RhoMicro.ParserGenerator 1.0.1

ParserGenerator

Overview

This is a parser generator for generating parsers that are suitable for use in Roslyn source generators. It allows for faster development of Domain Specific Languages (DSL) by automating the development of parsing infrastructure. For example, DSLs that transpile to C# are suitable for implementation as a source generator using this library.

Installation

Terminal:

dotnet add package RhoMicro.ParserGenerator --version 1.0.0

PackageReference:

<PropertyGroup>
    <PackageReference Include="RhoMicro.ParserGenerator" Version="1.0.0"/>
</PropertyGroup>

Quick Start

Define a grammar for which to generate a parser:

[assembly: ParserGenerator.GenerateParser(
    """
    type-parameter-template = HEAD arity-hole TAIL
    arity-hole = "{" arity-range / ARITY "}"
    arity-range = ARITY "-" ARITY
    HEAD = "a"-"z" / "A"-"Z" TAIL
    TAIL = *("a"-"z" / "A"-"Z" / "0"-"9" / "_")
    ARITY = "1"-"9" / ("1"-"9" "0"-"9") / ("1" / "2" "0"-"4" "0"-"9") / ( "2" "5" "0"-"6")

    """)]

Here is a reference for the syntax of grammars (they use a dialect of ABNF): <table> <thead> <tr> <th>Name</th> <th>Syntax</th> <th>Explanation</th> </tr> </thead> <tbody> <tr> <td>Rule List</td> <td>r = s<br/>s = p</td> <td> The rules r and s are defined.<br/> Rule definitions are terminated by line breaks.<br/> Ensure that the last rule is also terminated by a line break. </td> </tr> <tr> <td>Rule</td> <td>r = p</td> <td>The rule r is defined by the pattern p.</td> </tr> <tr> <td>Quoted Range</td> <td>"a"-"z"</td> <td>Matches a single character in the range [a,z].</td> </tr> <tr> <td>Hex Range</td> <td>%x61-7A</td> <td>Matches a single character in the range [a,z].</td> </tr> <tr> <td>Quoted Sequence</td> <td>"foo"</td> <td>Matches the character sequence foo.</td> </tr> <tr> <td>Hex Sequence</td> <td>%x66.6F.6F</td> <td>Matches the character sequence foo.</td> </tr> <tr> <td rowspan=4>Repetition</td> <td>n*mp</td> <td>Matches at least n and at most m repetitions of p.</td> </tr> <tr> <td>mp</td> <td>Matches at most m repetitions of p.</td> </tr> <tr> <td>np</td> <td>Matches at least n repetitions of p.</td> </tr> <tr> <td>*p</td> <td>Matches any amount of repetitions of p.</td> </tr> <tr> <td>Option</td> <td>[p]</td> <td>Optionally matches p.</td> </tr> <tr> <td>Concatenation</td> <td>p q</td> <td>Matches p followed by q.</td> </tr> <tr> <td>Alternative</td> <td>p / q</td> <td>Matches p or q.</td> </tr> <tr> <td>Group</td> <td>(p)</td> <td>Matches p.</td> </tr> <tr> <td>Wildcard</td> <td>.</td> <td>Matches any character.</td> </tr> <tr> <td>Reference</td> <td>r</td> <td>Matches the pattern defining the rule r.</td> </tr> </tbody> </table>

Utilize the generated lexer and parser on an input text:

var options = LexerOptions.Default;
var lexer = new Lexer(templateSource, options);
var parser = new Parser(lexer);
var parsing = parser.Parse(ct);

If the input text is not recognized by the parser, null is returned. The generated infrastructure does not emit diagnostics for invalid inputs or other failure modes. This is an area for future development.

Process the parsed syntax tree, e.g. using one of the provided visitors:

var stringBuilder = new StringBuilder();
var output = new StringWriter(stringBuilder);

output.WriteLine("/*");

parsing?.Accept(
    new SyntaxTreeWriter(
        SyntaxTreeWriterOptions.Default,
        new DescendantsCollector(DescendantsCollectorOptions.ChildNodesOnly),
        output),
    ct);

output.WriteLine("*/");