better-parse

• Fixed #51 (thanks a lot @InfectedBytes!)
• Update Kotlin dependencies to 1.6.20

• Fix reported range when an optional parser is lifted to an AST parser (thanks @Thomvis!)
• Fix InvalidMutabilityException in RegexToken (thanks @itegulov!)
• Use Kotlin 1.5.31

• Make the JVM RegexToken implementation thread-safe
• Release JS IR artifacts along with the JS legacy ones

• Updated to Kotlin 1.4.21 and new Multiplatform library publishing layout (no need to depend on platform modules or -metadata part)
• Fixed issue #28
• better-parse is now available on Maven Central

• better-parse is now a multiplatform library! All Kotlin targets are supported;
• Optimize tokenization and parsing, thanks @orangy for his help!
• Replace token with literalToken and regexToken

• Fixed Regex / Pattern converted to their string representations with loss of flags;
• Made Parsed<T>.remainder public;
• Optimized TokenizerMatchesSequence to avoid redundant objects allocation

What's new:

• Introducing syntax tree parsers that store the tree structure of the parsed input including the start/end positions of each tree node. Your parsers can automatically be transformed from Parser<T> to Parser<SyntaxTree<T>>, see the Syntax trees section;

  

• Tokenizer interface to customize the default tokenizing behavior; To support tokenizer implementations that do not use Java regexes, Token now stores a raw String instead of a Pattern. The default tokenizer implementation is now named DefaultTokenizer.

• Parser<T> properties now can be delegated inside Grammar, which will store them into the the new declaredParsers property;

• Optimizations of built-in combinators;

• Regex parameters of Token factory functions annoteted for IntelliJ IDEA regex language injection;

• Names for Tokens are no more necessary.

Migration notes:

• If you used Lexer explicitly, use DefaultTokenizer instead;

• If you used Token::pattern, change to using the raw pattern string;

• It is also good to change the val p = someParser declarations inside the Grammars to val p by someParser.

• Change JVM target to 1.6 to make better-parse compatible with Android

• Change the semantics of and chains: when there's only one non-skipped operand Parser<T>, the whole chain is now typed as Parser<T>, not Parser<Tuple1<T>>

• Add operator equivalents: a * b for a and b, and -a for skip(a).