1 stable release
new 1.0.0 | Apr 7, 2025 |
---|
#60 in #incremental-parser
23MB
536K
SLoC
tree-sitter-netlinx


NetLinx grammar for tree-sitter.
Contents 📖
What's Working ✅
- Expressions
- ✅ Binary Expressions
- ✅ Bitwise Expressions
- ✅ Bitwise Word Expressions (
band
,bor
,bxor
,bnot
,lshift
,rshift
) - ✅ Unary Expressions
- ✅ Update Expressions
- ✅ Assignment Expressions
- ✅ Devchan Expressions
- ✅ Devchan Range Expressions
- ✅ Comparison Expressions
- ✅ Logical Expressions
- ✅ Logical Word Expressions (
and
,or
,xor
,not
) - ✅ String Expressions
- ✅ Function Call Expressions
- ✅ Device Expressions (
0:first_local_port+1:1
,dvPort.NUMBER:dvPort.PORT:dvPort.SYSTEM
) - ✅ Parenthesized Expressions
- ✅ Subscript Expressions
- ✅ Comma Expressions
- ✅ Compiler Variables (
__file__
,__line__
,__date__
,__time__
, etc) - ✅ System Variables (
day
,date
,ldate
,time
, etc) - ✅ System Constants (
true
,false
, etc) - ✅ System Functions (all functions defined in
NetLinx.axi
) - ✅ System Types (all types defined in
NetLinx.axi
)
- Statements
- ✅ Expression Statements
- ✅ Compound Statements
- ✅ Return Statements
- ✅ Break Statements
- ✅ Continue Statements
- ✅ If Statements
- ✅ While Loops
- ✅ For Loops
- ✅ Switch/Case Statements
- ✅ Select/Active Statements
- ✅ Create Buffer Statements
- ✅ Create Multi Buffer Statements
- ✅ Clear Buffer Statements
- ✅ Wait Statements
- ✅ Wait Until Statements
- ✅ Cancel Wait Statements
- ✅ Cancel Wait Until Statements
- ✅ Cancel All Wait Statements
- ✅ Cancel All Wait Until Statements
- ✅ Break Statements
- ✅ Section Statements
- ✅ Program Name
- ✅ Module Name
- ✅ Send String Statements
- ✅ Send Command Statements
- ✅ Send Level Statements
- ✅ Devchan Operation Statements (
ON
,OFF
,TO
,MIN_TO
,PULSE
, etc) - ✅ Call Statements (for legacy
DEFINE_CALL
functions) - ✅ System Call Statements
- Declarations
- ✅ Define Function Definitions
- ✅ Define Library Function Declarations
- ✅ Define Call Definitions
- ✅ Variable Declarations
- ✅ Constants Declarations
- ✅ Type Declarations
- ✅ Module Definitions
- ✅ Combine Definitions
- ✅ Connect Level Definitions
- ✅ Toggling Definitions
- ✅ Mutually Exclusive Definitions
- Events
- ✅ Button Events
- ✅ Channel Events
- ✅ Level Events
- ✅ Data Events
- ✅ Timeline Events
- ✅ Custom Events
- ✅ Legacy Push
- ✅ Legacy Release
- Literals
- ✅ String Literals
- ✅ Single Quoted String Literals
- ✅ Escape Sequence for Single Quotes (
''
)
- ✅ Number Literals
- ✅ Decimal
- ✅ Hexadecimal
- ✅ Floating Point
- ✅ Device Literals
- ✅ String Literals
- Comments
- ✅ Single Line Comments
- ✅ Multi Line Comments (C Style
/* */
) - ✅ Pascal Comments (
(* *)
)
- Preprocessor
- ✅ Define
- ✅ Include
- ✅ Warn
- ✅ Disable Warning
- ✅ If Defined
- ✅ If Not Defined
Known Limitations ⚠️
Preprocessor Directives in Expressions
The NetLinx language allows preprocessor directives to be used within expressions, like:
(foo #IF_DEFINED BAR && baz #END_IF && foobar)
While this is valid NetLinx code that compiles correctly, tree-sitter has limitations when parsing these constructs due to the nature of preprocessor directives operating at a different level than normal syntax.
When encountering preprocessor directives within expressions, the parser will:
- Maintain the overall expression structure - The parenthesized expression remains intact
- Generate some error nodes - The preprocessor directives are marked as errors
- Preserve correct syntax highlighting - Despite the errors, tokens are still correctly identified
- Keep all identifiers and operators - Variable names and operators remain properly connected
Example Parse Tree
(source_file
(expression_statement
(parenthesized_expression
(ERROR
(identifier) // <- foo
(preproc_if_defined_keyword))
(binary_expression
left: (binary_expression
left: (identifier) // <- BAR
right: (identifier)) // <- baz
(ERROR
(preproc_end_if_keyword))
right: (identifier))))) // <- foobar
Implications
- Editor Experience: Syntax highlighting and code navigation should work normally
- Error Reports: Your editor may show these areas as errors, which can be safely ignored
- Alternative Approach: For cleaner parsing, consider restructuring complex conditional expressions to avoid embedding preprocessor directives within expressions, like:
#IF_DEFINED BAR (foo && baz && foobar) #ELSE (foo && foobar) #END_IF
Install ⚡
Node.js (npm)
For JavaScript/Node.js projects:
npm install tree-sitter-netlinx
# or
yarn add tree-sitter-netlinx
# or
pnpm add tree-sitter-netlinx
Rust (Cargo)
For Rust projects:
cargo add tree-sitter-netlinx
Python (pip)
For Python projects:
pip install tree-sitter-netlinx
Manual Installation
If you want to install the grammar manually, you can clone the repository and build it yourself:
git clone https://github.com/Norgate-AV/tree-sitter-netlinx
cd tree-sitter-netlinx
npm install
npx tree-sitter generate
Design 🎨
The grammar is designed to be as accurate as possible, while also being as flexible as possible.
Permissive Parsing
The grammar is intentionally permissive, allowing it to parse syntactically valid but semantically questionable code. This approach enables:
- Better error recovery during editing
- A more forgiving experience during development
- The ability to parse incomplete or incorrect code
- Better syntax highlighting and code navigation
Syntax vs. Semantics
As a parsing tool, tree-sitter focuses on syntactic structure rather than semantic validity:
- The parser will accept constructs that are syntactically correct but might fail during compilation
- Semantic validation should be handled by the NetLinx compiler or separate analysis tools
- This separation allows the grammar to be more stable and maintainable
Examples of Accepted Patterns
The parser will accept patterns that the NetLinx compiler might reject:
- Declarations with inconsistent or incomplete type specifiers
- Mixed implicit and explicit typings
- Unusual combinations of modifiers
- Devchan range expressions used with devchan operations
Flexibility Over Semantic Correctness
The parser deliberately prioritizes syntactic flexibility over strict semantic validation:
-
Section-Independent Parsing: Declarations can appear anywhere in the code, even outside their semantically correct sections. The parser doesn't enforce section-specific constraints that the NetLinx compiler would apply.
-
Context-Free Analysis: Device definitions, constants, and variables are parsed based on their syntactic structure rather than their semantic context. For example, device definitions in a
DEFINE_DEVICE
section are parsed as standard assignment expressions. -
Support for Implicit Typing: The parser accommodates NetLinx's implicit typing behaviors. In NetLinx, when type declarations are omitted, the compiler applies implicit types—
INTEGER
for regular variables andCHAR
for array variables.
Examples:
DEFINE_CONSTANT
FOO = 1 // Parsed as an assignment expression rather than a specialized constant declaration
DEFINE_VARIABLE
bar = 1 // Parsed as an assignment expression
// NetLinx compiler would implicitly type this as INTEGER
baz[10] // Parsed as an identifier with subscript
// NetLinx compiler would implicitly type this as CHAR array
This approach enables more resilient parsing during code editing and provides better syntax highlighting and tooling support, even for incomplete or semantically imperfect code. Semantic validation is intentionally left to the NetLinx compiler or separate analysis tools.
Team ⚽
This project is maintained by the following person(s) and a bunch of awesome contributors.
Damien Butt |
Contributors ✨
Thanks go to these awesome people (emoji key):
This project follows the all-contributors specification.
Contributions are welcome! Please fork and open a pull request if you have any suggestions or improvements.
Any help would be greatly appreciated.