Loom

A generic incremental parser framework for MoonBit — edit-aware lexing, a lossless green tree (CST), subtree reuse, error recovery, and a reactive pipeline. Any grammar plugs in via a single Grammar[T, K, Ast] value.

New here? Start with the dowdiness/loom package README for the API and a Quick Start, then browse the docs index.

Quick Start

Monorepo — no root moon.mod.json; run moon from each module's directory. incr, egraph, egglog, and event-graph-walker are git submodules, so clone with --recursive:

git clone --recursive https://github.com/dowdiness/loom.git && cd loom
moon update                               # fetch the package registry index
(cd loom && moon test)                    # framework module (loom/)
(cd examples/lambda && moon test)         # lambda example
(cd examples/lambda && moon bench --release)

Multi-module development workflow: docs/development/managing-modules.md.

Modules

Core framework (stable):

Module	Path	Purpose
dowdiness/diagnostic	diagnostic/	Parser-independent structured diagnostics, rendering, and atomic fixes
dowdiness/diagnostic_moji	diagnostic-moji/	Opt-in Unicode display-cell rendering adapter for structured diagnostics
dowdiness/diagnostic_pretty	diagnostic-pretty/	Opt-in width-aware Unicode report layouts with semantic annotations
dowdiness/loom	loom/	Parser framework: incremental parsing, CST building, grammar composition
dowdiness/seam	seam/	Language-agnostic CST infrastructure (CstNode / SyntaxNode)
dowdiness/incr	incr/	Salsa-inspired reactive inputs / derived cells

Sibling modules (see each module's README for scope and status):

Module	Path	Purpose
dowdiness/pretty	pretty/	Wadler-Lindig pretty-printer (generic Layout[A], annotations) — used by examples/json
dowdiness/egraph	egraph/	Equality graph for equality saturation
dowdiness/egglog	egglog/	Relational e-graph engine (Datalog + equality saturation)
dowdiness/text_change	text-change/	Pure contiguous text-change utilities (migrated from canopy 2026-05, #147)
dowdiness/moji	moji/	UAX #29 grapheme cluster + word boundary segmentation, UTF-16 indexed (migrated from canopy 2026-05, #147)

Examples

Example	Path	Purpose
Lambda Calculus	examples/lambda/	Reference grammar — typed SyntaxNode views, error recovery, CRDT exploration
JSON	examples/json/	Step-based prefix_lexer + block_reparse_spec — exercises every Grammar::new option
Markdown	examples/markdown/	Mode-aware lexing via ModeLexer — line-start / inline / fenced code contexts
JSX	examples/jsx/	Streaming-prefix error recovery — every EOF truncation keeps already-parsed content (generative-UI foundation)
MoonBit	examples/moonbit/	Skeleton official MoonBit lexer adapter + coarse Loom CST grammar
Graph DSL	examples/graph-dsl/	Source-map/token-role graph authoring example with graph-operation lowering

Documentation

• docs/README.md — navigation index (start here, architecture, API, archive)
• ROADMAP.md — phase status and future work
• CHANGELOG.md — user-facing changes