io7m-com / jsx Goto Github PK

View Code? Open in Web Editor NEW

1.0 3.0 0.0 2.69 MB

S-expression parsing

Home Page: https://www.io7m.com/software/jsx/

License: ISC License

Java 98.70% Assembly 1.26% CSS 0.04%

java s-expression parser pretty-print

jsx's Introduction

jsx

JVM	Platform	Status
OpenJDK (Temurin) Current	Linux
OpenJDK (Temurin) LTS	Linux
OpenJDK (Temurin) Current	Windows
OpenJDK (Temurin) LTS	Windows

jsx

A general-purpose, configurable S-expression parser.

Features

Hand-coded lexer and parser with full support for tokens using characters outside of the Unicode BMP.
Optional square brackets [f (g [x y])].
Optional multi-line strings.
Configurable comment characters (#, %, or ;).
Configurable pretty printing of expressions.
High coverage test suite.
OSGi-ready
JPMS-ready
ISC license.

Documentation

See the user manual.

Usage

Parsing

Give the configuration for the lexer:

var squareBrackets = true;
var newlinesInQuotedStrings = true;
var startAtLine = 1;

final var lexConfiguration =
  new JSXLexerConfiguration(
    squareBrackets,
    newlinesInQuotedStrings,
    Optional.of(URI.create("file.txt")),
    EnumSet.noneOf(JSXLexerComment.class),
    startAtLine
  );

Instantiate a parser using a parser and lexer configuration:

var preserveLexicalInfo = true;

final var parserConfig =
  new JSXParserConfiguration(preserveLexicalInfo);

final JSXParserSupplierType parsers =
  new JSXParserSupplier();

final var parser =
  parsers.createFromStreamUTF8(
    parserConfig,
    lexConfiguration,
    lexers,
    System.in
  );

Parse expressions:

final var exprOpt = parser.parseExpressionOrEOF();
if (exprOpt.isPresent()) {
  final var e = exprOpt.get();
  System.out.println(e);
}

jsx's People

Contributors

Stargazers

Watchers

jsx's Issues

Do not extend java.util.List

Unfortunately, the SExpressionListType currently extends the java.util.List interface. This effectively implies that that implementations must be mutable. This also makes them somewhat painful to implement from Kotlin, which evidently currently has serious trouble delegating to java interfaces.

Publish services

The JSXLexerSupplier and JSXParserSupplier types should be services.

Add the ability to accept only square brackets as list delimiters

Currently, the parser only allows for accepting square brackets in addition to parentheses. It would be good to be able to specify that only square brackets can be list delimiters.

Add an ability to specify an upper bound on error counts

If the lexer or parser encounter more than N errors, they should stop. This avoids any potential non-termination.

Exception types should implement LexicalType

This was an oversight. They still define lexicalInformation() methods when there's an interface they could implement right there...

This means another major version increment to fix.

Serializer does not escape quote characters in quoted strings

Use less stack

io7m/kstructural#50

Add configurable comment tokens

Line comments are needed. The Lisp comment character is ; but I've also seen #, // and --. It would be nice if this was configurable somehow.

Misleading error message

The actual error is on line 97; the IdA1AdminSearchParameters record expression isn't closed. This is arguably a jsx bug; the lexical information it gives is the end of the expression rather than the start.

Admin1.txt

io7m-com/cedarbridge#12