I've attempted to parse some SGFs produced by LittleGolem for various games at large board sizes.

Eg https://www.littlegolem.net/jsp/game/game.jsp?gid=2205068, its sgf being at https://www.littlegolem.net/servlet/sgf/2205068/game2205068.sgf

As of the time of writing this comment, the first few moves look like this: “;B[?d];W[??];B[d?];W[dd];B[?c];W[?p]“. That's not just not standard SGF, I believe it's actively losing information: the first two moves aren't on the same line (in either direction).

You might think I'm using a bad text editor and it has mangled the text, but no. All the question marks in this snippet are actually 0x3F, which is an actual ascii question mark. This isn't a display issue on my side.

1. Rio, if you're reading this and not too busy, would you mind taking a look?

2. Is there any other way to load the list of moves for the game? There's the “Moves” box on the game page, but for large sized boards with many moves, this list gets interspersed with “…“, which further complicates downloading the list of moves. Any format is fine, but I'd like to get the whole game in one http request :)

Are you sure your text editor isn't mangling it? When I download it I get bytes above 0x7f where you're seeing ?'s.

I also get values in the range starting at 0x80.

With either curl or download.

Ahaha, you're of course right! With hexedit it's correct. In my defense, this is the first time $EDITOR is changing the characters' hex codes right under my nose. Off to learn $EDITOR and find out what's happening…

I wish there was a way to “mark as resolved” or close this topic :)