Developer Releases Open-Source CLI Tool to Automate Legacy Java Code Modernisation

·3 views

A software developer has built and released java-migrate, a command-line tool that automates the modernisation of legacy Java codebases to Java 17, 21, or 25. The tool uses a regex-based scanner to detect outdated patterns such as anonymous inner classes, verbose null checks, and old-style switch statements across Java source files. Detected patterns are then sent to Anthropic's Claude AI model via a structured prompt, which returns modernised equivalents along with explanatory migration notes. A dry-run mode lets developers preview a unified diff of all proposed changes before anything is written to disk. The project is designed to reduce what can be a week of repetitive manual refactoring across large codebases into a single command.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Python NameError hid for months by lazy evaluation, caught only by AST static analysis

A browser automation tool began throwing a Python NameError after a code refactor in version 1.6.1 moved a local variable into a new function, while a residual check in the original function continued referencing the now-out-of-scope name. Because Python resolves variable names only at runtime rather than at import time, the bug remained invisible during normal startup and SSH-based update paths. The error only surfaced when the specific browser-automation code path executed, meaning it went undetected for multiple release cycles. The fix promoted the variable to a module-level constant accessible by both functions. To prevent similar issues, the team added an AST-based static analysis test to automatically detect dangling variable references across function boundaries.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

SilentShare Lets You Transfer Files Peer-to-Peer Directly in Any Browser

A developer has built SilentShare, a browser-based peer-to-peer file sharing app designed to eliminate the need for USB cables, cloud storage, or app installations. The tool uses WebRTC to transfer files, text, code snippets, and images of up to 50 MB directly between devices without routing data through a central server. Users create a room, share a generated code or QR link, and begin transferring files instantly across phones, laptops, or any browser-enabled device. Optional end-to-end encryption via AES-GCM is available when a room password is set, and no user accounts or permanent data storage are involved. The app also supports live transfer progress, pause and resume functionality, and media previews for images, audio, video, and PDFs.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to automate Flutter and canvas apps when the DOM offers no elements

Developer tools like find and read_page fail entirely on Flutter Web and canvas-rendered apps because their interfaces are drawn as pixels on a single canvas, leaving the DOM and accessibility tree effectively empty. A developer used Claude to automate the Rive animation editor by relying solely on screen coordinates, since no element-based alternative exists in such environments. Key techniques include waiting for the viewport width to stabilize before clicking, zooming into canvas regions to read painted text, and verifying every click visually through a screenshot before proceeding. Standard double-click actions can misfire in canvas editors, so firing two rapid single clicks in one batched action is recommended to reliably trigger hierarchy interactions. The author notes the procedure — stabilize, zoom-read, evidence-check, rapid double-fire — is slower than human interaction but prevents cascading errors from undetected missed clicks.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

HeronSignal Aims to Bridge the Gap Between Deployment and Real-User Experience

A developer frustrated by the disconnect between green dashboards and actual user experience built a monitoring tool called HeronSignal. The tool tracks real visitor sessions to detect JavaScript errors, failed network requests, slow pages, and Core Web Vitals, then uses AI to rank issues by user impact rather than raw event count. Unlike traditional monitoring platforms that overwhelm teams with data and alert fatigue, HeronSignal is designed to guide users through three steps: identifying the problem, understanding it, and fixing it. It can also draft a pull request with reproduction context, allowing engineers to review and merge a suggested fix without the tool ever autonomously touching the main codebase. The tool is implemented via a single script tag, requiring no build step, and can integrate with coding agents to answer real-time production questions directly.

0 comments Read more at DEV Community