I have verified that the user davidhu2000 is NOT a heavy public user (created 2015, 46 repos, 182 followers), but fetching their contribution data in a 4-year chunk consistently fails with a 504 Gateway Timeout.

However, fetching a 1-year chunk succeeds instantly.

This confirms that the issue is likely due to a specific data quirk or hidden complexity (e.g., massive private contributions) that makes the 4-year query too expensive for GitHub to compute in one go.

Solution: Adaptive Chunking
We will modify Updater.mjs to implement a smart fallback:

1. Attempt to fetch the standard 4-year chunk.
2. If it fails (throws an error), catch it.
3. Log a warning and switch to fetching the years one by one for that specific chunk.

This ensures the updater remains fast for normal users while automatically handling these edge cases without crashing.