Faster Reports and Smoother Reading

This release focuses on keeping DevFeed fast and reliable as your data grows, while smoothing out a few high-impact reading and reporting issues.

Performance and reporting at scale

Several admin reports and feed options now run more efficiently on large datasets, while keeping the same results and accounting. This helps reports stay responsive as history grows and reduces slowdowns during heavy activity.

• Scheduler checks that only update an article’s next-check time no longer trigger full article card rebuilds, reducing unnecessary work.
• Token-usage, publication-review, and topic-inference reports reduce database work while preserving exact report results.
• Feed options now do less source-filter lookup work at large scale, while keeping source visibility, ordering, and combined filters intact.
• Old analysis payload cleanup now runs in bounded, resumable batches instead of repeatedly scanning history on every scheduler tick.

Overview and blocker reports now keep updating during continuous catalog refreshes. You will also see clearer status when the catalog is still catching up, with separate catalog progress so reports are not blocked by long-running catalog work.

Reader experience fixes

Returning to a background tab no longer resets your reading workspace. Loaded articles, Columns scroll positions, open editors, and the reading-streak panel are preserved across routine tab resume on web, Chrome, and Edge.

Article previews now keep your reading position stable when opening and closing, including delayed navigation behavior in Edge.

Leaderboards now include readers with recorded activity and public visibility even if they have not claimed a username yet. Readers without usernames appear with a display name (or a Reader fallback) and no profile link, while private accounts remain excluded.