parser: Unescaped [WIP]

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

basic/error.go

@@ -4,28 +4,19 @@ import (
 	"errors"
 
 	"asciigoat.org/core/lexer"
+	"asciigoat.org/ini/parser"
 )
 
 var (
 	errInvalidToken = errors.New("invalid token")
 )
 
-func newError(pos lexer.Position, content, hint string, err error) *lexer.Error {
-	return &lexer.Error{
-		Line:    pos.Line,
-		Column:  pos.Column,
-		Content: content,
-		Hint:    hint,
-		Err:     err,
-	}
-}
-
 func newErrInvalidToken(t *token) *lexer.Error {
-	return newError(t.pos, t.value, "", errInvalidToken)
+	return parser.NewError(t.pos, t.value, "", errInvalidToken)
 }
 
 func (dec *decoder) OnError(pos lexer.Position, content string, err error) error {
-	err = newError(pos, content, "", err)
+	err = parser.NewError(pos, content, "", err)
 	dec.executeFinal()
 	return err
 }

parser/comma_array.go

@@ -1,6 +1,8 @@
 package parser
 
 import (
+	"strings"
+
 	"asciigoat.org/core/lexer"
 )
 
@@ -46,18 +48,13 @@ func (p *commaArrayParser) lexWord() (lexer.StateFn, error) {
 			_, s := p.Emit()
 			p.out = append(p.out, s)
 			return nil, err
-		case r == RuneEscape:
+		case r == ',':
-			// escaped
+			// done
-			r2, _, err := p.ReadRune()
+			_, s := p.Emit()
-			switch {
+			// remove comma, trim and append to output
-			case err != nil:
+			s = strings.TrimRightFunc(s[:len(s)-1], IsSpace)
-				// incomplete
+			p.out = append(p.out, s)
-				return nil, NewErrIncompleteEscaped(&p.TextParser)
+			return p.lexStart, nil
-			case IsNewLine(r2):
-				// escaped new line
-				p.UnreadRune()
-				p.AcceptNewLine()
-			}
 		}
 	}
 }
@@ -79,7 +76,8 @@ func (p *commaArrayParser) Run() ([]string, error) {
 	return p.out, err
 }
 
-// SplitCommaArray splits
+// SplitCommaArray splits comma separated strings, removing whitespace
+// and respecting quoted literals.
 func SplitCommaArray(s string) ([]string, error) {
 	if s != "" {
 		var p commaArrayParser

parser/error.go

@@ -6,6 +6,17 @@ import (
 	"asciigoat.org/core/lexer"
 )
 
+// NewError creates a lexer.Error using a lexer.Position
+func NewError(pos lexer.Position, content, hint string, err error) *lexer.Error {
+	return &lexer.Error{
+		Line:    pos.Line,
+		Column:  pos.Column,
+		Content: content,
+		Hint:    hint,
+		Err:     err,
+	}
+}
+
 // ErrPlusPosition returns a copy of the given [lexer.Error]
 // offsetting the Line/Column information.
 func ErrPlusPosition(pos lexer.Position, e *lexer.Error) *lexer.Error {
@@ -14,13 +25,7 @@ func ErrPlusPosition(pos lexer.Position, e *lexer.Error) *lexer.Error {
 		Column: e.Column,
 	})
 
-	return &lexer.Error{
+	return NewError(pos, e.Content, e.Hint, e.Err)
-		Line:    pos.Line,
-		Column:  pos.Column,
-		Content: e.Content,
-		Hint:    e.Hint,
-		Err:     e.Err,
-	}
 }
 
 // NewErrIncompleteQuotedString returns a [lexer.Error]
@@ -41,11 +46,17 @@ func newErrIncomplete(p *TextParser, hint string) *lexer.Error {
 	pos, s := p.Emit()
 	pos.Add(GetPositionalLength(s))
 
-	return &lexer.Error{
+	return NewError(pos, s, hint, fs.ErrInvalid)
-		Line:    pos.Line,
+}
-		Column:  pos.Column,
+
-		Content: s,
+// NewErrInvalidEscapeSequence returns a [lexer.Error] indicating
-		Hint:    hint,
+// the specified sequence, at the end of the accepted buffer,
-		Err:     fs.ErrInvalid,
+// is invalid
-	}
+func NewErrInvalidEscapeSequence(p *TextParser, seq string) *lexer.Error {
+	pos, s := p.Position(), p.String()
+
+	s = s[:len(s)-len(seq)]
+	pos.Add(GetPositionalLength(s))
+
+	return NewError(pos, seq, "invalid escape character", fs.ErrInvalid)
 }

parser/parser.go

@@ -31,13 +31,7 @@ func defaultOnToken(pos lexer.Position, typ TokenType, value string) error {
 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{
+	return NewError(pos, content, "", err)
-		Line:   pos.Line,
-		Column: pos.Column,
-
-		Content: content,
-		Err:     err,
-	}
 }
 
 func (p *Parser) setDefaults() {

parser/text_quoted.go

@@ -1,11 +1,13 @@
 package parser
 
 import (
+	"strings"
+
 	"asciigoat.org/core/lexer"
 )
 
 // AcceptQuotedString consumes a quoted string from the source
-// and returns it unquoted.
+// and returns it unquoted and unescaped
 func (p *TextParser) AcceptQuotedString() (string, bool, error) {
 	r, _, err := p.ReadRune()
 	switch {
@@ -54,9 +56,10 @@ func lexQuotedStringNoEscape(p *TextParser) (string, bool, *lexer.Error) {
 			// end, just remove the quotes
 			s := p.String()
 			l := len(s)
-			return s[1 : l-2], true, nil
+			return s[1 : l-1], true, nil
 		case r == RuneEscape:
 			// things just got complicated...
+			p.UnreadRune()
 			return "", false, nil
 		case IsNewLine(r):
 			// new lines within quoted values are acceptable
@@ -68,26 +71,46 @@ func lexQuotedStringNoEscape(p *TextParser) (string, bool, *lexer.Error) {
 	}
 }
 
-func lexQuotedStringEscaped(*TextParser) (string, *lexer.Error) {
+// Unquoted removes quotes the content and unescapes the content
-	return "", nil
+func lexQuotedStringEscaped(p *TextParser) (string, *lexer.Error) {
-}
+	var result strings.Builder
 
-//	// escaped. append partial
+	// append what was accepted before the escape character
-//	mark = lexQuotedAppendPartial(p, &buf, mark)
+	_, _ = result.WriteString(p.String()[1:])
-//
+
-//	r2, _, err := p.ReadRune()
+	for {
-//	switch {
+		r, _, err := p.ReadRune()
-//	case err != nil:
+		switch {
-//		// incomplete
+		case err != nil:
-//		return "", NewErrIncompleteQuotedString(p)
+			// incomplete quoted
-//	case IsNewLine(r2):
+			return "", NewErrIncompleteQuotedString(p)
-//		// escaped new line
+		case r == RuneQuotes:
-//		p.UnreadRune()
+			// end
-//		p.AcceptNewLine()
+			return result.String(), nil
-//		mark = lexQuotedAppendNewLine(p, &buf, mark)
+		case r == RuneEscape:
-//	default:
+			// escaped
-//		// bad escaped
+			r2, _, err := p.ReadRune()
-//	}
+			switch {
+			case err != nil:
+				// incomplete escaped
+				return "", NewErrIncompleteEscaped(p)
+			case IsNewLine(r2):
+				// escaped new line, skip
+				p.UnreadRune()
+				p.AcceptNewLine()
+			default:
+				// TODO: check valid escape character and
+				// append to result
+				s := string([]rune{r, r2})
+				err := NewErrInvalidEscapeSequence(p, s)
+				return "", err
+			}
+		default:
+			// normal, append to result
+			_, _ = result.WriteRune(r)
+		}
+	}
+}
 
 // Unquoted removes quotes and unescapes the content
 func Unquoted(s string) (string, error) {