Friends of OpenJDK Today

ParparVM went from 4.21x slower than warmed Java 25 to geomean parity, with six of ten benchmarks at or below HotSpot and peak memory below the JVM’s. The architecture behind it: frameless C codegen, a BiBOP page heap, and a poor man’s Valhalla.

Table of Contents Why JVM Warmup and Steady State Should Not Share the Same Resource ProfilePodResourceProfiles: Vertical Scaling Without RestartsMeasuring JVM Warm-up Using Azul Prime JMX MetricsMetrics: What This AchievesHow Vertical and Horizontal Scaling Work TogetherTry It YourselfClosing ThoughtsAutoscaling on …

Table of Contents What pinning actually isThe two causes — and it really is just twoThe canonical bugWhat JDK 24 changed (JEP 491)How to catch itHow to fix itWhy I ended up automating the readThe incident that taught me to …

Table of Contents 🔤 Headline: BoxStringBuilder and Dramatically Faster String ConcatenationThe Performance StoryA First-Class BoxStringBuilder TypeOptimization 1: &= Does In-Place AppendOptimization 2: Compiler Self-Assignment RewriteOptimization 3: Compile-Time Literal FoldingOptimization 4: Auto-Switching Runtime Concat StrategyJava Interop Nuance🚀 Headline: Pluggable ClassLoader Factory …

Table of Contents The idea: Copilot-driven profilingStep 1 — Copilot ran it, and showed me exactly what it ranStep 2 — The verdict in plain languageStep 3 — One click hands the data back to CopilotStep 4 — A root …

Table of Contents YouTubePodcast AppsGuestsLinksContentFoojay Podcast hits episode 100. No plan, no roadmap. It just happened. To mark the occasion, Frank turned the microphone around and invited other podcasters: Adam Bien (airhacks.fm), Jennifer Reif (Breaktime Tech Talks), Kadi McKean and …

Table of Contents The pre-database era, replayedThe write-ahead log, but for your local variablesTransactions that span monthsThe same pattern, elsewhere in the stackWhat this means for your Java codeOnce upon a time, applications managed their own data files. Every program …

Event-Driven Architecture can unlock scalability, loose coupling and powerful integrations; but it also introduces retries, eventual consistency, observability challenges and operational overhead. This article explores seven situations where EDA may be the wrong choice, from simple CRUD applications to workflows that require strong consistency or immediate responses.

This week’s release adds car dashboards, motion sensors, desktop-class input, foldables, Commerce, Secrets, versioned builds, and real video generation while keeping last week’s open-source deal intact.

Table of Contents The problem: we all became distributed systems developersDurable execution: crash-proof, not crash-preventingHow it actually works: history replay, not snapshotsThe moving partsSeeing is believingWhy this changes how you write softwareWhen developers first hear about Temporal, they tend to …

Table of Contents Dynamic Sets – A First-Class CollectionRanges – Lazy, Typed, Extensible IntervalsInner Classes and Template ClassesClass References as Callable ConstructorsDataNavigator JSONPath SupportQuery Transformers – Own Your Result ShapeGlobal Query DefaultsCompanion Release: bx-mcp Is HereOther Notable AdditionsschedulerNew() BIFserver.webModeString BIFs: …

Table of Contents The collector doesn’t delete anythingMost objects die youngref = null is dead codeWhen ref = null can helpConclusionOne of the misconceptions that I continuously run into is that nulling out references in Java helps garbage collection. This …

Table of Contents Why It’s Important To Foojay CommunityHow the Premiere WorksOur AskEvery so often a story comes along that belongs to an entire community rather than any one company or person. The Java Story documentary by CultRepo is one …

Table of Contents Verify the signature of the JDK you downloadTencent Kona JDKInstall a JDK without making it the defaultSmarter version files: auto-detect the distributionFrom .sdkmanrcA much nicer Maven experienceQuieter logs by defaultNon-interactive settings.xmlToolchains that stop growing out of controlDon’t …

When we investigated unexpected performance variance in our Timefold Solver worker pods on our platform, we traced it to hyperthreading on AMD (x86-64) cloud instances. Switching to ARM (ARM64), where each vCPU maps to a physical core, eliminated the problem …

Table of Contents Improved Comparison WorkflowRendering FixesBuilt with Systematic AI CodingThe ResultThe CostWhat’s NextA Lottie library is only as good as its output looks. If an animation renders differently in Lottie4J than it does in the official web player, that’s …