Friends Of OpenJDK (foojay)

On Foojay, Frank Delporte explains what is coming in The Pi4J Project V5.0.0, the next major version of the open-source Java library for accessing Raspberry Pi hardware pins. Since 4.0.2, the main branch has seen about 300 commits and more than 80 merged pull requests. The main theme is cleanup, giving a leaner core and one clear way to create IO instances. Other changes include: - Fallback to mocks when no real hardware is available - A new ParallelPort IO type - Double values for PWM frequency and duty cycle - RISC-V support and better board detection - Java 25, Maven 4 and automated releases There is no release date yet, and several topics are still in progress. Snapshot builds are published for every change on main, and feedback is welcome. Read the full article: https://lnkd.in/ekjhx2EJ #Java #Pi4J #RaspberryPi

On Foojay, Shai Almog explains why the app team should decide whether to adopt a new iOS appearance, not Apple's release schedule. Codename One now lets you select Xcode 27 on its build servers with the returning ios.xcode_version hint. The iOS 27 theme is a separate setting, so you can update the SDK while keeping your current theme and CSS. The release also adds: - A JPA-inspired ORM that works in both the app and the backend - OpenTelemetry support from app to database - WebSocket support for Java backends - Push delivery callbacks with signed daily digests The article also covers removing OpenGL ES in favor of Metal and previews the follow-up posts planned for the week. Read the full article: https://lnkd.in/e_A8XSYD #Java #Kotlin #iOS #CodenameOne

On Foojay, Tim Kelly shows how to build a first SaaS with MongoDB Atlas Ephemeral Clusters. The tutorial builds a job-application tracker with a Kanban board, using Spring Boot, React with Vite, and MongoDB. An AI agent provisions a free, temporary Atlas M0 cluster with a single unauthenticated API call. It then uses the returned connection string to build the app. The cluster stays active for 48 hours and can be claimed in Atlas within 7 days of creation if the prototype is worth keeping. Ephemeral Clusters are currently in public preview and are meant for experiments and prototypes, not production. The article also covers keeping credentials out of source control and a detailed prompt you can give your coding agent. Read the full article: https://lnkd.in/eWK8AGig #MongoDB #Java #SpringBoot #AI

On Foojay today, Ted (Fyodor) Kupolov explains how to monitor a few Spring Boot applications on a small VPS without running Prometheus and Grafana. The article shows how StatLite reads the Actuator health and Micrometer metrics you already have. It presents health, requests, errors, latency, memory, uptime, and restarts in one dashboard. StatLite runs as a single Go binary and stores its history in SQLite. It also covers setup steps, compares it with raw Actuator endpoints, Spring Boot Admin, and Prometheus with Grafana, and explains when a broader observability platform is the better choice. Read the full article: https://lnkd.in/eBHYHXmC #SpringBoot #Java #Monitoring

On Foojay, Nicolas Fränkel 🇺🇦🇬🇪 looks at the different levels of bot protection he met while collecting genealogy data from public records sites, and what he used to get past each one. The article covers four levels: sites with little or no protection, where curl options like User-Agent, Referer, and session cookies are enough; then sites that require JavaScript (Anubis, Cloudflare JavaScript Detections); then sites that reject headless browsers (Turnstile, Imperva); and the signals they check. It also explains why a headed browser and a persistent profile help on some sites. He writes as a private person scraping for his own use, and reminds readers to scrape responsibly. Read the full article: https://lnkd.in/eHu2rJ_B #webscraping #botprotection #Playwright

On Foojay, Dominika T. explains what's in the October 2026 Azul Payara release: Payara Community 7.2026.10 and Azul Payara Server & Micro 7.5.0, 6.43.0, 5.92.0 and 4.1.2.191.60. For Azul Payara, the main reason to upgrade is security. Three security fixes ship across all branches, including a remote code execution vulnerability in Eclipse GlassFish and cross-site scripting in the Admin Console. For Payara Community, the focus is day-to-day operations: - SSH nodes no longer require elevated permissions - New asadmin commands: delete-jvm-option and delete-system-properties - Fixes for concurrent deployments and for apps failing to deploy after a domain restart - Three changes contributed by the community The release notes flag no breaking changes this cycle. Read the full article: https://lnkd.in/eiGXqpUU #Payara #JakartaEE #Java

On Foojay today, Somil Garg explains how he added semantic search to a Spring Boot job portal and tested whether storing two embeddings per job (one from the full posting, one from the requirements only) was worth the extra work. The setup uses Spring Boot 3.5 on Java 17, TiDB with its native VECTOR type, and gemini-embedding-001. The article explains how it extracts requirement sections with jsoup, how a SHA-256 hash marks stale vectors, and how it merges the two scores in plain Java. It also describes the evaluation harness he built, with 288 synthetic jobs and 40 hand-written queries, to check his guess instead of trusting a few manual searches. His short verdict: for the search mode he ships, it barely mattered. Read the article here: https://lnkd.in/e3cax9h6 #Java #SpringBoot #SemanticSearch #Embeddings

On Foojay, Sweety717 explains what was learned while building CodeGuard AI, a self-hosted AI code reviewer for GitHub Pull Requests, built with Java and Spring Boot. The main lesson: calling an LLM is the easy part. Most of the engineering sits around it: - Building review context from the PR changes - Defining review instructions - Turning raw AI output into structured findings with severity, risk and confidence - Posting the results back to GitHub - Supporting OpenAI, Google Gemini and Ollama The article also covers the limits of AI review. It works as a first pass, and human review still matters for business-critical logic and security-sensitive changes. Read the full article: https://lnkd.in/eWUMW7JV #Java #SpringBoot #AI #CodeReview

Error Prone is Google's open source static analysis tool that catches bugs at compile time, before they reach code review or production. Unlike traditional static analysis tools, Error Prone runs as a javac plugin, giving it access to all the compiler's type information. It flags code that's technically valid Java but likely incorrect, like trying to remove an element from a collection with an incompatible type, or using YYYY instead of yyyy in date patterns. The tool has been used on Google's Java codebase for years. Uber's NullAway runs as an Error Prone plugin, and Spring Framework 7 uses it in their build process. What makes it different: - No separate build step needed - Works with Maven, Gradle, Bazel, and Ant - Suggests fixes that can be applied automatically - Can be extended with custom rules using Refaster templates or BugCheckers Hüseyin Akdoğan explains how Error Prone works, shows real examples of bugs it catches, and walks through writing your own rules. https://lnkd.in/eUQ8wZ-3 #Java #StaticAnalysis #ErrorProne #CodeQuality

AI coding assistants have made writing code faster than ever. Reviewing it hasn't gotten any faster. That gap is where subtle bugs slip through: code that compiles, passes the tests, and still does the wrong thing. In my new article on Friends Of OpenJDK (foojay), I introduce Error Prone, the open source tool from Google that plugs straight into javac and catches these bugs at compile time, the cheapest moment to fix them. The article covers, with concrete examples: • How Error Prone's built-in checks catch code that compiles but is most likely wrong • How it differs from SpotBugs, PMD, Checkstyle and SonarQube • Writing transformation rules with Refaster • Writing a custom BugChecker that fails the build when sensitive data reaches a logger My takeaway: when code is produced faster than people can review it, catching bugs at compile time is no longer a luxury. It is a necessity. The article: https://lnkd.in/dvsU_wf9 Runnable examples: https://lnkd.in/d9gG2hDq #Java #ErrorProne #StaticAnalysis #SoftwareQuality #AI