feat(asm): symbol-valued DATA, division slash in symbols and plain semicolons

Assisted-by: GLM 5.3 Flash

parser/parser.go

@@ -32,9 +32,8 @@ func (e Error) Error() string {
 // returned file is usable even when errors is non-empty.
 func Parse(path, src string) (*ast.File, []error) {
 	tokens := lexer.Tokenize(src)
-	lines := splitLines(tokens)
 	p := &state{path: path}
-	p.parse(lines)
+	p.parse(statementLines(tokens))
 	return p.file, p.errs
 }
 
@@ -73,6 +72,46 @@ func splitLines(tokens []token.Token) [][]token.Token {
 	return lines
 }
 
+// statementLines turns the token stream into the logical lines the parser
+// reads: physical lines split at the ';' statement separators, exactly the
+// way the expansion path treats the expanded bodies.  The runtime writes
+// "ROLQ $3, DI; ROLQ $13, DI" and "REP; MOVSB" in plain files, and the
+// separator carries no meaning beyond the break.  Comments are statement
+// text, not structure: the lexer delivers a whole comment as one token, so
+// a ';' inside a comment is never a separator; a comment after a statement
+// stays on that statement's line; and a comment that sits between
+// statements (the runtime's "NO_LOCAL_POINTERS; /* … */" style) stands as
+// its own logical line, like a whole-line comment.
+func statementLines(tokens []token.Token) [][]token.Token {
+	var out [][]token.Token
+	var cur []token.Token
+	flush := func() {
+		if len(cur) > 0 {
+			out = append(out, cur)
+			cur = nil
+		}
+	}
+	for _, t := range tokens {
+		switch t.Kind {
+		case token.EOF:
+			// The stream's terminator is not statement content.
+		case token.Newline, token.Semicolon:
+			flush()
+		case token.Comment:
+			if len(cur) > 0 {
+				cur = append(cur, t)
+			} else {
+				out = append(out, []token.Token{t})
+			}
+			flush()
+		default:
+			cur = append(cur, t)
+		}
+	}
+	flush()
+	return out
+}
+
 func (p *state) parse(lines [][]token.Token) {
 	p.file = &ast.File{Path: p.path, Macros: map[string]bool{}}
 	for _, line := range lines {