@mjasnikovs/pi-task 0.40.30 → 0.40.32

package/dist/workers/docs-core.js

@@ -3,7 +3,7 @@ import * as fs from 'node:fs';
 import * as os from 'node:os';
 import * as path from 'node:path';
 import { openCache as defaultOpenCache } from './docs-cache.js';
-import { ensureIndexed as defaultEnsureIndexed } from './docs-index.js';
+import { collectFiles, ensureIndexed as defaultEnsureIndexed } from './docs-index.js';
 import { npmProfile, chooseEcosystem, defaultEcosystemIo, ECOSYSTEMS } from './docs-ecosystems.js';
 import { resolvePackage as defaultResolvePackage, ResolveError, resolveTypeSource, typesPackageName, splitRuntimeNamespace } from './docs-resolve.js';
 import { retrieveChunks as defaultRetrieveChunks, PACKAGE_RETRIEVE_LIMIT, RETRIEVE_CONTENT_BUDGET } from './docs-retrieve.js';
@@ -549,8 +549,12 @@ function docsRawCached(cache, pkg, profile, query, ensureIndexed, retrieveChunks
 }
 function docsRawUncached(pkg, profile, cacheError, autoInstalled) {
     const parts = [];
-    const surfaceFiles = walkSurfaceAlpha(pkg.root, profile);
-    const entryFirst = pkg.entry ? [pkg.entry, ...surfaceFiles.filter(f => f !== pkg.entry)] : surfaceFiles;
+    const surfaceFiles = collectFiles(pkg, profile).surface;
+    // Only an entry the rules KEPT goes first. A CJS-first package names its
+    // `.d.cts` in `types`, so the entry is exactly the twin `collectFiles` drops,
+    // and prepending it blind put it back at the head of the truncated blob.
+    const entry = pkg.entry !== null && surfaceFiles.includes(pkg.entry) ? pkg.entry : null;
+    const entryFirst = entry === null ? surfaceFiles : [entry, ...surfaceFiles.filter(f => f !== entry)];
     for (const abs of entryFirst) {
         let raw;
         try {
@@ -596,30 +600,6 @@ function docsRawUncached(pkg, profile, cacheError, autoInstalled) {
         autoInstalled: autoInstalled ? true : undefined
     };
 }
-function walkSurfaceAlpha(root, profile) {
-    const out = [];
-    const stack = [root];
-    while (stack.length) {
-        const dir = stack.pop();
-        let entries;
-        try {
-            entries = fs.readdirSync(dir, { withFileTypes: true });
-        }
-        catch {
-            continue;
-        }
-        for (const entry of entries) {
-            if (profile.skipDirs.includes(entry.name))
-                continue;
-            const full = path.join(dir, entry.name);
-            if (entry.isDirectory())
-                stack.push(full);
-            else if (entry.isFile() && profile.isSurfaceFile(entry.name))
-                out.push(full);
-        }
-    }
-    return out.sort();
-}
 function truncateHeadTail(s) {
     if (s.length <= NO_CACHE_TOTAL)
         return s;

package/dist/workers/docs-index.d.ts

@@ -7,6 +7,10 @@ export interface IndexResult {
     chunksWritten: number;
     contentHash: string;
 }
+export interface CollectedFiles {
+    surface: string[];
+    readme: string | null;
+}
 /**
  * The gate that decides whether a package needs re-indexing.
  *
@@ -48,4 +52,13 @@ export declare function chunkerFingerprint(): string;
  * `.d.cts` still has to be readable, and all 123 of zod's had a `.d.ts` twin.
  */
 export declare function dropParallelDeclarations(files: readonly string[]): string[];
+/**
+ * The files that make up a package's surface, after every selection rule.
+ *
+ * Exported because the uncached fallback assembles the same package by hand: a
+ * walk that skipped these rules put a Go module's whole subtree, and both halves
+ * of every parallel `.d.cts`, into the one degraded path where the budget is a
+ * single truncated blob.
+ */
+export declare function collectFiles(pkg: ResolvedPackage, profile: EcosystemProfile): CollectedFiles;
 export declare function ensureIndexed(cache: CacheHandle, pkg: ResolvedPackage, profile?: EcosystemProfile, supplements?: readonly ResolvedPackage[]): IndexResult;

package/dist/workers/docs-index.js

@@ -93,6 +93,9 @@ function computeContentHash(pkg, profile, supplements = []) {
 function walkSurface(root, profile) {
     const out = [];
     const stack = [root];
+    // A directory symlink is followed by its resolved target, so one pointing at
+    // an ancestor inside `root` walks the same subtree forever without this.
+    const walked = new Set([root]);
     while (stack.length) {
         const dir = stack.pop();
         let entries;
@@ -107,9 +110,11 @@ function walkSurface(root, profile) {
                 continue;
             const full = path.join(dir, entry.name);
             if (entry.isSymbolicLink()) {
+                let stat;
                 let realPath;
                 try {
                     realPath = fs.realpathSync(full);
+                    stat = fs.statSync(realPath);
                 }
                 catch {
                     continue;
@@ -117,15 +122,22 @@ function walkSurface(root, profile) {
                 const relReal = path.relative(root, realPath);
                 if (relReal.startsWith('..'))
                     continue;
-                const stat = fs.statSync(realPath);
-                if (stat.isDirectory())
+                if (stat.isDirectory()) {
+                    if (walked.has(realPath))
+                        continue;
+                    walked.add(realPath);
                     stack.push(realPath);
+                }
                 else if (stat.isFile() && profile.isSurfaceFile(realPath))
                     out.push(realPath);
                 continue;
             }
-            if (entry.isDirectory())
+            if (entry.isDirectory()) {
+                if (walked.has(full))
+                    continue;
+                walked.add(full);
                 stack.push(full);
+            }
             else if (entry.isFile() && profile.isSurfaceFile(entry.name))
                 out.push(full);
         }
@@ -177,7 +189,15 @@ function dropDeadMajors(files, root, version) {
     // own version is not shipping a dead major — it is shipping its API there.
     return kept.length > 0 ? kept : files;
 }
-function collectFiles(pkg, profile) {
+/**
+ * The files that make up a package's surface, after every selection rule.
+ *
+ * Exported because the uncached fallback assembles the same package by hand: a
+ * walk that skipped these rules put a Go module's whole subtree, and both halves
+ * of every parallel `.d.cts`, into the one degraded path where the budget is a
+ * single truncated blob.
+ */
+export function collectFiles(pkg, profile) {
     const walked = walkSurface(pkg.root, profile);
     const surface = dropDeadMajors(walked, pkg.root, pkg.version);
     return {

package/dist/workers/eco-go.d.ts

@@ -201,9 +201,11 @@ export declare function selectBuildVariants(files: readonly string[]): string[];
  * cannot fire here because it reads the major with `/^(\d+)\./` and Go spells its
  * versions `v1.12.0` and `go1.25.14`.
  *
- * A module whose root holds no Go files is not a package at all — the aws-sdk shape
- * — and keeps everything, for the same reason `dropDeadMajors` keeps a package whose
- * whole surface lives under one `vN/`.
+ * A root that declares nothing is not the API — the aws-sdk shape holds no Go files
+ * at all, and `cloud.google.com/go` holds a `doc.go` with a package clause and no
+ * imports while its whole surface lives in subdirectories. Either keeps everything,
+ * for the same reason `dropDeadMajors` keeps a package whose whole surface lives
+ * under one `vN/`.
  */
 export declare function selectOwnPackage(files: readonly string[], root: string, name: string, version: string): string[];
 /**

package/dist/workers/eco-go.js

@@ -25,7 +25,7 @@ import * as os from 'node:os';
 import * as path from 'node:path';
 import { ResolveError } from './docs-resolve.js';
 import { findAtOrAbove } from './eco-cargo.js';
-import { buildConstraint } from './go-surface.js';
+import { buildConstraint, splitGoItems } from './go-surface.js';
 import { readZip, readEntry, isUnsafeEntryName } from '../shared/zip.js';
 import { acquireStdlibPackage, findInGoroot, findSliced } from './go-stdlib.js';
 const PROXY = 'https://proxy.golang.org';
@@ -650,12 +650,25 @@ function holdsByDefault(src) {
         return true;
     return expr.split(/\s*&&\s*/).every(term => term.startsWith('!'));
 }
-/** The import paths a Go file's import block names. */
+/**
+ * The import paths a Go file's import block names.
+ *
+ * Scoped to the import DECLARATION, not the file: a sibling's import path also
+ * occurs as a plain string constant, in an error message and in a `go:generate`
+ * line, and reading those keeps the very subpackage `selectOwnPackage` exists to
+ * drop. A regex cannot draw that line — a column-0 `import (` also sits inside a
+ * generator template's raw string and inside a doc comment's example — so the
+ * split is the surface scanner's, which already skips comments and literals.
+ * An `ImportPath` is a `string_lit`, so the backtick form is legal Go and
+ * dropping it silently drops everything reachable only through it.
+ */
 function importsOf(src) {
     const out = [];
-    for (const m of src.matchAll(/"([^"\n]+)"/g)) {
-        if (/^[a-zA-Z0-9_.-]+(?:\/[a-zA-Z0-9_.~-]+)*$/.test(m[1]))
-            out.push(m[1]);
+    for (const item of splitGoItems(src)) {
+        if (!/^import\b/.test(item.text))
+            continue;
+        for (const m of item.text.matchAll(/"([^"\n]+)"|`([^`]+)`/g))
+            out.push(m[1] ?? m[2]);
     }
     return out;
 }
@@ -674,6 +687,17 @@ function goMajor(name, version) {
     const inVersion = /^(?:v|go)?(\d+)\./.exec(version);
     return inVersion ? Number(inVersion[1]) : 1;
 }
+/**
+ * Does this file declare anything a caller of the package could be handed?
+ *
+ * Read from the scanned declarations, not the raw text. `cloud.google.com/go`'s
+ * doc.go is one block comment, and a wrapped sentence beginning `type ` sits at
+ * column 0 — enough for a multiline regex to call the root the API and drop the
+ * subdirectories that hold all of it.
+ */
+function declaresApi(src) {
+    return splitGoItems(src).some(item => /^(?:func|type|var|const)\b/.test(item.text));
+}
 /**
  * Keep the package the caller asked for, and only the subpackages it can reach.
  *
@@ -701,9 +725,11 @@ function goMajor(name, version) {
  * cannot fire here because it reads the major with `/^(\d+)\./` and Go spells its
  * versions `v1.12.0` and `go1.25.14`.
  *
- * A module whose root holds no Go files is not a package at all — the aws-sdk shape
- * — and keeps everything, for the same reason `dropDeadMajors` keeps a package whose
- * whole surface lives under one `vN/`.
+ * A root that declares nothing is not the API — the aws-sdk shape holds no Go files
+ * at all, and `cloud.google.com/go` holds a `doc.go` with a package clause and no
+ * imports while its whole surface lives in subdirectories. Either keeps everything,
+ * for the same reason `dropDeadMajors` keeps a package whose whole surface lives
+ * under one `vN/`.
  */
 export function selectOwnPackage(files, root, name, version) {
     const rel = (f) => path.relative(root, f).replace(/\\/g, '/');
@@ -714,7 +740,16 @@ export function selectOwnPackage(files, root, name, version) {
     const byDir = new Map();
     for (const f of files)
         byDir.set(dirOf(f), [...(byDir.get(dirOf(f)) ?? []), f]);
-    if ((byDir.get('') ?? []).length === 0)
+    const sources = new Map();
+    const sourceOf = (f) => {
+        const seen = sources.get(f);
+        if (seen !== undefined)
+            return seen;
+        const text = safeRead(f) ?? '';
+        sources.set(f, text);
+        return text;
+    };
+    if (!(byDir.get('') ?? []).some(f => declaresApi(sourceOf(f))))
         return [...files];
     const major = goMajor(name, version);
     const reachable = new Set(['']);
@@ -722,7 +757,7 @@ export function selectOwnPackage(files, root, name, version) {
     while (queue.length > 0) {
         const dir = queue.shift();
         for (const f of byDir.get(dir) ?? []) {
-            for (const imp of importsOf(safeRead(f) ?? '')) {
+            for (const imp of importsOf(sourceOf(f))) {
                 if (!imp.startsWith(`${name}/`))
                     continue;
                 const sub = imp.slice(name.length + 1);
@@ -748,6 +783,7 @@ export function goContentFingerprintParts() {
         String(holdsByDefault),
         String(isWantedEntry),
         String(selectOwnPackage),
+        String(declaresApi),
         String(goMajor),
         String(importsOf)
     ];

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@mjasnikovs/pi-task",
-    "version": "0.40.30",
+    "version": "0.40.32",
     "description": "Deterministic task planning and spec-orchestration for local models — crash-safe /task pipelines with verify/enforce gates, a real-time remote web view, and web/docs/fetch/worker subagent tools.",
     "type": "module",
     "main": "./dist/index.js",