Gradle

📣 #Gradle 9.8.1 is out! https://lnkd.in/gVPYnb-m Fixes security issues and regressions reported against 9.8.0 Report new 🐛 on GitHub! #WhatsNew #Gradle

📣 #Gradle 8.14.6 is out! https://lnkd.in/e_WV9KUu Fixes security issues and regressions reported against 8.14.5 Report new 🐛 on GitHub! #WhatsNew #Gradle

Today we're shipping Agentic Gradle Plugin 1.0.0 which includes two skills: • gradle-best-practices checks your build against the official Gradle best practices and proposes fixes. • gradle-wrapper-upgrade upgrades your wrapper, pins the checksum, checks the build still configures, and rolls back if it doesn't. We tested both on seven models from five vendors. Every model did better with the skills, and cheaper models came close to frontier-model results. Install the skills and tell us what you think: https://lnkd.in/eTX_tgfs

📣 Gradle 9.8.0 is out! https://lnkd.in/e4jmQKuj Bug fixes and… ☕ Java 27 support 🪞Maven mirror settings reuse 🔗 Linked problem locations in build output Please report any 🐛 on GitHub!

📣 Gradle 9.8.0 RC2 is out and ready for testing 🔬 https://lnkd.in/eGnXGPpY Bug fixes and… ☕ Java 27 support 🪞Maven mirror settings reuse 🔗 Linked problem locations in build output Please report any 🐛 on GitHub!

At Gradle, we're building tooling and skills to help coding agents work well with Gradle. That means measuring what actually helps, not guessing. As part of that effort, Agentic Gradle, we pointed a new internal eval harness at a real, open GitHub issue and ran two coding agents against it under identical conditions: same prompt, same 50-turn limit, no follow-up. Both models, Fable 5.1 and Astra 6, fixed the bug. How they got there, and which fix we'd actually want to maintain, were not the same. Full breakdown — diffs, cost, token counts, and our call on which fix ships — is on the Gradle blog: https://lnkd.in/enktJs7D

📣 Gradle 9.8.0 RC1 is out and ready for testing 🔬 https://lnkd.in/e3JryMys Bug fixes and… ☕ Java 27 support 🪞Maven mirror settings reuse 🔗 Linked problem locations in build output Please report any 🐛 on GitHub!

Every AI agent that touches your codebase eventually has to run your build to check its own work. That means agent commit volume doesn't just add commits — it multiplies how many times your build runs. If your cache is quietly missing on a task that should be instant, that miss doesn't happen once anymore. It happens on every agent iteration, every retry, every verification pass. Waste that was tolerable at human commit rates isn't tolerable at agent commit rates. Worth asking your platform team: do we know which of our caches are actually working, or are we assuming? 🔧 #Develocity #BuildAcceleration #AI #PlatformEngineering

SolarWinds suffered a cyber attack in 2020 on its IT infrastructure management and monitoring software. The attackers got into the build system and injected malicious code during compilation, while the source stayed clean. Over 18,000 SolarWinds customers installed the compromised updates, which the attackers then used to reach into government agencies and enterprises. The code that got reviewed and the binary that got shipped were two different artifacts. You patch your dependencies. You patch your base images. You patch your CI runners. But who patches your build tool? Today, we're introducing the Gradle Security Subscription: prioritized security fixes under SLA, extended maintenance for versions past end-of-life, and embargoed vulnerability disclosure ahead of public release. The open source Gradle you know doesn't change. The release end-of-life policy has been in place for years, and security fixes continue under it exactly as before. The subscription is for organizations with formal compliance obligations, or production systems pinned to older versions, where best effort isn't enough. https://lnkd.in/eYCb3XpE

Flaky tests cost more than the minutes they burn in CI. They cost trust — once a team stops believing a red build means something is actually broken, they stop reacting to red builds at all. Flaky Tests: The Productive Developer's Guide covers practical strategies for reducing flaky-test impact and rebuilding a culture where a failing test still means something. 📘 Get the guide: https://lnkd.in/eM2PFpZa #Develocity #TestReliability #DPE #SoftwareQualit