Proposal for Google Summer of Code: Modular LilyPond Parser in C++
------------------------------------------------------------------------------------------------
I would like to propose a project to develop a modular LilyPond parser
in C++. The goal is a fault tolerant, context sensitive, and very fast
parser that can be used independently of the engraver system and
integrated into editors, analysis tools, and other applications. The
technical foundation is a trie handler pipeline for efficient tokenizing
and incremental re parsing.
Motivation
The current LilyPond parser is tightly coupled to the engraver system
and therefore difficult to reuse in external software. Modern editors
would benefit greatly from real time structural analysis of LilyPond
code. A standalone parser would make LilyPond more accessible to new
users and enable a wider range of applications.
A short video demonstrating my current editor is available here:
https://www.youtube.com/watch?v=nQNUPbI9swA
Project goals
1. Modularity: Lexer and parser as separate, reusable components.
2. Performance: Optimized for real time processing during text input.
3. Context sensitivity: Detection of which commands are valid in which
contexts; precise error reporting.
4. Fault tolerance: Robust handling of missing or mis nested brackets;
ability to read incomplete files.
5. Incremental parsing: Only modified regions need to be re parsed;
suitable for live editing.
Background
I am currently developing a LilyPond editor and have already implemented
a parser that reads LilyPond files into an AST on a block basis. This
parser works for well formed input, but it is not fault tolerant, cannot
handle missing or mismatched brackets, and is not incremental. Working
with these limitations has clarified which requirements a parser must
meet to be editor friendly and modular. A general purpose parser would
improve editor support and enable new tools for analyzing, converting,
or visualizing LilyPond files.
The goal is not to replace LilyPond’s existing parser, but to provide an
additional, editor oriented library that can be reused by external
tools. The initial GSoC scope would focus on a robust core (lexer,
parser, and AST for a well defined subset of LilyPond), with a clear
path for later extensions.
Additional note
My current family situation may or may not allow me to participate in
GSoC within the official time window. This may change by the time the
program starts. Regardless of GSoC, I intend to continue working on this
parser throughout the year and would like to align it with the needs and
expectations of the community. I am therefore sharing this proposal now
so that the project can be discussed independently of my availability
for GSoC.
Best regards,
cStr.
