I was reading this article and came across a new idea https://awesomeagents.ai/news/github-fake-stars-investigation/ GitHub doesn't verify users like other social services. However, not to ruin the verification badge like on X and to put that feature under a subscription, this needs to be changed (the verification badge will be accepted via a trusted group of maintainers). As a maintainer myself, I feel that these fake people can bring a large security risk, probably make it harder for us to decide what dependencies to use, for e.g. they can spread malicious packages through package indexes like PyPI, npmjs, jsr.io and crates.io. This thing can disintegrate the entire chain, making the system harder to recover.

Comments URL: https://news.ycombinator.com/item?id=47908718

Points: 1

# Comments: 1

Comments URL: https://news.ycombinator.com/item?id=47822301

Points: 27

# Comments: 5

Comments URL: https://news.ycombinator.com/item?id=47260423

Points: 3

# Comments: 0

Key features: - Full TypeScript support with excellent DX - Plugin architecture – attach only what you need (sorting, filtering, export, etc.) - Export to CSV, Excel, JSON, and PDF out of the box - Cell editing with formula support (=SUM, =AVG, etc.) - SQL query panel – query your grid data with SQL syntax - Clipboard support (Ctrl+C/X/V) - Master-detail rows, column pinning, row grouping - No external CSS framework required

It's designed to be simple to get started:

    import { WarperGrid } from '@itsmeadarsh/warper-grid';
    

Would love feedback from the community. What features do you look for in a data grid?

Comments URL: https://news.ycombinator.com/item?id=46853711

Points: 2

# Comments: 1

Why I built this:

I was working on a trading dashboard that needed to display 500k+ rows of live tick data. react-window worked fine at 10k rows, but at scale, every scroll event caused visible jank. Profiling showed the bottleneck was JavaScript - calculating visible ranges and item offsets was blocking the main thread for 8-12ms on each scroll.

I wondered: what if the math happened in WASM instead?

What it does:

Warper is a React hook (useVirtualizer) that renders only visible items. The difference from existing libraries:

Scroll position to visible range calculations happen in Rust/WASM Item offsets stored in a Fenwick tree for O(log n) updates when heights change Binary search for scroll-to-index is O(log n) instead of O(n) No JS object allocations during scroll = no GC pauses Technical details:

The WASM module maintains a Fenwick tree (binary indexed tree) of item heights. When you scroll:

JS passes scroll position to WASM WASM binary searches the tree to find the start index WASM returns [startIndex, endIndex, offsets[]] React renders only those items For variable heights, updating a single item's height is O(log n) instead of recalculating all subsequent offsets.

Benchmarks (Chrome, M1 Mac):

10k items: react-window 60 FPS, Warper 120 FPS 100k items: react-window ~40 FPS, Warper 120 FPS 1M items: react-window ~12 FPS, Warper 118 FPS At 10M items, react-window crashes. Warper holds at ~115 FPS.

Try it:

Live demos (no signup): https://warper.tech Stress test with 1M rows: https://warper.tech/one-million-rows Variable height chat: https://warper.tech/chat

GitHub: https://github.com/warper-org/warper npm: npm install @itsmeadarsh/warper (~45KB gzipped with WASM)

Limitations:

Requires top-level await or a bundler plugin for WASM Only React 18+ (uses useSyncExternalStore under the hood) Horizontal virtualization not yet implemented Happy to answer questions about the Fenwick tree implementation or WASM/JS interop challenges.

Comments URL: https://news.ycombinator.com/item?id=46512865

Points: 2

# Comments: 0

Live Demo: https://warper.tech

10 million rows at 120+ FPS. No lag. No dropped frames.

Existing libraries like react-window and react-virtuoso hit walls at scale. They rely on JavaScript binary search and O(n) scans for variable heights. Warper is different.

The engine uses:

- Fenwick trees for O(log n) variable-height lookups - O(1) arithmetic for fixed-height items - Zero-copy typed arrays between WASM and JS - Pre-allocated memory pools with zero GC during scroll

Benchmarks on M1 MacBook Pro with 1M rows: Warper hits 119 FPS. react-virtuoso drops to 58 FPS. At 10M rows Warper stays at 118 FPS while virtuoso falls to 31 FPS.

Its not marginally faster. Its 2-4x faster depending on the workload.

Trade-off is bundle size: around 45KB vs 6KB for react-window. Worth it if you need the performance.

Available through GitHub Sponsors at github.com/sponsors/itsmeadarsh2008

Happy to answer questions about the architecture.

Comments URL: https://news.ycombinator.com/item?id=46271032

Points: 3

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=44193954

Points: 1

# Comments: 2

Comments URL: https://news.ycombinator.com/item?id=43400287

Points: 1

# Comments: 0

Haze is a high-performance, easy-to-use Magic Link Authentication service for Python applications. Generate secure authentication links that work across devices with minimal setup.

Features

     Fast & Efficient - Optimized core with minimal overhead
     Ultra Secure - Modern cryptography with JWT
     Highly Configurable - Like Neovim, but for authentication
     Zero Daemon - No background processes required
     Minimal Dependencies - Lightweight core with optional extras
     Cross-Device Auth - Click on the phone, authenticate on the desktop
     Pluggable Storage - Use any database system
     Modern Defaults - NanoID, JWT, MsgPack by default

Comments URL: https://news.ycombinator.com/item?id=43391667

Points: 3

# Comments: 2

Comments URL: https://news.ycombinator.com/item?id=43317195

Points: 1

# Comments: 1

Comments URL: https://news.ycombinator.com/item?id=42803428

Points: 2

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=41356261

Points: 3

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=41320548

Points: 2

# Comments: 1