@vzn/vx 0.0.21 → 0.0.23

package/src/workspace/workspace.ts

@@ -2,7 +2,8 @@ import type { Dirent } from 'node:fs'
 import { readdir, stat } from 'node:fs/promises'
 import path from 'node:path'
 import type { ProjectConfig, WorkspaceConfig } from '../config.js'
-import { relPosix, UserError, normalizeGlob } from '../util/index.js'
+import { BUN_GLOB_WILDCARDS, relPosix, UserError, normalizeBunGlob } from '../util/index.js'
+import { type LoadReads, readOnce } from './load-reads.js'
 
 export interface PackageJson {
   name: string
@@ -38,7 +39,14 @@ export interface ProjectEntry {
   config: ProjectConfig
 }
 
-const CONFIG_FILENAMES = ['vx.config.ts', 'vx.config.mts', 'vx.config.js', 'vx.config.mjs']
+export const PROJECT_CONFIG_FILENAMES = [
+  'vx.config.ts',
+  'vx.config.mts',
+  'vx.config.js',
+  'vx.config.mjs',
+]
+
+const decoder = new TextDecoder()
 
 /**
  * Walk up from `start` to find the workspace root. A directory is a root
@@ -56,15 +64,20 @@ const CONFIG_FILENAMES = ['vx.config.ts', 'vx.config.mts', 'vx.config.js', 'vx.c
  * When no candidate claims `start` — a standalone package, or a subdirectory
  * of a single-project repo — the nearest candidate wins (the root itself IS
  * the project). Throws a `UserError` when there is no candidate before `/`.
+ * A load that goes on to `loadWorkspace` passes its `reads`, so the
+ * root's manifest is read once for both.
  */
-export async function findWorkspaceRoot(start: string): Promise<string> {
+export async function findWorkspaceRoot(
+  start: string,
+  reads: LoadReads = new Map(),
+): Promise<string> {
   let dir = path.resolve(start)
   const below: string[] = []
   let nearest: string | null = null
   while (true) {
     let globs: string[] | null
     try {
-      globs = await readPackageGlobs(dir)
+      globs = await readPackageGlobs(dir, reads)
     } catch {
       // An unparseable manifest is still a root SIGNAL (the pre-existing
       // behaviour probed only for existence); it just can't claim members.
@@ -74,10 +87,19 @@ export async function findWorkspaceRoot(start: string): Promise<string> {
     if (globs !== null) {
       nearest ??= dir
       if (claimsMember(dir, below, globs)) return dir
+      // pnpm takes the nearest `pnpm-workspace.yaml` as the root, listed or
+      // not. Walking past it, `apps/inner` resolved to the outer workspace
+      // while `apps/inner/pkgs/x` resolved to the inner one: two roots and
+      // two caches for one tree (item 990). Already read: `reads` holds it.
+      if ((await readOnce(reads, path.join(dir, 'pnpm-workspace.yaml'))) !== null) return dir
     }
     const parent = path.dirname(dir)
     if (parent === dir) break
-    below.push(dir)
+    // Only a directory holding a manifest can be a member: a glob that
+    // matched `packages/tools`, which has none, claimed the standalone
+    // package below it, and `packages/tools/standalone` ran as a stranger
+    // in a workspace that does not list it (item 989).
+    if (globs !== null) below.push(dir)
     dir = parent
   }
   if (nearest !== null) return nearest
@@ -117,7 +139,11 @@ function splitPackageGlobs(globs: readonly string[]): { positive: string[]; nega
   // `!./packages/legacy` and `!packages//legacy` excluded nothing, silently
   // (2026-09-10): the same spellings the input globs normalize.
   for (const raw of globs) {
-    const g = normalizeGlob(raw)
+    // A member glob names directories, so a trailing slash says nothing:
+    // npm and pnpm read `packages/*/` as `packages/*`. `normalizeBunGlob`
+    // gives it a task glob's meaning, `/**`, and `packages/*/` found every
+    // example and fixture package at any depth below the members (item 985).
+    const g = normalizeBunGlob(raw.replace(/(.)\/+$/, '$1'))
     if (g.startsWith('!')) negative.push(g.slice(1).replace(/\/+$/, ''))
     else positive.push(g)
   }
@@ -128,7 +154,13 @@ function excludedBy(rel: string, negative: readonly string[]): boolean {
   for (const neg of negative) {
     if (neg.length === 0) continue
     if (rel === neg || rel.startsWith(`${neg}/`)) return true
-    if (/[*?[\]{}]/.test(neg) && new Bun.Glob(neg).match(rel)) return true
+    if (!BUN_GLOB_WILDCARDS.test(neg)) continue
+    if (new Bun.Glob(neg).match(rel)) return true
+    // pnpm matches the MANIFEST (`<pattern>/package.json`), so its documented
+    // `!**/test/**` excludes `packages/test` itself; matched against the
+    // directory, `**/test/**` needs something below `test` and the package
+    // stayed a member (item 986).
+    if (new Bun.Glob(`${neg}/package.json`).match(`${rel}/package.json`)) return true
   }
   return false
 }
@@ -138,16 +170,29 @@ function excludedBy(rel: string, negative: readonly string[]): boolean {
  * candidate. A bare `package.json` (no `workspaces`) is single-project mode:
  * the root itself is the only project, hence `['.']`.
  */
-async function readPackageGlobs(dir: string): Promise<string[] | null> {
+async function readPackageGlobs(dir: string, reads?: LoadReads): Promise<string[] | null> {
   const yamlPath = path.join(dir, 'pnpm-workspace.yaml')
-  if (await Bun.file(yamlPath).exists()) {
-    const parsed = (parseManifest(await Bun.file(yamlPath).text(), yamlPath, Bun.YAML.parse) ??
-      {}) as { packages?: unknown }
-    return assertGlobList(parsed.packages ?? [], yamlPath, 'packages')
+  const yaml = await readOnce(reads, yamlPath)
+  let yamlWithoutPackages = false
+  if (yaml !== null) {
+    const parsed = parseManifest(decoder.decode(yaml), yamlPath, Bun.YAML.parse)
+    if (parsed !== null && parsed !== undefined) {
+      if (typeof parsed !== 'object' || Array.isArray(parsed)) {
+        throw new UserError(`${yamlPath}: must be a mapping (\`packages:\` and pnpm's settings)`)
+      }
+      const packages = (parsed as { packages?: unknown }).packages
+      if (packages !== undefined) return assertGlobList(packages, yamlPath, 'packages')
+    }
+    // No `packages:`: pnpm 10 keeps its settings and catalogs in this file
+    // for a single-package repo too. The root's package.json decides, as it
+    // would without the file; read as an empty list, the root found zero
+    // projects and every verb ran nothing and exited 0 (item 984).
+    yamlWithoutPackages = true
   }
   const pkgPath = path.join(dir, 'package.json')
-  if (!(await Bun.file(pkgPath).exists())) return null
-  const pkg = parseManifest(await Bun.file(pkgPath).text(), pkgPath, JSON.parse) as PackageJson
+  const pkgBytes = await readOnce(reads, pkgPath)
+  if (pkgBytes === null) return yamlWithoutPackages ? [] : null
+  const pkg = parsePackageJson(decoder.decode(pkgBytes), pkgPath)
   const ws = pkg.workspaces as unknown
   if (ws === undefined || ws === null) return ['.']
   if (ws && typeof ws === 'object' && !Array.isArray(ws) && 'packages' in ws) {
@@ -160,6 +205,49 @@ async function readPackageGlobs(dir: string): Promise<string[] | null> {
   return assertGlobList(ws, pkgPath, 'workspaces')
 }
 
+/**
+ * Package manifests the globs never reach, in single-project mode only: a
+ * root `package.json` with no `workspaces` field makes the root the one
+ * project, and a `packages/app/package.json` full of scripts beside it is
+ * then invisible — `vx init` said "no package.json scripts" and `vx run`
+ * said "run vx init" (2026-09-16). One shallow scan, two levels, on the
+ * failure path only; `node_modules` and dot directories are skipped.
+ */
+export async function unreachedPackages(workspace: Workspace): Promise<string[]> {
+  if (workspace.packageGlobs.length !== 1 || workspace.packageGlobs[0] !== '.') return []
+  const found: string[] = []
+  // Two scans: Bun.Glob does not expand a brace whose alternatives hold a
+  // slash (`{*,*/*}/package.json` matched nothing, measured 2026-09-16).
+  for (const pattern of ['*/package.json', '*/*/package.json']) {
+    for await (const rel of new Bun.Glob(pattern).scan({
+      cwd: workspace.root,
+      onlyFiles: true,
+      dot: false,
+    })) {
+      const posix = rel.split(path.sep).join('/')
+      if (posix.startsWith('node_modules/') || posix.includes('/node_modules/')) continue
+      found.push(posix.slice(0, -'/package.json'.length))
+    }
+  }
+  return found.sort()
+}
+
+/** The line for `unreachedPackages`' finding: the cause, the packages, the glob to add. */
+export function unreachedHint(unreached: readonly string[]): string {
+  const n = unreached.length
+  const shown = unreached.slice(0, 3).join(', ') + (n > 3 ? ` and ${n - 3} more` : '')
+  const globs = [
+    ...new Set(unreached.map((d) => (d.includes('/') ? `${path.posix.dirname(d)}/*` : d))),
+  ]
+    .map((g) => `"${g}"`)
+    .join(', ')
+  return (
+    `package.json declares no \`workspaces\` (and there is no pnpm-workspace.yaml), so the root is the only project ` +
+    `and ${n} package.json below it ${n === 1 ? 'is' : 'are'} not: ${shown}. ` +
+    `Add \`"workspaces": [${globs}]\` to package.json and re-run.`
+  )
+}
+
 /**
  * Parse a workspace manifest, naming the FILE on failure. The raw parser
  * errors (`Failed to parse JSON`) carry no path, so in a 1000-package
@@ -174,6 +262,25 @@ function parseManifest(text: string, file: string, parse: (t: string) => unknown
   }
 }
 
+/**
+ * A `package.json` as discovery reads it: a JSON object whose `name`, when
+ * present, is a string with no surrounding whitespace (npm refuses one).
+ * Parsed and cast unchecked, `null` crashed with a TypeError and a stack,
+ * `{"name":123}` planned `123#build`, and `" a"` beside `"a"` made two
+ * projects (item 988).
+ */
+function parsePackageJson(text: string, file: string): PackageJson {
+  const pkg = parseManifest(text, file, JSON.parse)
+  if (pkg === null || typeof pkg !== 'object' || Array.isArray(pkg)) {
+    throw new UserError(`${file}: must be a JSON object`)
+  }
+  const name = (pkg as { name?: unknown }).name
+  if (name !== undefined && (typeof name !== 'string' || name.trim() !== name)) {
+    throw new UserError(`${file}: "name" must be a string with no surrounding whitespace`)
+  }
+  return pkg as PackageJson
+}
+
 /**
  * A glob list is user input from a manifest: `packages: "packages/*"`
  * (a bare string instead of a list) is an easy YAML slip, and without
@@ -183,9 +290,46 @@ function assertGlobList(value: unknown, file: string, field: string): string[] {
   if (!Array.isArray(value) || value.some((p) => typeof p !== 'string')) {
     throw new UserError(`${file}: \`${field}\` must be an array of glob strings`)
   }
+  for (const pattern of value as string[]) {
+    if (EXTGLOB.test(pattern)) throw extglobRefusal(pattern, file, field)
+  }
   return value as string[]
 }
 
+/**
+ * An extglob group: `!(…)`, `@(…)`, `+(…)`, `*(…)`, `?(…)`. npm (minimatch)
+ * and yarn (micromatch) read `packages/!(x)` as "every directory in
+ * packages/ but x"; `Bun.Glob` has no extglob, and its scan read the
+ * segment as a wildcard, so the excluded package became a project
+ * (turborepo#3766) while its match read the same text literally. vx cannot
+ * honour the pattern, so it refuses it rather than read it either way.
+ */
+const EXTGLOB = /[!@+*?]\(/
+
+function extglobRefusal(pattern: string, file: string, field: string): UserError {
+  const head = `${file}: \`${field}\` entry "${pattern}" is an extglob, which vx's glob engine does not read`
+  // The common shape, a whole segment excluding plain names, has an exact
+  // rewrite in the `!` entries every package manager takes.
+  const segments = pattern.split('/')
+  const at = segments.findIndex((s) => /^!\([^()]*\)$/.test(s))
+  const names = at === -1 ? [] : segments[at]!.slice(2, -1).split('|')
+  const plain = names.every((n) => n !== '' && !BUN_GLOB_WILDCARDS.test(n) && !/[()!@+]/.test(n))
+  const rest = segments.filter((_s, i) => i !== at).join('/')
+  if (at === -1 || !plain || EXTGLOB.test(rest)) {
+    return new UserError(
+      `${head}. List the members with \`*\`, braces or \`!\` exclusions instead.`,
+    )
+  }
+  const before = segments.slice(0, at)
+  const rewrite = [
+    [...before, '*', ...segments.slice(at + 1)].join('/'),
+    ...names.map((n) => `!${[...before, n].join('/')}`),
+  ]
+  return new UserError(
+    `${head} (npm and yarn read \`!(…)\` as an exclusion). List the exclusion as its own entry: [${rewrite.map((r) => JSON.stringify(r)).join(', ')}].`,
+  )
+}
+
 /**
  * Resolve the cache directory for a workspace. Respects the user's
  * `defineWorkspace({ cacheDir })` override (relative to the workspace
@@ -206,8 +350,8 @@ export function resolveCacheDir(root: string, config: WorkspaceConfig | null): s
  * If a `package.json` exists with no `workspaces` field, the root
  * itself is treated as a single-project workspace.
  */
-export async function loadWorkspace(root: string): Promise<Workspace> {
-  const packageGlobs = await readPackageGlobs(root)
+export async function loadWorkspace(root: string, reads?: LoadReads): Promise<Workspace> {
+  const packageGlobs = await readPackageGlobs(root, reads)
   if (packageGlobs === null) {
     // Should be unreachable: findWorkspaceRoot only returns dirs that
     // pass at least one of the two existence checks.
@@ -242,8 +386,9 @@ export function memberBaseDirs(workspace: Workspace): string[] {
  * The directories a workspace glob names. For the `<dir>/*` shape this is
  * one readdir of `<dir>` — the same answer `Bun.Glob` gives, at a third of
  * the cost (measured 2026-09-02: 25 ms → ~2 ms for 1000 members). A
- * symlinked member is followed when it points at a directory, as the glob
- * did. Dot-directories are skipped (`dot: false`), so are `node_modules`.
+ * symlinked member is followed when it points at a directory, and so is
+ * it by the scan any other shape takes, unless the glob holds `**` (item
+ * 987). Dot-directories are skipped (`dot: false`), so are `node_modules`.
  */
 async function memberDirs(root: string, pattern: string): Promise<string[]> {
   // Normalize: `"."` -> the root itself; `"foo/"` -> `"foo"`.
@@ -274,7 +419,12 @@ async function memberDirs(root: string, pattern: string): Promise<string[]> {
   const globPattern = normalized === '' ? 'package.json' : `${normalized}/package.json`
   const glob = new Bun.Glob(globPattern)
   const dirs: string[] = []
-  for await (const rel of glob.scan({ cwd: root, onlyFiles: true, dot: false })) {
+  // A linked member is a member, as the readdir path above has it:
+  // `packages/*` found `packages/b -> ../ext/b` while `packages/{a,b}` and
+  // `pack*/*` did not (item 987). Followed only where the depth is bounded:
+  // under `**` the scan would walk every pnpm `node_modules` link.
+  const followSymlinks = !normalized.includes('**')
+  for await (const rel of glob.scan({ cwd: root, onlyFiles: true, dot: false, followSymlinks })) {
     // Skip nested node_modules — workspace package globs shouldn't
     // ever reach into them, but a pathological pattern like `**`
     // would. Avoid splitting the path on the hot loop.
@@ -286,7 +436,7 @@ async function memberDirs(root: string, pattern: string): Promise<string[]> {
 }
 
 /**
- * The project's config file, by `CONFIG_FILENAMES` precedence, or null.
+ * The project's config file, by `PROJECT_CONFIG_FILENAMES` precedence, or null.
  *
  * Two shapes, one per platform, each the measured winner there across
  * 1,000 projects with the manifest read in flight alongside:
@@ -309,7 +459,7 @@ async function findConfigFile(dir: string): Promise<string | null> {
       return null
     }
     const byName = new Map(entries.map((e) => [e.name, e]))
-    for (const name of CONFIG_FILENAMES) {
+    for (const name of PROJECT_CONFIG_FILENAMES) {
       const entry = byName.get(name)
       if (entry === undefined) continue
       const candidate = path.join(dir, name)
@@ -318,7 +468,7 @@ async function findConfigFile(dir: string): Promise<string | null> {
     }
     return null
   }
-  for (const name of CONFIG_FILENAMES) {
+  for (const name of PROJECT_CONFIG_FILENAMES) {
     const candidate = path.join(dir, name)
     if (await isFile(candidate)) return candidate
   }
@@ -364,7 +514,7 @@ export async function listProjects(workspace: Workspace): Promise<ProjectMeta[]>
           .catch(() => null),
       ])
       if (text === null) return null
-      const pkg = parseManifest(text, pkgJsonPath, JSON.parse) as PackageJson
+      const pkg = parsePackageJson(text, pkgJsonPath)
       return { dir, pkg, configPath }
     }),
   )