HTML Parser not working

[struckchure]

I'm building something like an HTML template parser, but I can't it to work.

// ast.go

type Component struct {
	HTML []Node `parser:"@@*"`
}

type Node interface{ node() }

type TextNode struct {
	Value string `parser:"@Text"`
}

func (TextNode) node() {}

type ExpressionNode struct {
	Value string `parser:"LeftCurlyB @@ RightCurlyB"`
}

type ElementNode struct {
	Tag       string      `parser:"LeftAngleB @Ident"`
	Attribute []Attribute `parser:"@@*"`
	SelfClose bool        `parser:"( SelfCloseTag"`
	Children  []Node      `parser:"| RightAngleB @@* CloseTag @Ident RightAngleB )"`
}

func (ElementNode) node() {}

type Attribute struct {
	Name  string `parser:"@Ident"`
	Value string `parser:"Equals @String"`
}

// lexer.go

var componentLexer = lexer.MustSimple([]lexer.SimpleRule{
	{Name: "CloseTag", Pattern: `</`},
	{Name: "SelfCloseTag", Pattern: `/>`},

	{Name: "Float", Pattern: `[-+]?(?:\d+\.\d*|\.\d+)(?:[eE][-+]?\d+)?`},
	{Name: "Int", Pattern: `[-+]?\d+`},

	{Name: "String", Pattern: `"(\\"|[^"])*"`},

	{Name: "Equals", Pattern: `=`},
	{Name: "Comma", Pattern: `,`},
	{Name: "SemiColon", Pattern: `;`},
	{Name: "Colon", Pattern: `:`},
	{Name: "LeftSquareB", Pattern: `\[`},
	{Name: "RightSquareB", Pattern: `\]`},
	{Name: "LeftCurlyB", Pattern: `\{`},
	{Name: "RightCurlyB", Pattern: `\}`},
	{Name: "LeftAngleB", Pattern: `<`},
	{Name: "RightAngleB", Pattern: `>`},

	{Name: "Ident", Pattern: `[a-zA-Z_]\w*`},
	{Name: "Text", Pattern: `[^<>{}]+`},

	{Name: "Whitespace", Pattern: `[ \t\r\n]+`},
})

// parser.go

func Parse(filename, code string) (*Component, error) {
	parser, err := participle.Build[Component](
		participle.Union[Node](TextNode{}, ElementNode{}),
		participle.Elide("Whitespace", "SemiColon"),
		participle.Unquote("String"),
		participle.Lexer(componentLexer),
	)
	if err != nil {
		return nil, err
	}

	return parser.ParseString(filename, code)
}

// main.go

func main() {
	componentFile := "./sample.html"
	code, _ := os.ReadFile(componentFile)

	blocks, err := Parse(componentFile, string(code))
	if err != nil {
		fmt.Println(err)
		return
	}

	fmt.Println(gelt.Prettify(blocks))
}

// sample.html

<p>Hello <b>world</b></p>

I keep getting this error

sample.html:1:4: unexpected token "Hello" (expected <closetag> <ident> <rightangleb>)

[alecthomas]

I can't really tell because I can't reproduce it without the lexer, but this is incorrect:

	Children  []Node      `parser:"| RightAngleB @@* CloseTag @Ident RightAngleB )"`

You can only use @@ here, using @Ident is invalid. You'd need to capture that into a closing tag name.

[struckchure]

I just added the lexer, please check again.

[alecthomas]

The issue is that two of your tokens have overlapping regex patterns: Ident and Text.

You're matching "Text" in TextNode, but Ident has higher precedence in the lexer rules, so Text will never be produced. You can fix this by either changing TextNode to this:

	Value string `parser:"(@Ident | @Text)"`

Or fixing your lexer rules to remove the ambiguity. I'd suggest doing this.

[struckchure]

Thank you!