webidl-rs

A parser for WebIDL in Rust.

Documentation

Example

Lexing

```rust use webidl::*;

let lexer = Lexer::new("/* Example taken from emscripten site */\n\ enum EnumClassEnumWithinClass {\n\ \"EnumClass::eval\"\n\ };"); asserteq!(lexer.collect::>(), vec![Ok((41, Token::Enum, 45)), Ok((46, Token::Identifier("EnumClassEnumWithinClass".tostring()), 71)), Ok((72, Token::LeftBrace, 73)), Ok((74, Token::StringLiteral("EnumClass::eval".to_string()), 92)), Ok((93, Token::RightBrace, 94)), Ok((94, Token::Semicolon, 95))]); ```

Parsing

```rust use webidl::; use webidl::ast::;

let parser = Parser::new(); let result = parser.parse_string("[Attribute] interface Node { };");

asserteq!(result, Ok(vec![Definition::Interface(Interface::NonPartial(NonPartialInterface { extendedattributes: vec![ Box::new(ExtendedAttribute::NoArguments( Other::Identifier("Attribute".tostring())))], inherits: None, members: vec![], name: "Node".tostring() }))])); ```

Pretty printing AST

An example of a visitor implementation can be found here. Below is an example of how it is used:

```rust use webidl::ast::; use webidl::visitor::;

let ast = vec![Definition::Interface(Interface::NonPartial(NonPartialInterface { extendedattributes: vec![ Box::new(ExtendedAttribute::NoArguments( Other::Identifier("Attribute".tostring())))], inherits: None, members: vec![InterfaceMember::Attribute(Attribute::Regular(RegularAttribute { extendedattributes: vec![], inherits: false, name: "attr".tostring(), readonly: true, type: Box::new(Type { extendedattributes: vec![], kind: TypeKind::SignedLong, nullable: true }) }))], name: "Node".tostring() }))]; let mut visitor = PrettyPrintVisitor::new(); visitor.visit(&ast); asserteq!(visitor.getoutput(), "[Attribute]\ninterface Node {\n readonly attribute long? attr;\n};\n\n"); ```

Conformance

The parser is conformant with regards to the WebIDL grammar except for one point. Extended attributes, as described by the grammar, are not supported due to their lack of semantic meaning when parsed. Instead, limited forms are supported (as shown in the table). This parser allows a bit more flexibility when parsing extended attributes of the form A=B. The specification states that A and B must be identifiers, but this parser accepts B as any token. If you would like for any extended attributes to be parsed (essentially any sequences of tokens), please consider looking at #8 to help resolve the problem with doing so.