parser: add initial Parser emitting non-whitespace tokens

Signed-off-by: Alejandro Mery <amery@jpi.io>

go.mod

@@ -3,6 +3,7 @@ module asciigoat.org/ini
 go 1.19
 
 require (
+	asciigoat.org/core v0.3.6
 	github.com/mgechev/revive v1.3.3
 	golang.org/x/tools v0.12.0
 )

go.sum

@@ -1,3 +1,5 @@
+asciigoat.org/core v0.3.6 h1:b1vL090OxylmSOwLQryjrmC8FhhCtktMyeJSy1e6LwI=
+asciigoat.org/core v0.3.6/go.mod h1:tXj+JUutxRbcO40ZQRuUVaZ4rnYz1kAZ0nblisV8u74=
 github.com/BurntSushi/toml v1.3.2 h1:o7IhLm0Msx3BaB+n3Ag7L8EVlByGnpq14C4YWiu/gL8=
 github.com/BurntSushi/toml v1.3.2/go.mod h1:CxXYINrC8qIiEnFrOxCa7Jy5BFHlXnUU2pbicEuybxQ=
 github.com/chavacava/garif v0.0.0-20230608123814-4bd63c2919ab h1:5JxePczlyGAtj6R1MUEFZ/UFud6FfsOejq7xLC2ZIb0=

parser/lexer.go

@@ -0,0 +1,67 @@
+package parser
+
+import "asciigoat.org/core/lexer"
+
+// Run parses the source
+func (p *Parser) Run() error {
+	p.setDefaults()
+	p.pos.Reset()
+
+	return lexer.Run(p.lexStart)
+}
+
+func (p *Parser) lexStart() (lexer.StateFn, error) {
+	for {
+		r, _, err := p.src.ReadRune()
+		switch {
+		case err != nil:
+			return p.emitError("", err)
+		case IsNewLine(r):
+			// new line
+			p.lexMoreNewLine(r)
+			p.stepLine()
+		case IsSpace(r):
+			// whitespace
+			p.stepRune()
+		default:
+			p.src.UnreadRune()
+			return p.lexToken, nil
+		}
+	}
+}
+
+func (p *Parser) lexMoreNewLine(r1 rune) {
+	// r1 is warrantied to be either '\r' or '\n'
+	r2, _, err := p.src.ReadRune()
+	switch r1 {
+	case '\n':
+		switch {
+		case r2 == '\r':
+			// LN CR
+		case err == nil:
+			// LN
+			p.src.UnreadRune()
+		default:
+			// LN EOF
+		}
+	case '\r':
+		switch {
+		case r2 == '\n':
+			// CR LN
+		case err == nil:
+			// CR
+			p.src.UnreadRune()
+		default:
+			// CR EOF
+		}
+	default:
+		panic("unreachable")
+	}
+}
+
+func (p *Parser) lexToken() (lexer.StateFn, error) {
+	p.src.AcceptAll(IsNotSpace)
+
+	err := p.emitString(TokenUnknown)
+	return p.lexStart, err
+}

parser/lexer_runes.go

@@ -0,0 +1,16 @@
+package parser
+
+import (
+	"asciigoat.org/core/lexer"
+)
+
+var (
+	// IsNewLine tells if the rune indicates a line break or the start of one
+	IsNewLine = lexer.NewIsIn("\r\n")
+	// IsNotNewLine tells if the rune is not a line break nor the start of one
+	IsNotNewLine = lexer.NewIsNot(IsNewLine)
+	// IsSpace tells if the rune is considered whitespace by Unicode
+	IsSpace = lexer.IsSpace
+	// IsNotSpace tells if the rune is not considered whitespace by Unicode
+	IsNotSpace = lexer.NewIsNot(IsSpace)
+)

parser/parser.go

@@ -0,0 +1,105 @@
+// Package parser parses dosini-style files
+package parser
+
+import (
+	"io"
+	"log"
+
+	"asciigoat.org/core/lexer"
+)
+
+// Parser parses a dosini-style document
+type Parser struct {
+	src *lexer.Reader
+	pos lexer.Position
+
+	// OnToken is called for each identified token. if it returns an error
+	// parsing is interrupted.
+	OnToken func(pos lexer.Position, typ TokenType, value string) error
+
+	// OnError is called in case of a parsing error, and it's allowed
+	// to replace the error returned by [Parser.Run].
+	// OnError is called for io.EOF, but [Parser.Run] will consider it
+	// normal termination.
+	OnError func(pos lexer.Position, content string, err error) error
+}
+
+func defaultOnToken(pos lexer.Position, typ TokenType, value string) error {
+	log.Printf("%s:%v:%v: %q", typ, pos.Line, pos.Column, value)
+	return nil
+}
+
+func defaultOnError(pos lexer.Position, content string, err error) error {
+	log.Printf("%s:%v:%v: %q: %s", "error", pos.Line, pos.Column, content, err)
+
+	return lexer.Error{
+		Line:   pos.Line,
+		Column: pos.Column,
+
+		Content: content,
+		Err:     err,
+	}
+}
+
+func (p *Parser) setDefaults() {
+	if p.OnToken == nil {
+		p.OnToken = defaultOnToken
+	}
+	if p.OnError == nil {
+		p.OnError = defaultOnError
+	}
+}
+
+func (p *Parser) emitString(typ TokenType) error {
+	s := p.src.Emit()
+	err := p.OnToken(p.pos, typ, s)
+	p.pos.StepN(len(s))
+
+	return err
+}
+
+func (p *Parser) emitError(content string, err error) (lexer.StateFn, error) {
+	err2 := p.OnError(p.pos, content, err)
+	switch {
+	case err2 != nil:
+		// return wrapped error
+		return nil, err2
+	default:
+		// return original error
+		return nil, err
+	}
+}
+
+// stepLine discards the data and moves the position
+// to the next line.
+func (p *Parser) stepLine() {
+	p.src.Discard()
+	p.pos.StepLine()
+}
+
+// stepRune discards the data and moves the position
+// one rune forward on the same line.
+func (p *Parser) stepRune() {
+	p.src.Discard()
+	p.pos.Step()
+}
+
+// stepString discards the data and moves the position
+// forward on the same line the length of the discarded
+// content.
+func (p *Parser) stepString() {
+	s := p.src.Emit()
+	p.pos.StepN(len(s))
+}
+
+// NewParser creates a dosini-style parser using
+// an [io.Reader] as source
+func NewParser(r io.Reader) *Parser {
+	if r == nil {
+		return nil
+	}
+
+	return &Parser{
+		src: lexer.NewReader(r),
+	}
+}

parser/token.go

@@ -0,0 +1,31 @@
+package parser
+
+//go:generate go run golang.org/x/tools/cmd/stringer -type=TokenType
+
+// A TokenType is a type of Token
+type TokenType uint
+
+const (
+	// TokenUnknown represents a Token that hasn't been identified
+	TokenUnknown TokenType = iota
+	// TokenSectionStart indicates the opening marker of a section declaration.
+	// The left squared bracket.
+	TokenSectionStart
+	// TokenSectionEnd indicates the closing marker of a section declaration.
+	// The right squared bracket.
+	TokenSectionEnd
+	// TokenSectionName represents the section name between the squared brackets
+	TokenSectionName
+	// TokenSectionSubname represents a secondary name in the section represented
+	// between quotes after the section name.
+	// e.g.
+	// [section_name "section_subname"]
+	TokenSectionSubname
+	// TokenComment represents a comment, including the initial ';' or '#' until
+	// the end of the line.
+	TokenComment
+	// TokenFieldKey represents a field name in a `key = value` entry
+	TokenFieldKey
+	// TokenFieldValue represents a field value in a `key = value` entry
+	TokenFieldValue
+)

parser/tokentype_string.go

@@ -0,0 +1,30 @@
+// Code generated by "stringer -type=TokenType"; DO NOT EDIT.
+
+package parser
+
+import "strconv"
+
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[TokenUnknown-0]
+	_ = x[TokenSectionStart-1]
+	_ = x[TokenSectionEnd-2]
+	_ = x[TokenSectionName-3]
+	_ = x[TokenSectionSubname-4]
+	_ = x[TokenComment-5]
+	_ = x[TokenFieldKey-6]
+	_ = x[TokenFieldValue-7]
+}
+
+const _TokenType_name = "TokenUnknownTokenSectionStartTokenSectionEndTokenSectionNameTokenSectionSubnameTokenCommentTokenFieldKeyTokenFieldValue"
+
+var _TokenType_index = [...]uint8{0, 12, 29, 44, 60, 79, 91, 104, 119}
+
+func (i TokenType) String() string {
+	if i >= TokenType(len(_TokenType_index)-1) {
+		return "TokenType(" + strconv.FormatInt(int64(i), 10) + ")"
+	}
+	return _TokenType_name[_TokenType_index[i]:_TokenType_index[i+1]]
+}