On-line engine blitz tourney March

[Henk]

Ridiculous. Now already the sequence below makes Skipper play a move with white. When issuing the time command

Code: Select all

new
c2c4
go
time 11795

Should be easy to fix. Why am I posting this? Because it is ridiculous. Hardly changed anything today.
So this 'feature' should have been there for years?? Just ignore this post.

By the way ny eyes getting worse.

Probably even the smallest change giving strange features.

[Henk]

Not simple at all when you forget or do not know that black's first move starts after new <newline> <move> <newline>.
Now i know so should be simpel. Better not predict for I am often wrong. ChatGPT saved my day. Have not found myself a clear discription of the winboard protocol that I can translate easily into software. So I use chatGPT.

Better not post here for nobody interested.

[Henk]

My version of the winboard protocal handling was different from what was given.
Why? Because I wanted to run two engines playing together without winboard using a unit test.
Looks like that is too complicated for me for I forgot how it worked. So I should simplify it using only one engine/robot/search object
or whatever it is. But looks like I have to start all over again with the winboard driver/program/console ui or what is its name

Don't know whether I will finish it at all. Already taking too much time.

[flok]

My version of the winboard protocal handling was different from what was given.
Why? Because I wanted to run two engines playing together without winboard using a unit test.
Looks like that is too complicated for me for I forgot how it worked. So I should simplify it using only one engine/robot/search object
or whatever it is. But looks like I have to start all over again with the winboard driver/program/console ui or what is its name

Don't know whether I will finish it at all. Already taking too much time.

Maybe you should try UCI instead: much easier.

[Henk]

My version of the winboard protocal handling was different from what was given.
Why? Because I wanted to run two engines playing together without winboard using a unit test.
Looks like that is too complicated for me for I forgot how it worked. So I should simplify it using only one engine/robot/search object
or whatever it is. But looks like I have to start all over again with the winboard driver/program/console ui or what is its name

Don't know whether I will finish it at all. Already taking too much time.

Maybe you should try UCI instead: much easier.

There is an UML statechart for winboard protocol.
If it is correct I can use it. I remember from school that you can translate it automatically into code.So I should be able to do that manually.
https://www.chessprogramming.org/Chess_ ... n_Protocol

[jdart]

I have implemented the Winboard protocol, but it is quite complex, especially if including pondering and in analysis or edit mode. There are many corner cases that can cause bugs: for example receiving a move or a ponder hit after the game has ended. My implementation is solid now, but that is after quite a few bug fixes.