feat: Add adaptive timeouts and WASM memory error recovery for robust parsing

Implements file-size-based timeouts (base 10s + 10s per 100KB), more frequent worker recycling (250 files), and automatic retry logic for WASM memory corruption failures. Workers now crash immediately on memory errors to prevent cascading failures, with failed files automatically retried on fresh workers with clean heaps.
This commit is contained in:
Colby McHenry
2026-04-04 22:49:36 -05:00
parent c19c0ca8fd
commit 1271ad9161
3 changed files with 83 additions and 7 deletions
+65 -6
View File
@@ -45,7 +45,7 @@ const PARSE_TIMEOUT_MS = 10_000;
* V8 isolate by terminating the worker thread and spawning a fresh one.
* This interval balances memory usage against the cost of reloading grammars.
*/
const WORKER_RECYCLE_INTERVAL = 500;
const WORKER_RECYCLE_INTERVAL = 250;
/**
* Progress callback for indexing operations
@@ -521,8 +521,12 @@ export class ExtractionOrchestrator {
logWarn('Parse worker exited unexpectedly', { code });
rejectAllPending(`Worker exited with code ${code}`);
}
// Clear reference so we know to respawn
if (parseWorker === w) parseWorker = null;
// Clear reference so we know to respawn, reset count so
// the fresh worker gets a full cycle before recycling.
if (parseWorker === w) {
parseWorker = null;
workerParseCount = 0;
}
});
}
@@ -580,17 +584,20 @@ export class ExtractionOrchestrator {
const id = nextId++;
workerParseCount++;
// Scale timeout for large files: base 10s + 10s per 100KB
const timeoutMs = PARSE_TIMEOUT_MS + Math.floor(content.length / 100_000) * 10_000;
return new Promise<ExtractionResult>((resolve, reject) => {
const timer = setTimeout(() => {
pendingParses.delete(id);
log(`TIMEOUT: ${filePath} exceeded ${PARSE_TIMEOUT_MS}ms — killing worker`);
log(`TIMEOUT: ${filePath} exceeded ${timeoutMs}ms — killing worker`);
// Reject FIRST — worker.terminate() can hang if WASM is stuck
parseWorker = null;
workerParseCount = 0;
reject(new Error(`Parse timed out after ${PARSE_TIMEOUT_MS}ms`));
reject(new Error(`Parse timed out after ${timeoutMs}ms`));
// Fire-and-forget: kill the stuck worker in the background
worker.terminate().catch(() => {});
}, PARSE_TIMEOUT_MS);
}, timeoutMs);
pendingParses.set(id, { resolve, reject, timer });
worker.postMessage({ type: 'parse', id, filePath, content });
@@ -712,6 +719,58 @@ export class ExtractionOrchestrator {
}
}
// Retry pass: files that failed due to WASM memory corruption may succeed
// on a fresh worker with a clean heap. Collect retryable failures, recycle
// the worker, and try each one individually.
const retryableErrors = errors.filter(
(e) => e.code === 'parse_error' && e.filePath &&
(e.message.includes('Worker exited') || e.message.includes('memory access out of bounds'))
);
if (retryableErrors.length > 0 && WorkerClass) {
log(`Retrying ${retryableErrors.length} files that failed due to WASM memory errors...`);
// Force a fresh worker
recycleWorker();
for (const errEntry of retryableErrors) {
const filePath = errEntry.filePath!;
if (signal?.aborted) break;
let content: string;
try {
const fullPath = validatePathWithinRoot(this.rootDir, filePath);
if (!fullPath) continue;
content = await fsp.readFile(fullPath, 'utf-8');
} catch {
continue; // Skip files we can't read
}
let result: ExtractionResult;
try {
result = await requestParse(filePath, content);
} catch {
continue; // Still failing — leave as errored
}
if (result.nodes.length > 0 || result.errors.length === 0) {
// Success on retry — store result and fix counts
const language = detectLanguage(filePath);
const stats = await fsp.stat(path.join(this.rootDir, filePath));
this.storeExtractionResult(filePath, content, language, stats, result);
// Remove the original error and update counts
const idx = errors.indexOf(errEntry);
if (idx >= 0) errors.splice(idx, 1);
filesErrored--;
filesIndexed++;
totalNodes += result.nodes.length;
totalEdges += result.edges.length;
log(`Retry OK: ${filePath} (${result.nodes.length} nodes)`);
}
}
}
// Shut down parse worker and clear any pending timers
rejectAllPending('Indexing complete');
if (parseWorker) {
+8
View File
@@ -33,6 +33,14 @@ parentPort!.on('message', async (msg: { type: string; id?: number; filePath?: st
parentPort!.postMessage({ type: 'parse-result', id, result });
} catch (err) {
const message = err instanceof Error ? err.message : String(err);
// WASM memory errors leave the module in a corrupted state — all
// subsequent parses would also fail (cascading failures). Crash the
// worker so the main thread spawns a fresh one with a clean heap.
if (message.includes('memory access out of bounds') || message.includes('out of memory')) {
process.exit(1);
}
parentPort!.postMessage({
type: 'parse-result',
id,
+10 -1
View File
@@ -172,8 +172,17 @@ export class TreeSitterExtractor {
this.visitNode(this.tree.rootNode);
this.nodeStack.pop();
} catch (error) {
const msg = error instanceof Error ? error.message : String(error);
// WASM memory errors leave the module in a corrupted state — all subsequent
// parses would also fail. Re-throw so the worker can detect and crash,
// forcing a clean restart with a fresh heap.
if (msg.includes('memory access out of bounds') || msg.includes('out of memory')) {
throw error;
}
this.errors.push({
message: `Parse error: ${error instanceof Error ? error.message : String(error)}`,
message: `Parse error: ${msg}`,
filePath: this.filePath,
severity: 'error',
code: 'parse_error',