eyaltoledano/claude-task-master · error · StreamingError

STREAM_PROCESSING_FAILED

STREAM_PROCESSING_FAILED

Error message

Failed to process AI text stream: ${error.message}

What it means

handleStreamError is the StreamParser's central error wrapper, invoked from processStream. Any non-StreamingError thrown while reading or processing the AI text stream (I/O failures, unexpected stream termination, JSON decode explosions) is wrapped in a StreamingError with code STREAM_PROCESSING_FAILED, preserving the original message. StreamingError instances are re-thrown untouched so callers can branch on the specific code.

Source

Thrown at src/utils/stream-parser.js:396

			this.progressTracker.addText(chunk);
			this.jsonParser.write(chunk);
		});

		try {
			await processor.process(textStream);
		} catch (streamError) {
			this.handleStreamError(streamError);
		}

		this.jsonParser.end();
	}

	handleStreamError(error) {
		// Re-throw StreamingError as-is, wrap other errors
		if (error instanceof StreamingError) {
			throw error;
		}
		throw new StreamingError(
			`Failed to process AI text stream: ${error.message}`,
			STREAMING_ERROR_CODES.STREAM_PROCESSING_FAILED
		);
	}

	async waitForParsingCompletion() {
		// Wait for final parsing to complete (JSON parser may still be processing)
		await new Promise((resolve) => setTimeout(resolve, 100));
	}

	async attemptFallbackIfNeeded() {
		const fallbackItems = await this.fallbackParser.attemptParsing();
		return fallbackItems.length > 0;
	}

	buildResult(usedFallback) {
		const metadata = this.progressTracker.getMetadata();

View on GitHub (pinned to c0c98d367c)

Solutions

1. Read error.message to identify the underlying cause, and check error.code === 'STREAM_PROCESSING_FAILED' to confirm it came from this wrapper.
2. Add retry logic around the whole stream fetch + parse operation for transient network drops.
3. Validate the stream source (encoding, availability) before passing it to parse().
4. If the cause is a bug in your own progressTracker/config callbacks, fix the callback rather than handling the wrapper.
5. Catch StreamingError specifically so other errors (like [510]'s plain Error) remain distinct.

Example fix

// before
const result = await parser.parse(stream);
// after
try {
  const result = await parser.parse(stream);
} catch (err) {
  if (err.code === 'STREAM_PROCESSING_FAILED') {
    console.error('Stream processing failed, retrying:', err.message);
    return await retryWithBackoff(() => parser.parse(fetchStream()));
  }
  throw err;
}

function isUsableStream(stream) {
  return stream != null &&
    (typeof stream[Symbol.asyncIterator] === 'function' || typeof stream.on === 'function');
}
if (!isUsableStream(stream)) {
  throw new TypeError('parse() expects a string or readable/async-iterable stream');
}

function isStreamingError(err) {
  return err instanceof Error && typeof err.code === 'string' && err.name === 'StreamingError';
}

import { StreamingError } from './streaming-error.js';
try {
  await parser.parse(stream);
} catch (err) {
  if (err instanceof StreamingError && err.code === 'STREAM_PROCESSING_FAILED') {
    console.error('Underlying cause:', err.message);
    // retry with fresh stream or fall back to non-streaming request
  } else {
    throw err;
  }
}

When it happens

Trigger: The text stream emits malformed bytes that crash both the JSONStreamParser and FallbackParser; the underlying stream emits an 'error' event (connection dropped mid-response); an exception inside progressTracker callbacks or waitForParsingCompletion; any bug thrown inside processStream's iteration loop.

Common situations: LLM providers closing the connection mid-stream under load; proxy/timeouts truncating the response; the parser receiving a stream in an unexpected encoding; upgrading the AI SDK so the stream object no longer implements the async-iterator interface the parser expects.

Related errors

• ${this.getName()} provider error: ${errorMessage}
• START_FLOW_FAILED
• BACKEND_UNREACHABLE
• CODE_AUTH_FAILED
• PKCE_FAILED

AI-assisted analysis of eyaltoledano/claude-task-master@c0c98d367c (2026-08-29). Data as JSON: /api/errors/87c234ddde00dc01. Report an issue: GitHub.