Save files may contain invalid UTF-8

[robojumper]

#26 provided a
persist.map.json that contains the null-terminated byte sequence 72 6F 6F 80 00 as an object name, presumably a room name. Room names are generated by taking the string roo and appending a single ASCII character to it. The game seems to have run out of room names by going outside of the ASCII character range with the 80. This makes the object name invalid UTF-8 and both implementations reject it.

I'm inclined to not fix this. This could possibly be be fixed in the Java impl here

DarkestDungeonSaveEditor/src/main/java/de/robojumper/ddsavereader/file/DsonFile.java

Lines 205 to 216 in f580c05

	static String readName(byte[] data, int start, int len) throws ParseException {
	// Field names can be UTF-8
	byte[] str = Arrays.copyOfRange(data, start, start + len);
	String name = new String(str, StandardCharsets.UTF_8);
	if (!Arrays.equals(name.getBytes(StandardCharsets.UTF_8), str) || data[start + len] != 0) {
	throw new ParseException(
	String.format("%d: Wrong name length: Name %s, expected %d but has null bytes in wrong place",
	start, name, len),
	start);
	}
	return name;
	}

and in the Rust impl here

DarkestDungeonSaveEditor/rust/ddsaveedit/ddsavelib/src/file/bin.rs

Lines 450 to 453 in f580c05

	let name = {
	let cs = CStr::from_bytes_with_nul(&field_name)?.to_str()?;
	NameType::from(cs)
	};

and here (need support for byte escape sequences)

DarkestDungeonSaveEditor/rust/ddsaveedit/ddsavelib/src/util.rs

Lines 10 to 67 in f580c05

	pub fn escape(arg: &str) -> Cow<str> {
	if arg
	.chars()
	.any(|c| matches!(c, '\x08' | '\x0C' | '\n' | '\r' | '\t' | '\\' | '"'))
	{
	let mut s = String::new();
	for c in arg.chars() {
	match c {
	'\x08' => s.push_str("\\b"),
	'\x0C' => s.push_str("\\f"),
	'\n' => s.push_str("\\n"),
	'\r' => s.push_str("\\r"),
	'\t' => s.push_str("\\t"),
	'"' => s.push_str("\\\""),
	'\\' => s.push_str("\\\\"),
	_ => s.push(c),
	}
	}
	Cow::Owned(s)
	} else {
	Cow::Borrowed(arg)
	}
	}
	pub fn unescape(arg: &str) -> Option<Cow<str>> {
	// Bare control characters are disallowed
	if arg
	.chars()
	.any(|c| matches!(c, '\x08' | '\x0C' | '\n' | '\r' | '\t'))
	{
	return None;
	}
	if arg.find('\\').is_some() {
	let mut s = String::new();
	let mut it = arg.chars();
	while let Some(c) = it.next() {
	match c {
	'\\' => {}
	c => {
	s.push(c);
	continue;
	}
	}
	s.push(match it.next() {
	Some('b') => '\x08',
	Some('f') => '\x0C',
	Some('n') => '\n',
	Some('r') => '\r',
	Some('t') => '\t',
	Some('"') => '\"',
	_ => return None,
	});
	}
	return Some(Cow::Owned(s));
	}
	Some(Cow::Borrowed(arg))
	}