Merge pull request 'parser: Unquoted(), AcceptQuotedString() and SplitCommaArray' (#9)

Reviewed-on: #9

parser/comma_array.go

@@ -0,0 +1,88 @@
+package parser
+
+import (
+	"strings"
+
+	"asciigoat.org/core/lexer"
+)
+
+type commaArrayParser struct {
+	TextParser
+
+	out []string
+}
+
+func (p *commaArrayParser) lexStart() (lexer.StateFn, error) {
+	for {
+		r, _, err := p.ReadRune()
+		switch {
+		case err != nil:
+			// EOF
+			return nil, err
+		case r == RuneQuotes:
+			// Quoted Value
+			return p.lexQuotedString, nil
+		case IsNewLine(r):
+			// new lines are acceptable when parsing a string for
+			// comma delimited arrays. but make sure we discard it
+			// complete
+			p.UnreadRune()
+			p.AcceptNewLine()
+			p.Discard()
+		case lexer.IsSpace(r):
+			// discard whitespace outside quotes
+			p.Discard()
+		default:
+			p.UnreadRune()
+			return p.lexWord, nil
+		}
+	}
+}
+
+func (p *commaArrayParser) lexWord() (lexer.StateFn, error) {
+	for {
+		r, _, err := p.ReadRune()
+		switch {
+		case err != nil:
+			// done. store what we got and move on
+			_, s := p.Emit()
+			p.out = append(p.out, s)
+			return nil, err
+		case r == ',':
+			// done
+			_, s := p.Emit()
+			// remove comma, trim and append to output
+			s = strings.TrimRightFunc(s[:len(s)-1], IsSpace)
+			p.out = append(p.out, s)
+			return p.lexStart, nil
+		}
+	}
+}
+
+func (p *commaArrayParser) lexQuotedString() (lexer.StateFn, error) {
+	s, err := lexQuotedString(&p.TextParser)
+	if err != nil {
+		return nil, err
+	}
+
+	p.Discard()
+	p.out = append(p.out, s)
+	return p.lexStart, nil
+}
+
+func (p *commaArrayParser) Run() ([]string, error) {
+	err := lexer.Run(p.lexStart)
+
+	return p.out, err
+}
+
+// SplitCommaArray splits comma separated strings, removing whitespace
+// and respecting quoted literals.
+func SplitCommaArray(s string) ([]string, error) {
+	if s != "" {
+		var p commaArrayParser
+		p.InitString(s)
+		return p.Run()
+	}
+	return nil, nil
+}

parser/error.go

@@ -1,6 +1,8 @@
 package parser
 
 import (
+	"io/fs"
+
 	"asciigoat.org/core/lexer"
 )
 
@@ -25,3 +27,17 @@ func ErrPlusPosition(pos lexer.Position, e *lexer.Error) *lexer.Error {
 
 	return NewError(pos, e.Content, e.Hint, e.Err)
 }
+
+// NewErrIncompleteQuotedString returns a [lexer.Error]
+// indicating the quoted string being parsed wasn't correctly
+// terminated
+func NewErrIncompleteQuotedString(p *TextParser) *lexer.Error {
+	return newErrIncomplete(p, "incomplete quoted string")
+}
+
+func newErrIncomplete(p *TextParser, hint string) *lexer.Error {
+	pos, s := p.Emit()
+	pos.Add(GetPositionalLength(s))
+
+	return NewError(pos, s, hint, fs.ErrInvalid)
+}

parser/lexer_runes.go

@@ -7,11 +7,13 @@ import (
 )
 
 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
+	RuneComment      = ';'  // RuneComment is the standard INI comment character
+	RuneCommentExtra = '#'  // RuneCommentExtra is UNIX shell's comment character
+	RuneSectionStart = '['  // RuneSectionStart indicates the start of a section declaration
+	RuneSectionEnd   = ']'  // RuneSectionEnd indicates the end of a section declaration
+	RuneFieldEqual   = '='  // RuneFieldEqual separates field keys from their values
+	RuneQuotes       = '"'  // RuneQuotes indicates the start and end of a quoted value
+	RuneEscape       = '\\' // RuneEscape indicates the next rune is escaped
 )
 
 var (
@@ -27,6 +29,8 @@ var (
 		RuneSectionStart,
 		RuneSectionEnd,
 		RuneFieldEqual,
+		RuneQuotes,
+		RuneEscape,
 	})
 )
 

parser/text_quoted.go

@@ -0,0 +1,97 @@
+package parser
+
+import (
+	"strings"
+
+	"asciigoat.org/core/lexer"
+)
+
+// AcceptQuotedString consumes a quoted string from the source
+// and returns it unquoted and unescaped
+func (p *TextParser) AcceptQuotedString() (string, bool, error) {
+	r, _, err := p.ReadRune()
+	switch {
+	case err != nil:
+		// nothing here
+		return "", false, err
+	case r != RuneQuotes:
+		// not for us
+		p.UnreadRune()
+		return "", false, nil
+	default:
+		// let's roll
+		s, err := lexQuotedString(p)
+		switch {
+		case err != nil:
+			// bad quoted string
+			return "", false, err
+		default:
+			// success
+			return s, true, nil
+		}
+	}
+}
+
+func lexQuotedString(p *TextParser) (string, *lexer.Error) {
+	for {
+		r, _, err := p.ReadRune()
+		switch {
+		case err != nil:
+			// incomplete
+			return "", NewErrIncompleteQuotedString(p)
+		case r == RuneQuotes:
+			// end, remove quotes and process escaped characters
+			return lexReturnUnescapedQuotedString(p)
+		case r == RuneEscape:
+			// escaped, take another
+			_, _, err := p.ReadRune()
+			if err != nil {
+				// incomplete
+				return "", NewErrIncompleteQuotedString(p)
+			}
+		case IsNewLine(r):
+			// new lines within quoted values are acceptable
+			p.UnreadRune()
+			p.AcceptNewLine()
+		default:
+			// continue
+		}
+	}
+}
+
+func lexReturnUnescapedQuotedString(p *TextParser) (string, *lexer.Error) {
+	// remove quotes
+	s := p.String()
+	l := len(s)
+	s = s[1 : l-1]
+
+	if strings.ContainsRune(s, RuneEscape) {
+		// TODO: implement unescaping
+		err := NewError(p.Position(), s, "escaped characters", lexer.ErrNotImplemented)
+		return "", err
+	}
+
+	return s, nil
+}
+
+// Unquoted removes quotes the content and unescapes the content
+func Unquoted(s string) (string, error) {
+	var p TextParser
+	if s == "" {
+		return "", nil
+	}
+
+	p.InitString(s)
+	unquoted, ok, err := p.AcceptQuotedString()
+	switch {
+	case err != nil:
+		// bad string
+		return "", err
+	case ok:
+		// success
+		return unquoted, nil
+	default:
+		// not quoted
+		return s, nil
+	}
+}