@wrongstack/tools 0.303.0 → 0.305.1

package/dist/codebase-index/codebase-index-tool.d.ts

@@ -6,6 +6,12 @@ interface CodebaseIndexInput {
 }
 interface CodebaseIndexOutput {
     filesIndexed: number;
+    fileOutcomes?: {
+        parsed: number;
+        skipped: number;
+        empty: number;
+        failed: number;
+    } | undefined;
     symbolsIndexed: number;
     langStats: Record<string, number>;
     durationMs: number;

package/dist/codebase-index/index.d.ts

@@ -31,6 +31,7 @@ export { internalKindToLspKind, lspKindToInternalKind, } from './lsp-kind.js';
 export type { ProjectIndexServerClientHealth, ProjectIndexServerConnectionState, ProjectIndexServerConnectionStatus, } from './project-server-client.js';
 export type { ProjectIndexServerActivity, ProjectIndexServerHealth, } from './project-server-protocol.js';
 export type { CallSite, CodeMapGraph, FileMeta, FileSymbols, GraphEdge, GraphNode, IndexResult, IndexStats, SearchResult, Symbol, SymbolKind, SymbolLang, } from './schema.js';
+export { projectIndexServerEndpoint, projectIndexServerMetadataPath, } from './project-server-endpoint.js';
 export { SCHEMA_VERSION } from './schema.js';
 export { codebaseIndexDirOverride, IndexStore, resolveIndexDir } from './writer.js';
 //# sourceMappingURL=index.d.ts.map

package/dist/codebase-index/index.js

@@ -2815,7 +2815,6 @@ import * as fs4 from "node:fs";
 import * as os from "node:os";
 import * as path5 from "node:path";
 import { fileURLToPath } from "node:url";
-import { assertUnixSocketPathWithinLimit } from "@wrongstack/core/utils";
 
 // src/codebase-index/writer.ts
 import { expectDefined } from "@wrongstack/core/utils";
@@ -4665,7 +4664,7 @@ var IndexStore = class _IndexStore {
       const ftsSchema = this.stmt(
         "SELECT sql FROM sqlite_master WHERE type='table' AND name='symbols_fts'"
       ).get();
-      if (ftsSchema?.sql && ftsSchema.sql.includes("unicode61")) {
+      if (ftsSchema?.sql?.includes("unicode61")) {
         this.db.exec("DROP TABLE IF EXISTS symbols_fts");
       }
       this.db.exec(SYMBOLS_FTS_SQL);
@@ -5158,9 +5157,13 @@ var IndexStore = class _IndexStore {
         sim: cosineSimilarity(queryVec, decodeVector(r.vector))
       })).sort((a, b) => b.sim - a.sim);
       const bm25Rank = /* @__PURE__ */ new Map();
-      bm25Rows.forEach((r, i) => bm25Rank.set(r.id, i));
+      bm25Rows.forEach((r, i) => {
+        bm25Rank.set(r.id, i);
+      });
       const vecRank = /* @__PURE__ */ new Map();
-      vecScores.forEach((r, i) => vecRank.set(r.id, i));
+      vecScores.forEach((r, i) => {
+        vecRank.set(r.id, i);
+      });
       const fused = reciprocalRankFusion(bm25Rank, vecRank, 60);
       const fusedScore = new Map(fused);
       const sorted = [...bm25Rows].sort(
@@ -6563,6 +6566,86 @@ import {
   isFrugalPerf
 } from "@wrongstack/core/utils";
 
+// src/codebase-index/content-hash.ts
+var PRIME64_1 = 0x9e3779b185ebca87n;
+var PRIME64_2 = 0xc2b2ae3d27d4eb4fn;
+var PRIME64_3 = 0x165667b19e3779f9n;
+var PRIME64_4 = 0x85ebca77c2b2ae63n;
+var PRIME64_5 = 0x27d4eb2f165667c5n;
+var MASK64 = 0xffffffffffffffffn;
+function mul64(a, b) {
+  return (a & MASK64) * (b & MASK64) & MASK64;
+}
+function rotl64(x, n) {
+  const v = x & MASK64;
+  return (v << BigInt(n) | v >> BigInt(64 - n)) & MASK64;
+}
+function readU64LE(buf, off) {
+  let v = 0n;
+  for (let i = 7; i >= 0; i--) {
+    v = v << 8n | BigInt(buf[off + i] ?? 0);
+  }
+  return v & MASK64;
+}
+function readU32LE(buf, off) {
+  return (BigInt(buf[off] ?? 0) | BigInt(buf[off + 1] ?? 0) << 8n | BigInt(buf[off + 2] ?? 0) << 16n | BigInt(buf[off + 3] ?? 0) << 24n) & MASK64;
+}
+function xxh64Round(acc, lane) {
+  return mul64(rotl64(acc + mul64(lane, PRIME64_2) & MASK64, 31), PRIME64_1);
+}
+function xxh64MergeRound(acc, val) {
+  return mul64(acc ^ xxh64Round(0n, val), PRIME64_1) + PRIME64_4 & MASK64;
+}
+function xxhash64Hex(buf, explicitLen) {
+  const length = explicitLen ?? buf.length;
+  let h;
+  let off = 0;
+  if (length >= 32) {
+    let v1 = PRIME64_1 + PRIME64_2 & MASK64;
+    let v2 = PRIME64_2;
+    let v3 = 0n;
+    let v4 = 0n - PRIME64_1 & MASK64;
+    const end32 = length - 32;
+    while (off <= end32) {
+      v1 = xxh64Round(v1, readU64LE(buf, off));
+      v2 = xxh64Round(v2, readU64LE(buf, off + 8));
+      v3 = xxh64Round(v3, readU64LE(buf, off + 16));
+      v4 = xxh64Round(v4, readU64LE(buf, off + 24));
+      off += 32;
+    }
+    h = rotl64(v1, 1) + rotl64(v2, 7) + rotl64(v3, 12) + rotl64(v4, 18) & MASK64;
+    h = xxh64MergeRound(h, v1);
+    h = xxh64MergeRound(h, v2);
+    h = xxh64MergeRound(h, v3);
+    h = xxh64MergeRound(h, v4);
+  } else {
+    h = PRIME64_5;
+  }
+  h = h + BigInt(length) & MASK64;
+  while (off + 8 <= length) {
+    const k1 = xxh64Round(0n, readU64LE(buf, off));
+    h = mul64(rotl64(h ^ k1, 27), PRIME64_1) + PRIME64_4 & MASK64;
+    off += 8;
+  }
+  if (off + 4 <= length) {
+    h = mul64(rotl64(h ^ mul64(readU32LE(buf, off), PRIME64_1), 23), PRIME64_2) + PRIME64_3 & MASK64;
+    off += 4;
+  }
+  while (off < length) {
+    h = mul64(rotl64(h ^ mul64(BigInt(buf[off] ?? 0), PRIME64_5), 11), PRIME64_1) & MASK64;
+    off += 1;
+  }
+  h = (h ^ h >> 33n) & MASK64;
+  h = mul64(h, PRIME64_2);
+  h = (h ^ h >> 29n) & MASK64;
+  h = mul64(h, PRIME64_3);
+  h = (h ^ h >> 32n) & MASK64;
+  return h.toString(16).padStart(16, "0");
+}
+function xxhash64String(content) {
+  return xxhash64Hex(new TextEncoder().encode(content));
+}
+
 // src/codebase-index/gitignore.ts
 import * as fs6 from "node:fs/promises";
 import * as path6 from "node:path";
@@ -7290,91 +7373,14 @@ function getParserPool() {
   return _pool;
 }
 
-// src/codebase-index/content-hash.ts
-var PRIME64_1 = 0x9e3779b185ebca87n;
-var PRIME64_2 = 0xc2b2ae3d27d4eb4fn;
-var PRIME64_3 = 0x165667b19e3779f9n;
-var PRIME64_4 = 0x85ebca77c2b2ae63n;
-var PRIME64_5 = 0x27d4eb2f165667c5n;
-var MASK64 = 0xffffffffffffffffn;
-function mul64(a, b) {
-  return (a & MASK64) * (b & MASK64) & MASK64;
-}
-function rotl64(x, n) {
-  const v = x & MASK64;
-  return (v << BigInt(n) | v >> BigInt(64 - n)) & MASK64;
-}
-function readU64LE(buf, off) {
-  let v = 0n;
-  for (let i = 7; i >= 0; i--) {
-    v = v << 8n | BigInt(buf[off + i] ?? 0);
-  }
-  return v & MASK64;
-}
-function readU32LE(buf, off) {
-  return (BigInt(buf[off] ?? 0) | BigInt(buf[off + 1] ?? 0) << 8n | BigInt(buf[off + 2] ?? 0) << 16n | BigInt(buf[off + 3] ?? 0) << 24n) & MASK64;
-}
-function xxh64Round(acc, lane) {
-  return mul64(rotl64(acc + mul64(lane, PRIME64_2) & MASK64, 31), PRIME64_1);
-}
-function xxh64MergeRound(acc, val) {
-  return mul64(acc ^ xxh64Round(0n, val), PRIME64_1) + PRIME64_4 & MASK64;
-}
-function xxhash64Hex(buf, explicitLen) {
-  const length = explicitLen ?? buf.length;
-  let h;
-  let off = 0;
-  if (length >= 32) {
-    let v1 = PRIME64_1 + PRIME64_2 & MASK64;
-    let v2 = PRIME64_2;
-    let v3 = 0n;
-    let v4 = 0n - PRIME64_1 & MASK64;
-    const end32 = length - 32;
-    while (off <= end32) {
-      v1 = xxh64Round(v1, readU64LE(buf, off));
-      v2 = xxh64Round(v2, readU64LE(buf, off + 8));
-      v3 = xxh64Round(v3, readU64LE(buf, off + 16));
-      v4 = xxh64Round(v4, readU64LE(buf, off + 24));
-      off += 32;
-    }
-    h = rotl64(v1, 1) + rotl64(v2, 7) + rotl64(v3, 12) + rotl64(v4, 18) & MASK64;
-    h = xxh64MergeRound(h, v1);
-    h = xxh64MergeRound(h, v2);
-    h = xxh64MergeRound(h, v3);
-    h = xxh64MergeRound(h, v4);
-  } else {
-    h = PRIME64_5;
-  }
-  h = h + BigInt(length) & MASK64;
-  while (off + 8 <= length) {
-    const k1 = xxh64Round(0n, readU64LE(buf, off));
-    h = mul64(rotl64(h ^ k1, 27), PRIME64_1) + PRIME64_4 & MASK64;
-    off += 8;
-  }
-  if (off + 4 <= length) {
-    h = mul64(rotl64(h ^ mul64(readU32LE(buf, off), PRIME64_1), 23), PRIME64_2) + PRIME64_3 & MASK64;
-    off += 4;
-  }
-  while (off < length) {
-    h = mul64(rotl64(h ^ mul64(BigInt(buf[off] ?? 0), PRIME64_5), 11), PRIME64_1) & MASK64;
-    off += 1;
-  }
-  h = (h ^ h >> 33n) & MASK64;
-  h = mul64(h, PRIME64_2);
-  h = (h ^ h >> 29n) & MASK64;
-  h = mul64(h, PRIME64_3);
-  h = (h ^ h >> 32n) & MASK64;
-  return h.toString(16).padStart(16, "0");
-}
-function xxhash64String(content) {
-  return xxhash64Hex(new TextEncoder().encode(content));
-}
-
 // src/codebase-index/indexer.ts
 var YIELD_EVERY_N = 50;
 function resolveParallelBatch() {
   return indexParallelBatchSize(availableParallelism());
 }
+function shouldUseParserWorkerPool(candidateFileCount, parseBatchCount) {
+  return !isFrugalPerf() && candidateFileCount >= WORKER_POOL_THRESHOLD && parseBatchCount > 1;
+}
 function yieldEventLoop() {
   return new Promise((resolve4) => setImmediate(resolve4));
 }
@@ -7551,11 +7557,7 @@ async function resolveProjectRelations(store, projectRoot, opts) {
     const structure = await detectModuleRoots(projectRoot, indexedFiles);
     if (opts.signal?.aborted) return;
     store.setFilePackages(assignPackageLabels(structure, indexedFiles));
-    const resolver = new ModuleResolver(
-      structure,
-      indexedFiles,
-      store.getNamespaceDeclarations()
-    );
+    const resolver = new ModuleResolver(structure, indexedFiles, store.getNamespaceDeclarations());
     const pending2 = store.getUnresolvedImports(opts.onlyFiles);
     const resolutions = [];
     for (const entry of pending2) {
@@ -7591,6 +7593,10 @@ async function runIndexerWithStore(store, opts) {
   const errors = [];
   const langStats = {};
   let filesIndexed = 0;
+  let filesParsed = 0;
+  let filesSkipped = 0;
+  let filesEmpty = 0;
+  let filesFailed = 0;
   let symbolsIndexed = 0;
   const isGitIgnored = await loadGitignoreMatcher(projectRoot);
   let files;
@@ -7632,12 +7638,14 @@ async function runIndexerWithStore(store, opts) {
       langStats[meta.lang] = (langStats[meta.lang] ?? 0) + meta.symbolCount;
       symbolsIndexed += meta.symbolCount;
       filesIndexed++;
+      filesSkipped++;
       filesPreSkipped++;
       return false;
     });
     if (filesPreSkipped > 0) opts.onProgress?.(filesPreSkipped, totalFilesForProgress);
   }
   const parallelBatch = resolveParallelBatch();
+  const parserPoolCandidateCount = files.length;
   let filesSinceLastYield = 0;
   for (let batchStart = 0; batchStart < files.length; batchStart += parallelBatch) {
     const batchEnd = Math.min(batchStart + parallelBatch, files.length);
@@ -7730,7 +7738,7 @@ async function runIndexerWithStore(store, opts) {
       });
     }
     if (toParse.length > 0) {
-      let pool = toParse.length >= WORKER_POOL_THRESHOLD ? getParserPool() : null;
+      let pool = shouldUseParserWorkerPool(parserPoolCandidateCount, toParse.length) ? getParserPool() : null;
       if (pool) {
         try {
           await pool.ensureReady();
@@ -7780,12 +7788,14 @@ async function runIndexerWithStore(store, opts) {
         const err = settled.reason;
         if (err instanceof Error && isAbortError(err)) throw err;
         errors.push(`batch error: ${file}: ${err instanceof Error ? err.message : String(err)}`);
+        filesFailed++;
         continue;
       }
       const result = settled.value;
       if (result.error) {
         if (result.missing) store.deleteFile(file);
         errors.push(`${file}: ${result.error}`);
+        filesFailed++;
         continue;
       }
       const { stat: stat2, lang, parsed } = result;
@@ -7793,6 +7803,7 @@ async function runIndexerWithStore(store, opts) {
         langStats[lang] = (langStats[lang] ?? 0) + result.skippedMeta.symbolCount;
         symbolsIndexed += result.skippedMeta.symbolCount;
         filesIndexed++;
+        filesSkipped++;
         const stored = existingMeta.get(file);
         if (stored && stored.mtimeMs !== result.skippedMeta.mtimeMs) {
           store.upsertFile({
@@ -7817,6 +7828,7 @@ async function runIndexerWithStore(store, opts) {
             contentHash: result.contentHash ?? ""
           });
           filesIndexed++;
+          filesEmpty++;
         }
         continue;
       }
@@ -7830,6 +7842,7 @@ async function runIndexerWithStore(store, opts) {
           contentHash: result.contentHash ?? ""
         });
         filesIndexed++;
+        filesEmpty++;
         continue;
       }
       batchEntries.push({
@@ -7851,6 +7864,7 @@ async function runIndexerWithStore(store, opts) {
           symbolsIndexed += count;
           langStats[entry.lang] = (langStats[entry.lang] ?? 0) + count;
           filesIndexed++;
+          filesParsed++;
         }
       } catch (err) {
         const message = err instanceof Error ? err.message : String(err);
@@ -7863,6 +7877,7 @@ async function runIndexerWithStore(store, opts) {
             symbolsIndexed += symbolsWithIds.length;
             langStats[entry.lang] = (langStats[entry.lang] ?? 0) + symbolsWithIds.length;
             filesIndexed++;
+            filesParsed++;
             if (entry.refs.length > 0 && symbolsWithIds.length > 0) {
               const fallbackBatch = assignRefsToSymbols2(entry.refs, symbolsWithIds);
               if (fallbackBatch.length > 0) store.insertRefsBatch(fallbackBatch);
@@ -7880,6 +7895,7 @@ async function runIndexerWithStore(store, opts) {
               contentHash: entry.contentHash
             });
           } catch (innerErr) {
+            filesFailed++;
             errors.push(
               `fallback write failed: ${entry.file}: ${innerErr instanceof Error ? innerErr.message : String(innerErr)}`
             );
@@ -7912,6 +7928,12 @@ async function runIndexerWithStore(store, opts) {
   const durationMs = Date.now() - startMs;
   return {
     filesIndexed,
+    fileOutcomes: {
+      parsed: filesParsed,
+      skipped: filesSkipped,
+      empty: filesEmpty,
+      failed: filesFailed
+    },
     symbolsIndexed,
     langStats,
     durationMs,
@@ -9466,6 +9488,8 @@ export {
   onIndexStateChange,
   outgoingCallsService2 as outgoingCallsService,
   packageGraphService2 as packageGraphService,
+  projectIndexServerEndpoint,
+  projectIndexServerMetadataPath,
   resetIndexCircuitBreaker,
   resolveIndexDir,
   resolveProjectIndexDaemonAvailability,

package/dist/codebase-index/indexer.d.ts

@@ -6,6 +6,12 @@ import { IndexStore } from './writer.js';
  * Re-resolved at the start of each index run so env profile changes apply.
  */
 export declare function resolveParallelBatch(): number;
+/**
+ * Pool startup is amortized across the complete index run, not one outer
+ * batch. Balanced batches are capped at 40 files, so comparing the per-batch
+ * parse count with the 500-file threshold made the worker path unreachable.
+ */
+export declare function shouldUseParserWorkerPool(candidateFileCount: number, parseBatchCount: number): boolean;
 interface IndexerOptions {
     projectRoot: string;
     files?: string[] | undefined;

package/dist/codebase-index/project-server-endpoint.d.ts

@@ -3,7 +3,7 @@ export declare const PROJECT_INDEX_SERVER_METADATA_FILE = "server.json";
 /**
  * Short directory name that owns the per-project Unix socket on Linux.
  *
- * The directory is created `0o700` by `ensureProjectIndexSocketDirectory`: on
+ * The directory is created `0o700` by `bindProjectEndpoint`: on
  * multi-user Linux `/tmp` the kernel's sticky bit otherwise lets a local
  * attacker pre-bind a predictable socket name (project paths are guessable
  * via the public `buildId` handshake), hijacking clients whose `bind()` hits
@@ -48,5 +48,4 @@ export declare function projectIndexServerKey(projectRoot: string, indexDir?: st
  */
 export declare function projectIndexServerEndpoint(projectRoot: string, indexDir?: string): string;
 export declare function projectIndexServerMetadataPath(projectRoot: string, indexDir?: string): string;
-export declare function ensureProjectIndexSocketDirectory(endpoint: string): void;
 //# sourceMappingURL=project-server-endpoint.d.ts.map