add unicode escape sequences

parse/lex.ha

@@ -246,6 +246,9 @@ fn scanchar(lex: *lexer) (rune | error) = {
 		unget(lex, rnn);
 		if (isspace(rnn)) {
 			return rn;
+		} else {
+			if (rn == 'x') {
+				return scanescape2(lex);
 			} else {
 				memio::appendrune(&namebuf, rn)!;
 				memio::concat(&namebuf, scanword(lex)?)!;
@@ -257,6 +260,7 @@ fn scanchar(lex: *lexer) (rune | error) = {
 				};
 				return lex.loc: invalid;
 			};
+		};
 	case io::EOF =>
 		return rn;
 	};
@@ -282,6 +286,56 @@ fn scanescape(lex: *lexer) (rune | error) = {
 	case 'v' => return '\v';
 	case '\\' => return '\\';
 	case '"' => return '"';
+	case 'x' => return scanescape2(lex)?;
+	case =>
+		return lex.loc: invalid;
+	};
+};
+
+fn scanescape2(lex: *lexer) (rune | error) = {
+	// This handles the `\xhh...;` family of escapes.
+	// It's on a separate function since both [[scanescape]] and
+	// [[scanchar]] make use of it. Much like how [[scanescape]] assumes
+	// that the backslash has already been consumed, this one assumes that
+	// the leading character has been consumed prior to entering this
+	// function.
+
+	const rn = match (nextrune(lex)?) {
+	case let rn: rune =>
+		yield rn;
+	case io::EOF =>
+		return ("escape sequence", lex.loc.0, lex.loc.1): unterminated;
+	};
+
+	const buf: [6]u8 = [0...];
+	let buf = memio::fixed(buf);
+	memio::appendrune(&buf, rn)!;
+
+	let count = 1z;
+	for (true) {
+		const rn = match (nextrune(lex)?) {
+		case let rn: rune =>
+			yield rn;
+		case io::EOF =>
+			return ("escape sequence", lex.loc.0, lex.loc.1): unterminated;
+		};
+
+		count += 1;
+
+		if (count > 6) {
+			return lex.loc: invalid;
+		} else if (rn == ';') {
+			break;
+		} else {
+			memio::appendrune(&buf, rn)!;
+		};
+	};
+
+	const buf = memio::string(&buf)!;
+
+	return match (strconv::stou32(buf, strconv::base::HEX)) {
+	case let codepoint: u32 =>
+		return codepoint: rune;
 	case =>
 		return lex.loc: invalid;
 	};