NppTreeSitter

A Notepad++ plugin that provides syntax highlighting and code folding powered by tree-sitter grammars.

From the Notepad++ menu:

Language -> T -> treesitter.*

Instead of hand-written lexer rules, NppTreeSitter uses tree-sitter's incremental parsing to produce a full syntax tree, then maps tree-sitter highlight captures to Scintilla style IDs. This gives accurate, language-aware highlighting that understands the actual structure of the code.

How it works

The plugin implements Scintilla's ILexer5 interface and exports the Lexilla API (GetLexerCount, GetLexerName, CreateLexer). Notepad++ discovers these at startup and registers each grammar as an external language (e.g. treesitter.python) in the Language menu.

Each language needs two files in the TreeSitterGrammars directory alongside the plugin DLL:

• Grammar DLL (e.g. tree-sitter-python.dll) -- compiled tree-sitter grammar exporting tree_sitter_<lang>()
• Highlight query (e.g. python-highlights.scm) -- standard tree-sitter highlights.scm with @capture names mapped to style categories

The tree-sitter core library is compiled directly into the plugin DLL (not a separate dependency).

Supported languages

The grammar manifest (grammars/grammars.json) lists a curated bundled set of common languages and common project/config formats that are intended to ship in release packages.

Release packages currently split them into two groups:

• preinstalled in TreeSitterGrammars: Python, Rust, Go, C, C++, C#, F#, JavaScript, TypeScript, Java, JSON, HTML, CSS, Bash, Ruby, PHP, XML
• bundled for install-on-demand in BundledTreeSitterGrammars: R, Lua, YAML, TOML, Markdown, CMake, Fortran, Pascal, MATLAB

Grammar DLLs are built from source using grammars/build-grammars.ps1.

The wider tree-sitter ecosystem has many more parsers. See the upstream parser list:

• https://github.com/tree-sitter/tree-sitter/wiki/List-of-parsers

Not every parser should be bundled by default. The bundled set should stay focused on broadly useful languages and avoid unusually large or niche parser packs.

The practical policy is:

• bundle mainstream languages and common project/config formats
• prefer official or well-maintained parser repositories
• avoid shipping very large parser collections or low-value niche grammars by default
• let additional grammars be added separately when needed

Adding more grammars

There are two different levels of support for additional grammars:

1. Add a grammar package for manual use
2. Add a grammar package plus auto-detect mapping

Manual grammar install

The plugin discovers grammars dynamically at runtime from TreeSitterGrammars.

If you place these files alongside the plugin:

• TreeSitterGrammars\tree-sitter-<language>.dll
• TreeSitterGrammars\<language>-highlights.scm
• optional: TreeSitterGrammars\<language>-locals.scm
• optional: TreeSitterGrammars\<language>-tags.scm
• optional: TreeSitterGrammars\<language>-injections.scm

then Notepad++ can expose that lexer as treesitter.<language> without rebuilding the plugin DLL.

Add a grammar to this repository

1. Choose a parser from the upstream list above.
2. Prefer repositories with:
   • a valid tree-sitter.json
   • either generated src/parser.c or a grammar.js that can be generated during the build
   • queries/highlights.scm
   • release tags or a stable tag to pin
3. Add the repo and tag to grammars/grammars.json.
4. Run grammars\build-grammars.ps1 for the desired platform.
5. Copy the generated tree-sitter-*.dll and *-*.scm files into TreeSitterGrammars.

Add auto-detect for a grammar

Dynamic grammar discovery is already supported, but file-extension auto-detect is still configured in code.

To make a newly added grammar auto-detect by file extension, update:

• src/dllmain.cpp in DetectTreeSitterLanguageForFile
• src/dllmain.cpp in PreferBuiltInLexerForExtension

This keeps the runtime flexible while still letting bundled languages have predictable defaults.

Installing bundled grammars

Release packages now include a BundledTreeSitterGrammars payload alongside the active TreeSitterGrammars directory.

The Install Missing Bundled Grammar command copies the matching per-architecture grammar files from that bundled payload into TreeSitterGrammars, then activates the lexer for the current buffer.

This is meant to work in an installed Notepad++ plugin directory, not only from a developer checkout.

If a language is already present in TreeSitterGrammars, the command stays disabled because there is nothing to install.

Building

Prerequisites

• Visual Studio 2022 or newer (v143+ toolset)
• Windows SDK 10.0
• PowerShell (for building grammar DLLs)
• Git (for the tree-sitter submodule)

Build the plugin

git clone --recurse-submodules https://github.com/<owner>/NppTreeSitter.git
cd NppTreeSitter
MSBuild NppTreeSitter.vcxproj -p:Configuration=Release -p:Platform=x64
MSBuild NppTreeSitter.vcxproj -p:Configuration=Release -p:Platform=Win32
MSBuild NppTreeSitter.vcxproj -p:Configuration=Release -p:Platform=ARM64

Outputs:

• build\x64\Release\TreeSitterLexer.dll
• build\Win32\Release\TreeSitterLexer.dll
• build\ARM64\Release\TreeSitterLexer.dll

Build grammar DLLs

cd grammars
.\build-grammars.ps1 -Platform x64
.\build-grammars.ps1 -Platform Win32
.\build-grammars.ps1 -Platform ARM64

This downloads each grammar repo listed in grammars.json, compiles the parser source into a DLL, and copies the query files used by the plugin (highlights, locals, injections, tags) when they are available.

The builder now refuses to bless incomplete cached source trees, refreshes broken extractions, and generates src/parser.c on demand for grammars that ship grammar.js instead of committing generated C sources.

Installation

Copy files into your Notepad++ installation:

<Notepad++>\plugins\TreeSitterLexer\TreeSitterLexer.dll
<Notepad++>\plugins\TreeSitterLexer\TreeSitterGrammars\tree-sitter-python.dll
<Notepad++>\plugins\TreeSitterLexer\TreeSitterGrammars\python-highlights.scm
<Notepad++>\plugins\TreeSitterLexer\TreeSitterGrammars\...  (other languages)
<Notepad++>\plugins\TreeSitterLexer\BundledTreeSitterGrammars\...  (bundled install payload)
<Notepad++>\plugins\Config\TreeSitterLexer.xml

Or run deploy.ps1 -Platform x64, deploy.ps1 -Platform Win32, or deploy.ps1 -Platform ARM64 (requires admin privileges for writing to Program Files).

After restarting Notepad++, the tree-sitter languages appear at the bottom of the Language menu.

Style mapping

Tree-sitter highlight captures are mapped to 14 Scintilla style IDs:

Style ID	Name	Typical captures
0	Default	(unmarked nodes)
1	Comment	@comment
2	Comment Doc	@comment.documentation
3	Keyword	@keyword, @keyword.*
4	String	@string, @string.*
5	Number	@number, @float
6	Operator	@operator
7	Function	@function, @function.*, @method
8	Type	@type, @type.*
9	Preprocessor	@preproc, @attribute
10	Variable	@variable, @variable.*, @parameter
11	Constant	@constant, @constant.*, @boolean
12	Punctuation	@punctuation.*
13	Tag	@tag, @tag.*

Colors and fonts for each style are configured in config/TreeSitterLexer.xml and can be customized through Notepad++'s Settings > Style Configurator.

Project structure

NppTreeSitter/
  src/
    dllmain.cpp          # DLL entry, N++ plugin stubs, Lexilla exports
    TreeSitterLexer.h     # ILexer5 implementation, grammar loader, style map
    TreeSitterLexer.cpp   # Full implementation (~730 lines)
  config/
    TreeSitterLexer.xml   # N++ style/language config for all lexers
  grammars/
    grammars.json         # Grammar manifest (repos + tags)
    build-grammars.ps1    # Build script for grammar DLLs
  external/
    tree-sitter/          # git submodule: tree-sitter core library
    scintilla/include/    # Vendored Scintilla headers (ILexer.h, Sci_Position.h)
  deploy.ps1              # Deployment script
  NppTreeSitter.vcxproj   # MSVC project file

License

GPL-2.0-or-later