subscript

What's Changed

• Pluggable language features
• Standardized common syntax
• Added assignment operators += -= *= /= %=
• Flatten code - no set, no wrappers
• Generic access node handling
• Simplified primitives wrapping [,true] vs ['',true]
• a?.valueOf?.() context calls support
• Misc. fixes

Full Changelog: https://github.com/dy/subscript/compare/v7.4.0...v8.0.0

• Fixed nary
• More complete tests
• Added skip argument to nary
• Better error display
• Type definitions
• Rassoc precedence is half, not full

Full Changelog: https://github.com/dy/subscript/compare/v7.4.4...v7.4.5

• Add explicit parse.unary, parse.binary and parse.nary registerers
• Add redefinable parse.id and parse.space

• Reintroduces AST/calltree as intermediary layer for better optimizations, extensibility and interop.
• Better registering fn.
• +10% parsing performance (170ms → 150ms)
• +20% eval performance (17ms → 13ms)
• +30% size (1kb → 1.3kb)

Full Changelog: https://github.com/spectjs/subscript/compare/v6.4.0...v7.0.0

• Collect & expose found arguments
• Better error messages
• Removed tagged literal

Full Changelog: https://github.com/spectjs/subscript/compare/v6.0.0...v6.1.0

What's Changed

• Direct evals as result of parsing.
• No lisp tree layer.
• Tokens and comments are parsed from lookup table.
• Single set method for defining operators or parsers.
• Postfix unaries are now custom parsers.
• Removed code, char functions.
• No fancy literal wrappers or string prefixes.
• Size -20% (2.5 → 2kb)
• Parsing perf +20% (200 → 160ms)
• Eval perf +50% (30 → 15ms)

Full Changelog: https://github.com/spectjs/subscript/compare/v5.5.2...v6.0.0

What's Changed

• ×3 faster eval due to flat reducer & fn type detection
• evaluate.operator is a function on par with parse.operator
• split core and subscript and justin configurators
• string, literal tokens are unary operators
• literals handling
• word operators are externalized
• more tests
• strings are prefixed with @ only, no "" tokens

Full Changelog: https://github.com/spectjs/subscript/compare/v5.0.0...v5.5.2

What's Changed

• Direct lookup table instead of precedence ascend
• Precedences match MDN
• Added operator function to create any operators or tokens
• Group/array are operators, not tokens
• Performance +30% (260 → 200ms)
• Size -10% (2.8 → 2.5kb)

Full Changelog: https://github.com/spectjs/subscript/compare/v4.0.0...v5.0.0

What's changed

• Functional parsers precedence table;
• Fast char lookup table;
• Improved performance by 25%;
• Space/comments;
• Correct ternary parser;
• Correct errors detection;
• Optimized expression parsing algorithm (merged prefix, postfix & binary);
• All JSEP tests;

Full Changelog: https://github.com/spectjs/subscript/compare/v3.0.0...v4.0.0

What's Changed

• Significantly increased performance (on par with jsep or faster for major cases)
• Extensible functional parsers instead of declarative configs
• Operators precedence matches JS
• Better source files layout
• Postfix operators support
• Smaller size
• Node tests

Full Changelog: https://github.com/spectjs/subscript/compare/v2.0.0...v3.0.0

What's Changed

• Generalized & improved performance
• Operator evaluators are separate from precedence dicts
• Streamlined expression algorithm to real recursive descent
• API is consolidated under parse.* and evaluate.* scopes
• Reduced size

Full Changelog: https://github.com/spectjs/subscript/compare/1.0.4...v2.0.0

• Parser performance based on stripped JSEP engine
• Frisk-compatible tree
• 1.4kb size

Full Changelog: https://github.com/spectjs/subscript/commits/1.0.4