v0 imports as@v1) and a placeholder for the cloud overlay.src/v1/tsconfig.json).src/v1.@v1entry.prop-types, antd, component.css, explicitany.ant-class prefix, PropTypes assignments, component style files outsidestyles/.components/ui/may import onlylib/andstyles/, and pages may not import other pages,app/,test/(except from test files) or the v1 entry.typecheck:v1,lint:v1andtest:v1scripts, and runstypecheck:v1andlint:v1in the frontend CItypecheckjob. v0 scripts are unchanged.Why
How
../,@/,src/), v0 ↛ v1 except@v1noRestrictedImportsin twooverridesinbiome.json. This already runs inbiome ci src/, pre-commit and editorsprop-types,antd,@ant-design/*, component.css, explicitanysrc/v1/**override@v1aliassrc/v1/tsconfig.json. It lives in the folder so editors pick it up; v0tsconfig.json/jsconfig.jsonexcludesrc/v1styles/, folder taxonomyscripts/check-v1.mjs --files, run first bytypecheck:v1ant-class prefixscripts/check-v1.mjs --code, run bylint:v1Why some rules live in a script rather than Biome:
v1-base-rules-still-applyfixture checks that the shared rules still apply under v1.allowJs: falsemakestscskip a.jsfile rather than fail on it. That is why the--filescheck runs beforetsc../,@v1/, or a package name.ant-check is temporary. Remove it once the shadcn migration completes.test-fixtures/v1-boundary/holds one case per rule. For each case, the runner builds a throwaway frontend tree, overlays the case's files and runs the realpackage.jsonscripts. It usesnode --test, andtest-fixtures/**is excluded from vitest, so the v0 suite never collects it.Can this PR break any existing features. If yes, please list possible items. If no, please explain why. (PS: Admins do not merge the PR without this section filled)
src/v1, and nothing in v0 imports from there today.src/plugins/tree, which cloud CI lints with thisbiome.json. That is intended: cloud v0 plugins must not reach into v1 either.Database Migrations
Env Config
Relevant Docs
frontend/src/v1/README.md: the rules, the folder contract table and the scripts.Related Issues or PRs
typecheck:v1,lint:v1andtest:v1all run in the frontend CItypecheckjob; the remaining CI gates are P0-T10 (UN-4363).Dependencies Versions
Notes on Testing
v0 no-regression, before (
main@ 6e14fa7) and after:vite build: sha256 of every non-sourcemap output file (239 files)76925c7e…bun run test --runbun run typecheckbiome ci src/v1:
bun run typecheck:v1andbun run lint:v1pass on the empty scaffold.bun run test:v1passes 46/46, in CI as well: one case per rule, generated tests over every folder pair, import form and outside spelling (v1 and v0), plus checks that every case folder is listed and that no v1 or fixture file is git-ignored.Two other fixes the fixtures turned up:
.gitignore: the root Pythonlib/rule matchedsrc/v1/lib/, so its files would never have been committed. The PR re-includes it, and a test guards against it happening again.check-jsonpre-commit hook: the exclude now coversfrontend/src/v1/tsconfig.json, which contains comments, matching what the hook already does forfrontend/tsconfig.json.Screenshots
N/A
Checklist
I have read and understood the Contribution Guidelines.