2 ms·
(Author here, my wife posted this thread; it has no bearing on the question but I am told it is good to disclaim.) You could quibble about what counts but the
by nyuutsu 2mo ago
(Author here, my wife posted this thread; it has no bearing on the question but I am told it is good to disclaim.)
You could quibble about what counts but the core is "yeah".
Ppf3-with-undo stores literal "hey, we expect byte n to be 'foo'. on apply, replace it with 'bar'. on undo, put 'foo' back. remember: the copyrighted byte is 'foo'". Touch every byte in some way and your undo data is the entire original.
Some patches literally touch one byte. Their undo is, indirectly or indirectly, a recipe for retrieving one byte of copyrighted data. Is that fragmentary enough that it is entirely in the clear? I have no idea.
- gblargg 2mo agoFor some reason an XOR scheme escaped me. It's the perfect way, doesn't take any extra space over the size of the modified regions, and avoids even storing the copyrighted work of the ROM patch author. You could distribute the patch without anyone's permission (maybe)! Just include a strong checksum of the affected regions to be sure you're starting with the correct version, and checksums of the resulting data to give the same guarantee for reversing the process.