Merge pull request 'parser: implement basic dosini parsing' (#2)

Reviewed-on: #2

go.mod

@@ -2,8 +2,6 @@ module asciigoat.org/ini
 
 go 1.19
 
-replace asciigoat.org/core => ../core
-
 require (
 	asciigoat.org/core v0.3.6
 	github.com/mgechev/revive v1.3.3

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

@@ -15,21 +15,30 @@ func (p *Parser) lexStart() (lexer.StateFn, error) {
 		r, _, err := p.src.ReadRune()
 		switch {
 		case err != nil:
-			err = p.emitError("", err)
+			return p.emitError("", err)
-			return nil, err
 		case IsNewLine(r):
-			p.lexNewLine(r)
+			// new line
+			p.lexMoreNewLine(r)
 			p.stepLine()
 		case IsSpace(r):
+			// whitespace
 			p.stepRune()
-		default:
+		case IsCommentStart(r):
+			// switch to comment lexer
 			p.src.UnreadRune()
-			return p.lexToken, nil
+			return p.lexComment, nil
+		case IsSectionStart(r):
+			// section
+			return p.lexSectionStart, nil
+		default:
+			// entry
+			p.src.UnreadRune()
+			return p.lexEntryStart, nil
 		}
 	}
 }
 
-func (p *Parser) lexNewLine(r1 rune) {
+func (p *Parser) lexMoreNewLine(r1 rune) {
 	// r1 is warrantied to be either '\r' or '\n'
 	r2, _, err := p.src.ReadRune()
 	switch r1 {
@@ -58,9 +67,78 @@ func (p *Parser) lexNewLine(r1 rune) {
 	}
 }
 
-func (p *Parser) lexToken() (lexer.StateFn, error) {
+func (p *Parser) lexComment() (lexer.StateFn, error) {
-	p.src.AcceptAll(IsNotSpace)
+	// until the end of the line
+	p.src.AcceptAll(IsNotNewLine)
+
+	err := p.emitString(TokenComment)
+	return p.lexStart, err
+}
+
+func (p *Parser) lexSectionStart() (lexer.StateFn, error) {
+	if err := p.emitString(TokenSectionStart); err != nil {
+		return nil, err
+	}
+
+	// remove whitespace between `[` and the name
+	if p.src.AcceptAll(IsSpaceNotNewLine) {
+		p.stepString()
+	}
+
+	if !p.src.AcceptAll(IsName) {
+		// no name
+		return p.emitError("section name missing", lexer.ErrUnacceptableRune)
+	}
+
+	if err := p.emitString(TokenSectionName); err != nil {
+		return nil, err
+	}
+
+	// remove whitespace between the name andthe closing `]`
+	if p.src.AcceptAll(IsSpaceNotNewLine) {
+		p.stepString()
+	}
+
+	r, _, err := p.src.ReadRune()
+	switch {
+	case err != nil:
+		return p.emitError("", err)
+	case IsSectionEnd(r):
+		err := p.emitString(TokenSectionEnd)
+		return p.lexStart, err
+	default:
+		return p.emitInvalidRune(r)
+	}
+}
+
+func (p *Parser) lexEntryStart() (lexer.StateFn, error) {
+	p.src.AcceptAll(IsName)
+	if err := p.emitString(TokenFieldKey); err != nil {
+		return nil, err
+	}
+
+	// ignore whitespace between key and the '=' sign
+	if p.src.AcceptAll(IsSpaceNotNewLine) {
+		p.stepString()
+	}
+
+	r, _, err := p.src.ReadRune()
+	switch {
+	case err != nil:
+		return p.emitError("", err)
+	case r != RuneFieldEqual:
+		return p.emitInvalidRune(r)
+	}
+
+	// ignore whitespace between the '=' and the value
+	if p.src.AcceptAll(IsSpaceNotNewLine) {
+		p.stepString()
+	}
+
+	p.src.AcceptAll(IsNotNewLine)
+	if err := p.emitString(TokenFieldValue); err != nil {
+		return nil, err
+	}
 
-	err := p.emitString(TokenUnknown)
 	return p.lexStart, err
 }

parser/lexer_runes.go

@@ -1,9 +1,35 @@
 package parser
 
 import (
+	"strings"
+
 	"asciigoat.org/core/lexer"
 )
 
+const (
+	RuneComment      = ';' // RuneComment is the standard dosini comment character
+	RuneCommentExtra = '#' // RuneCommentExtra is UNIX shell's comment character
+	RuneSectionStart = '[' // RuneSectionStart indicates the start of a section declaration
+	RuneSectionEnd   = ']' // RuneSectionEnd indiciates the end of a section declaration
+	RuneFieldEqual   = '=' // RuneFieldEqual separates field keys from their values
+)
+
+var (
+	// RunesComment is a string containing all runes acceptable to start comments
+	RunesComment = string([]rune{
+		RuneComment,
+		RuneCommentExtra,
+	})
+	// RunesSpecial is a string containing all the runes with special meaning
+	RunesSpecial = string([]rune{
+		RuneComment,
+		RuneCommentExtra,
+		RuneSectionStart,
+		RuneSectionEnd,
+		RuneFieldEqual,
+	})
+)
+
 var (
 	// IsNewLine tells if the rune indicates a line break or the start of one
 	IsNewLine = lexer.NewIsIn("\r\n")
@@ -13,4 +39,29 @@ var (
 	IsSpace = lexer.IsSpace
 	// IsNotSpace tells if the rune is not considered whitespace by Unicode
 	IsNotSpace = lexer.NewIsNot(IsSpace)
+	// IsCommentStart ...
+	IsCommentStart = lexer.NewIsIn(RunesComment)
 )
+
+// IsSpaceNotNewLine indicates a rune is whitespace but not a new line
+func IsSpaceNotNewLine(r rune) bool {
+	return IsSpace(r) && !IsNewLine(r)
+}
+
+// IsSectionStart indicates the rune starts the section declaration
+func IsSectionStart(r rune) bool { return r == RuneSectionStart }
+
+// IsSectionEnd indicates the rune ends the section declaration
+func IsSectionEnd(r rune) bool { return r == RuneSectionEnd }
+
+// IsName indicates a rune is acceptable for section or field names
+func IsName(r rune) bool {
+	switch {
+	case IsSpace(r):
+		return false
+	case strings.ContainsRune(RunesSpecial, r):
+		return false
+	default:
+		return true
+	}
+}

parser/parser.go

@@ -51,27 +51,29 @@ func (p *Parser) setDefaults() {
 }
 
 func (p *Parser) emitString(typ TokenType) error {
-	var err error
-
 	s := p.src.Emit()
-	err = p.OnToken(p.pos, typ, s)
+	err := p.OnToken(p.pos, typ, s)
 	p.pos.StepN(len(s))
 
 	return err
 }
 
-func (p *Parser) emitError(content string, err error) error {
+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 err2
+		return nil, err2
 	default:
 		// return original error
-		return err
+		return nil, err
 	}
 }
 
+func (p *Parser) emitInvalidRune(r rune) (lexer.StateFn, error) {
+	return p.emitError(string([]rune{r}), lexer.ErrUnacceptableRune)
+}
+
 // stepLine discards the data and moves the position
 // to the next line.
 func (p *Parser) stepLine() {