parser: SplitCommaArray [WIP]

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

parser/comma_array.go

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

parser/error.go

@@ -48,15 +48,3 @@ func newErrIncomplete(p *TextParser, hint string) *lexer.Error {
 
 	return NewError(pos, s, hint, fs.ErrInvalid)
 }
-
-// NewErrInvalidEscapeSequence returns a [lexer.Error] indicating
-// the specified sequence, at the end of the accepted buffer,
-// 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/text_quoted.go

@@ -1,13 +1,11 @@
 package parser
 
 import (
-	"strings"
-
 	"asciigoat.org/core/lexer"
 )
 
 // AcceptQuotedString consumes a quoted string from the source
-// and returns it unquoted and unescaped
+// and returns it unquoted.
 func (p *TextParser) AcceptQuotedString() (string, bool, error) {
 	r, _, err := p.ReadRune()
 	switch {
@@ -56,10 +54,9 @@ func lexQuotedStringNoEscape(p *TextParser) (string, bool, *lexer.Error) {
 			// end, just remove the quotes
 			s := p.String()
 			l := len(s)
-			return s[1 : l-1], true, nil
+			return s[1 : l-2], true, nil
 		case r == RuneEscape:
 			// things just got complicated...
-			p.UnreadRune()
 			return "", false, nil
 		case IsNewLine(r):
 			// new lines within quoted values are acceptable
@@ -71,45 +68,8 @@ func lexQuotedStringNoEscape(p *TextParser) (string, bool, *lexer.Error) {
 	}
 }
 
-// Unquoted removes quotes the content and unescapes the content
+func lexQuotedStringEscaped(*TextParser) (string, *lexer.Error) {
-func lexQuotedStringEscaped(p *TextParser) (string, *lexer.Error) {
+	panic("not implemented")
-	var result strings.Builder
-
-	// append what was accepted before the escape character
-	_, _ = result.WriteString(p.String()[1:])
-
-	for {
-		r, _, err := p.ReadRune()
-		switch {
-		case err != nil:
-			// incomplete quoted
-			return "", NewErrIncompleteQuotedString(p)
-		case r == RuneQuotes:
-			// end
-			return result.String(), nil
-		case r == RuneEscape:
-			// 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