first commit

2026-08-09 18:03:09 +01:00
commit d7ca591b76
169 changed files with 51272 additions and 0 deletions
@@ -0,0 +1,131 @@
+// Package sieve implements a Sieve (RFC 5228) interpreter covering the
+// common mail-filtering subset: header tests (:contains, :is), if/elsif/else,
+// and the fileinto/discard/keep/stop actions. Not full RFC 5228 — no
+// extensions (vacation, reject, notify), no envelope/size/address tests,
+// no allof/anyof boolean combinators. This covers what real users actually
+// write for "move mail matching X to folder Y" / "discard mail from Z",
+// which is the overwhelming majority of real-world Sieve scripts; broader
+// grammar support is a natural follow-up once client compatibility testing
+// calls for it.
+package sieve
+
+import (
+	"fmt"
+	"strings"
+	"unicode"
+)
+
+type tokenKind int
+
+const (
+	tokIdent tokenKind = iota
+	tokString
+	tokTag // :contains, :is, etc.
+	tokSemicolon
+	tokLBrace
+	tokRBrace
+	tokEOF
+)
+
+type token struct {
+	kind  tokenKind
+	value string
+}
+
+type lexer struct {
+	input []rune
+	pos   int
+}
+
+func newLexer(script string) *lexer {
+	return &lexer{input: []rune(script)}
+}
+
+func (l *lexer) next() (token, error) {
+	l.skipWhitespaceAndComments()
+	if l.pos >= len(l.input) {
+		return token{kind: tokEOF}, nil
+	}
+
+	c := l.input[l.pos]
+	switch {
+	case c == ';':
+		l.pos++
+		return token{kind: tokSemicolon, value: ";"}, nil
+	case c == '{':
+		l.pos++
+		return token{kind: tokLBrace, value: "{"}, nil
+	case c == '}':
+		l.pos++
+		return token{kind: tokRBrace, value: "}"}, nil
+	case c == '"':
+		return l.readString()
+	case c == ':':
+		return l.readTag()
+	case unicode.IsLetter(c):
+		return l.readIdent()
+	default:
+		return token{}, fmt.Errorf("unexpected character %q at position %d", c, l.pos)
+	}
+}
+
+func (l *lexer) skipWhitespaceAndComments() {
+	for l.pos < len(l.input) {
+		c := l.input[l.pos]
+		if unicode.IsSpace(c) {
+			l.pos++
+			continue
+		}
+		// Single-line comment: # ... end of line
+		if c == '#' {
+			for l.pos < len(l.input) && l.input[l.pos] != '\n' {
+				l.pos++
+			}
+			continue
+		}
+		// Bracketed comment: /* ... */
+		if c == '/' && l.pos+1 < len(l.input) && l.input[l.pos+1] == '*' {
+			l.pos += 2
+			for l.pos+1 < len(l.input) && !(l.input[l.pos] == '*' && l.input[l.pos+1] == '/') {
+				l.pos++
+			}
+			l.pos += 2
+			continue
+		}
+		break
+	}
+}
+
+func (l *lexer) readString() (token, error) {
+	l.pos++ // skip opening quote
+	var sb strings.Builder
+	for l.pos < len(l.input) && l.input[l.pos] != '"' {
+		if l.input[l.pos] == '\\' && l.pos+1 < len(l.input) {
+			l.pos++
+		}
+		sb.WriteRune(l.input[l.pos])
+		l.pos++
+	}
+	if l.pos >= len(l.input) {
+		return token{}, fmt.Errorf("unterminated string literal")
+	}
+	l.pos++ // skip closing quote
+	return token{kind: tokString, value: sb.String()}, nil
+}
+
+func (l *lexer) readTag() (token, error) {
+	start := l.pos
+	l.pos++ // skip ':'
+	for l.pos < len(l.input) && (unicode.IsLetter(l.input[l.pos]) || l.input[l.pos] == '-') {
+		l.pos++
+	}
+	return token{kind: tokTag, value: string(l.input[start:l.pos])}, nil
+}
+
+func (l *lexer) readIdent() (token, error) {
+	start := l.pos
+	for l.pos < len(l.input) && (unicode.IsLetter(l.input[l.pos]) || unicode.IsDigit(l.input[l.pos]) || l.input[l.pos] == '_') {
+		l.pos++
+	}
+	return token{kind: tokIdent, value: string(l.input[start:l.pos])}, nil
+}