Blender 5.0's upcoming Linux release will experimentally support High Dynamic Range (HDR) displays, but only when using Wayland and Vulkan. This means you'll need an HDR monitor, a Wayland desktop environment, and Vulkan acceleration to use it. The feature is currently under testing and its inclusion in the final release is not guaranteed.
Apple's Swift programming language is expanding into Android app development. While Android primarily uses Kotlin, a newly formed Swift Android Working Group aims to make Android an officially supported platform. This group's goals include improving Android support for Swift, optimizing core Swift packages for Android's idioms, defining supported API levels and architectures, and establishing best practices for bridging Swift with Android's Java SDK. Although third-party tools already enabled Swift for Android development, Apple's move signifies a strategic expansion in mobile development.
This tool leverages AI (GPT-4) and Faker to generate realistic datasets for demos, learning, and dashboards. Customize datasets via a conversational prompt builder, preview data in real-time, and export as CSV or SQL. It integrates with Metabase for easy data exploration. OpenAI API calls (costing ~$0.05) are only made for data previews; CSV/SQL downloads are free, generating more rows based on the preview's schema. Supports both One Big Table (OBT) and Star Schema for flexible analytics.
This article showcases the integration of the uv package manager with the Ray compute engine, addressing the challenges of dependency management in distributed Python applications. Traditional containerization methods slow down iteration speeds. The uv + Ray combination allows for rapid creation and synchronization of consistent Python environments across a cluster, dramatically improving development efficiency. By setting the environment variable `RAY_RUNTIME_ENV_HOOK`, Ray automatically detects the uv environment and applies it to all worker processes, ensuring consistent code execution. The article demonstrates its ease of use with examples using Ray Data and LLM integration, and covers advanced usage and best practices.
typr is a command-line typing test tool inspired by Keybr's word selection algorithm. It optimizes your typing speed by weighting accuracy, letter frequency, and typing speed. Featuring a cool curses-based TUI, it saves your data in a JSON file. Customize your test with options for time limits, word counts, or even endless practice.
This proposal suggests adding a declarative templating API to the web platform. The author argues that while the existing DOM API is powerful, it lacks ergonomic templating capabilities, forcing developers to rely on various frameworks and libraries, increasing development overhead and application size. Declarative templating offers superior code readability, security, performance, and maintainability, aligning with modern web development trends. The proposal suggests leveraging existing Tagged Template Literals, incrementally implementing the API, ultimately supporting HTML templates, thereby enhancing the web development experience.
Tired of syncing files between two Windows machines? Sink offers a revolutionary solution! Bypass the cloud, email, and USB drives. Running on your local network, Sink automatically detects other machines running Sink and syncs files near-instantly. It handles conflicts, preventing data loss, and allows you to ignore specific files and folders using a .sinkignore file (similar to .gitignore). Currently a work in progress, future plans include a UI, custom paths, system tray integration, and support for more than two devices.
Tired of complex build systems like Next.js, the author sought a simpler way to build static websites. After exploring various options, they discovered XSLT, a language for transforming XML into HTML. By storing website data as XML and using XSLT for transformation, they created a 'client-side' build system requiring no extra tools. The browser handles the XSLT transformation directly, generating the final HTML—a zero-config solution for static site generation. While not a perfect replacement for all scenarios, it offers developers a powerful alternative.
ISC is thrilled to announce the release of Kea 3.0.0, the first Long-Term Support (LTS) version! This major release opensources twelve previously commercially licensed Kea hooks, enhancing security and simplifying the installation process and client classification. Kea 3.0 also features a modernized Meson build system and native API access via HTTP and TLS, eliminating the need for the Kea Control Agent. This release marks a significant step towards a more open and maintainable Kea.
The author's Rust website took 4 minutes to build, requiring a rebuild, copy, and restart for every change. To speed things up, Docker containers and cargo-chef were used for dependency pre-building, but with limited success. Using rustc's self-profiling and LLVM trace data, the author identified link-time optimization (LTO) and LLVM module code generation as major bottlenecks. Through a series of optimizations, including adjusting LTO levels, optimization levels, and breaking up large async functions, build time was reduced from 4 minutes to 32.3 seconds.
Matrix 1.15 is here, boasting improvements to authentication, room summaries, and rich topics! This release incorporates 10 MSCs, highlighting the implementation of a next-generation authentication system – a significant step towards Matrix 2.0. New features also include enhanced room summaries, providing clients with richer room information, and support for rich text in room topics. These improvements enhance Matrix's security, user experience, and functionality.
Tired of bloated RSS readers? The author took a different approach, building a custom RSS reader using Deno and a simple text file. It displays only the titles and links of the three latest posts, without local full-text storage or read/unread markers, and is automatically updated daily via GitHub Actions. The code is concise, easy to understand and extend, a minimalist's dream.
Cubic's AI code review agent, initially plagued by excessive noise and false positives, underwent a significant transformation. Three key architectural revisions led to a 51% reduction in false positives. These improvements included requiring explicit reasoning from the AI before feedback, streamlining the toolset to essential components, and employing specialized micro-agents for focused tasks. The result? A smoother, more efficient code review process with increased developer trust and productivity.
Snow is an open-source Macintosh emulator written in Rust, aiming for hardware-level accuracy in emulating classic Motorola 680x0-based Macintosh computers. Unlike emulators that patch the ROM or intercept system calls, Snow focuses on low-level hardware emulation. Currently, it supports the Macintosh 128K, 512K, Plus, SE, Classic, and II. While under development, bleeding-edge builds and a limited online demo (emulated machine only) are available.
This article critiques the 'ticket-driven development' model where developers are overwhelmed with tasks, stifling thought and innovation. This leads to declining code quality, accumulating technical debt, and low team morale. The author advocates for developers to take ownership, think proactively, and strive for excellence, not just speed, ultimately focusing on building systems that work rather than simply closing tickets.
This article clears up the misconceptions surrounding mixed-DPI configuration support in the X11 windowing system. The author delves into the history of X11, from its origins with single X screens, through the Xinerama extension, to the modern XRANDR extension. The article reveals that X11 has always had the capability to support mixed DPI, but the key lies in client applications correctly utilizing the DPI information provided by the XRANDR extension. Different approaches are examined, along with solutions for various scenarios, including using the Qt toolkit for automatic mixed-DPI support and workarounds for toolkits lacking support, such as GTK+. Ultimately, the author calls on developers to improve toolkits and applications to fully leverage X11's mixed-DPI capabilities.
Windows 11's Snipping Tool just got a major upgrade! Version 11.2505.21.0 now lets you export screen recordings as GIFs, making sharing quick captures super easy. Record a video using Win + Shift + R or the app's screen recording mode. After recording, click the 'Export GIF' button, choose low or high quality, and save to your files or copy to your clipboard. Note: GIF export is limited to videos 30 seconds or shorter.
This article demonstrates how to generate structured JSON outputs from Llamafile, a locally runnable LLM. By leveraging LangChain's JsonOutputParser and PromptTemplate, and defining a custom Answer class to specify the desired JSON structure, the author chains together prompt, LLM, and parser components. This cleverly bypasses Llamafile's lack of built-in structured output functionality. A practical example using Llama-3.2-1B-Instruct-Q8_0.llamafile is provided, along with a link to the complete source code.
Xor, a graphics programmer, is passionate about crafting compact GLSL shader programs and sharing them on Twitter. He uses 'code golfing' techniques to squeeze complex visual effects, like simulated galaxies and voxel raytracers, into under 280 characters. This not only challenges his coding skills but has also connected him with a vibrant community of artists and programmers. Xor details his creative process and code golfing tips, encouraging others to experiment.
In early 1982, Apple's Lisa team tracked engineer productivity by lines of code. Bill Atkinson, QuickDraw's creator, found this metric absurd, prioritizing concise, efficient code. He optimized QuickDraw's region calculation, achieving a six-fold speed increase while reducing code by 2000 lines. On the productivity form, he famously reported '-2000'. Management wisely stopped using this flawed metric.
Fedora Linux's proposal to drop 32-bit support has sparked controversy, particularly threatening the popular handheld gaming distribution Bazzite. Bazzite's creator, Kyle Gospodnetich, strongly opposes the change, arguing it would kill projects like Bazzite and damage Fedora's public image. He points out that even with built Steam packages, basic use cases would break, and Flatpak wouldn't solve issues with Bazzite's reliance on 32-bit architecture for Steam Big Picture Mode. Currently, the proposal is unlikely to pass for Fedora 44, but ideally, Valve would port the Steam client to 64-bit, resolving many problems. The issue also impacts OBS Studio game capturing and FEX.
This article details the author's journey building a CUDA-based ray tracer that surpasses a Vulkan/RTX implementation—sometimes by over 3x—on identical hardware. Starting with a naive CUDA port, the author systematically optimized the renderer, tackling recursion, register pressure, memory layouts, and branching inefficiencies. Techniques like explicit stacks, structure of arrays, early ray termination, and Russian roulette were employed, resulting in a frame time reduction from 2.5 seconds to 9 milliseconds. The article dives deep into CUDA performance bottlenecks and offers practical optimization strategies. Benchmarks showcase the significant performance gains achieved on an RTX 3080.
This article guides you through building a simple web tool that leverages JavaScript's SpeechSynthesis API for text-to-speech (TTS) functionality and dynamic sentence highlighting. It explains how to control speech playback, set voice parameters, and track speech events. The article also details implementing sentence-level highlighting using CSS and JavaScript, culminating in a fully functional interactive reader with play, pause, resume, stop buttons, and voice selection. The complete HTML, CSS, and JavaScript code is provided.
Iroh is a novel network communication framework built on QUIC, enabling public key dialing and automatic discovery and maintenance of the fastest connection. Iroh attempts direct connections, falling back to public relay servers if necessary, with continuous monitoring for optimal speed. It offers pre-built protocols like iroh-blobs for large file transfers and iroh-gossip for building scalable publish-subscribe networks, simplifying development. Primarily written in Rust, Iroh also provides FFI bindings for other languages.
LM Studio 0.3.17 introduces Model Context Protocol (MCP) support, enabling users to connect various MCP servers and utilize them with local models. This release also adds support for 11 new languages, fixes numerous bugs, and improves the UI with a new 'Solarized Dark' theme. MCP allows large language models access to tools and resources; users should exercise caution when installing MCP servers from untrusted sources.
Howdy brings Windows Hello-like facial recognition login to Linux. Using your built-in IR emitters and camera, it provides convenient authentication. Integrating with PAM, it works everywhere you'd normally use a password: login, lock screen, sudo, su, etc. Howdy supports Debian/Ubuntu, Arch Linux, Fedora, and openSUSE, with detailed installation instructions covering .deb packages, AUR, Fedora COPR, and building from source. Important: Howdy's security is not absolute and should not be used as your sole authentication method.
Anthropic has launched the ability to build, host, and share interactive AI-powered apps directly within the Claude app. Developers can now iterate faster on their AI apps without worrying about scaling complexities and costs. Claude creates artifacts that interact via its API, turning them into shareable AI apps where usage is billed against the user's Claude subscription, not the developer's. Claude automatically writes code, handling prompt engineering, error handling, and orchestration logic. Features include using a Claude API within artifacts, processing files, creating rich React UIs, and forking/customizing artifacts. Current limitations include no external API calls, no persistent storage, and a text-based completion API. This beta feature is available to Free, Pro, and Max plan users.
The author bought a Nanoleaf Pegboard Desk Dock, a USB hub with RGB LEDs and gadget hooks, and decided to write a Linux driver for it since it only supports Windows and macOS. After obtaining protocol documentation from the vendor, the author used Rust and the libusb library to create a user-space driver. The article details the process, including using lsusb to get device information, handling USB interfaces, and using udev rules for user permissions. The author successfully created a simple driver to control the Desk Dock's RGB LEDs and handle device interrupts. Challenges like kernel driver occupancy and interrupt handling were overcome, leading to a deeper understanding of writing simple USB device drivers. While a simple proof of concept, it highlights that writing basic device drivers isn't as hard as it seems.
Bereket Engida, a self-taught programmer from Ethiopia, has quietly built Better Auth, an open-source authentication framework lauded by developers as the best they've ever used. This solo founder's startup recently raised $5 million in seed funding. Better Auth addresses limitations in existing tools, offering flexibility and on-premise data storage, appealing to AI startups and others concerned about data security. Boasting 150,000+ weekly downloads and 15,000+ GitHub stars, Better Auth is a testament to Engida's skill and a source of inspiration for aspiring African founders.
Libxml2, a widely used XML parser, highlights the successes and failures of the open-source model. After years of being relied upon by major corporations like Apple, Google, and Microsoft without adequate support, its maintainer, Nick Wellnhofer, is rejecting security embargoes and treating security vulnerabilities like regular bugs. This decision, driven by burnout and insufficient funding, sparked a debate about corporate responsibility in open source and the need for sustainable maintenance models. Wellnhofer's actions suggest a growing frustration within the open-source community and could signal a shift in how maintainers engage with large companies.