Hi,
Thanks for the reply. I underestimated the Edit Book feature before. Now
I see it is genius. I was simply testing it before I use it. :)
On 10/24/25 14:35, h.g.muller wrote:
There indeed is a bit of a bug here; the loop that adds the moves of a
game to the book terminates one move early, by testing i<forwardMostMove
instead of <=. As a result the last (half)move of a game is ignored.
Ok. Good to know. Workaround is to bump the -bookDepth, yes?
In
your test-case the second (half)move of each game is also ignored,
because these are black moves, and black lost. A Polyglot book only
counts the wins.
Excellent, thanks. I forgot to set Result "1/2-1/2". I don't need to go
back and look at that.
As to entering moves already in the book: yes, this could be considered
a bit sloppy, but why would one ever do that?
Human error. Mouse slip. Inattention. Not enough coffee.
When I am testing software I like to see what happens if I do something
stupid. I am used to writing software for business users where if they
do something stupid it is a loss for the business, and my manager would
blame both them and me. So I test for stupidity.
And if we want XBoard to > flag it, what should be done instead? Refuse
the move with an error
message?
It's what I hope for. I can fix it manually so it's not tragic to put
the duplicate in there. But clicking OK on a dialog is simpler. Nice to
put the move SAN in the dialog, because if I am stupid enough to put a
duplicate in there, I might be even more stupid and not know which move
I was doing, say working at 3:00 AM.
--
Alan
