v0.5.0 release

Features:

• Can now parse TLA+ snippets without an enclosing module declaration
• Updated highlighting queries with a focus on more advanced reference highlighting

Breaking changes:

• single_quantifier_bound rule merged with quantifier_bound rule
• Changed identifier in record_value and except rules to be identifier_ref

v0.4.1 release

• Fixed bug with parsing of strong & weak fairness
• PlusCal blocks now use their own external scanner parsing context
• Updated tree-sitter CLI version and regenerated parser
• Updates to queries

v0.4.0 release

Added PlusCal support
Changed parsing of block comments to support PlusCal parsing
Added PlusCal highlighting queries
Added local/reference highlighting for Neovim

v0.3.0 release

• Nat, Int, and Real can now use the alternative ℕ, ℤ, and ℝ symbols respectively
• Converted primitive_value_set node to a supertype, added nodes for each primitive value set type
• Fixed not specifying ⟩, : and → as right delimiters in external scanner
• For dev dependencies, use fork of tree-sitter-cli with get_column method modified to count codepoints

v0.2.1 release

• Added alternative unicode symbols
• Removed intermediate postfix_op_symbol child node from bound_postfix_op

v0.2.0 release

This release includes a number of changes to the parser's node structure:

• unit node is now hidden, but exposed as a supertype
• number node is now hidden, but exposed as a supertype
• Nodes that were previously hidden children of number are now visible
• Added string escape character node
• Renamed single_line nodes in module header to header_line
• Added keyword extraction

In addition, this release also includes the tree-sitter-tlaplus.wasm file in the NPM package.

v0.1.2 release

This release ensures the grammar builds on macOS by removing C++11 language features from the external scanner.

v0.1.1 release

This release makes minor changes to the parsing of extramodular text and block comments, to better facilitate code folding. Consider a common form of block comment:

(****************************)
(* text text text text text *)
(* text text text text text *)
(* text text text text text *)
(****************************)

Previously, the grammar would parse this as a series of block comments (one for each pair of (* and *)). Now the external scanner will look ahead after the end of a block comment to see whether another block comment starts on the very next (or same) line. If so, it will group that block comment in with the current one. Thus large block comments can now be folded.

v0.1.0 initial release

Grammar is fully-functional but parse tree layout is subject to change.