$ show Floorp-Projects/Floorp
Floorp is a Firefox-based browser that extends Gecko with a comprehensive feature layer including workspaces, split-view tiling, panel sidebar, PWA/SSB management, mouse gestures, keyboard shortcuts, command palette, tab stacks, design theming, private containers, external browser integration, notes, profile manager, and more. The architecture is hybrid: chrome components are built with SolidJS and injected via a custom loader into Firefox's XUL, while settings, new tab, notes, welcome wizard, and profile manager pages are standalone React/Vite web applications served within the browser. Configuration is stored in Firefox preferences with io-ts runtime validation, and cross-process communication uses a birpc bridge. The build system is Deno-based, orchestrating Vite builds, Firefox runtime patching, and platform-specific packaging (macOS App Shim in C++, Windows Tauri stub installer). The codebase demonstrates strong engineering practices with extensive test coverage (colocated browser tests, unit tests, integration tests), clean reactive state management patterns, and well-structured feature modules. However, it carries technical debt from multiple legacy migration paths, duplicated UI components across page apps, and tight coupling to Firefox internal APIs that require version-specific workarounds.
The LLM identified 7 files worth highlighting for reuse:
Self-contained fuzzy string matching algorithm with scoring. Implements a character-level fuzzy match that returns relevance scores for search results. Has zero dependencies on Firefox internals or the Floorp ecosystem, making it directly extractable into any TypeScript project needing fuzzy search functionality.
Implementation of the $1 gesture recognizer algorithm, which matches 2D mouse/touch gestures against a library of predefined templates using angular similarity. Pure algorithm with no external dependencies, reusable in any application requiring gesture recognition (browsers, drawing apps, accessibility tools).
Custom HTML-to-Markdown converter with rule-based transformation, HTML parsing, whitespace collapse, and commonmark compliance. Includes fingerprinting for content deduplication. Self-contained module that could be extracted for any project needing server-side HTML-to-Markdown conversion without relying on external services.
CJK text segmenter using the BudouX algorithm for splitting Japanese and Chinese text into meaningful segments for search indexing. Pure utility with no browser dependencies, reusable in any application needing CJK text segmentation for search or NLP pipelines.
HTML content fingerprinting algorithm that generates stable, content-based hashes of web page sections for deduplication and change detection. Uses DOM structure and text content to create fingerprints resilient to minor formatting changes. Self-contained and reusable in web scraping, archiving, or monitoring systems.
Zone computation algorithm for determining drop zones in a split-view/tiling layout. Calculates optimal pane positions and sizes based on container dimensions and existing pane layout. Pure geometry algorithm with no external dependencies, reusable in any tiling window manager or split-pane UI system.
Utility for computing character-level highlight ranges in search result text. Given a query and a result string, produces an array of [start, end] ranges indicating which characters match the query for UI highlighting. Self-contained algorithm reusable in any search interface needing result highlighting.