helix-editor/regex-cursor

regex-cursor is a Rust library that lets you search for text patterns (regular expressions) across data that arrives in chunks rather than one continuous string. It was built for the Helix text editor, where the underlying text representation stores content in pieces rather than as a single block. Most regex tools assume you hand them a complete string. But text editors and similar applications often work with data structures that keep text in multiple segments. This project adapts the standard Rust regex engine to handle that chunked structure through a "cursor" interface that moves forward through chunks and can backtrack when needed. The practical use case is a code editor searching through a document that might be too large to hold in memory as one piece, or a tool working with text stored in a tree structure. Rather than copying everything into a single buffer just to run a search, you can search directly across the segmented data. One notable limitation: the backtracking requirement means this approach doesn't work for truly streaming scenarios like network data arriving in real time. The README discusses that most regex functionality needs to look backward at previous chunks, which requires buffering. Fully unbuffered streams like TCP connections would need a different approach, and the project documents why that's hard. The project is explicitly a prototype aimed at eventually adding streaming regex support to the upstream Rust regex ecosystem. The author notes that regex backtracks more than expected, which constrains what's possible for fully streaming searches.