Carbec wrote: ↑Sun Jan 30, 2022 9:05 am
About TSCP, i indeed thought that the protocol was perhaps the problem. But modifying the protocol didn't change anything :
It changed something, since the error messages are not the same.
Maybe TSCP doesn't support the tournament book. Many engines have this problem.
[roland@localhost src (main)]$ ./zangdar
Entrez une commande : [help] uci
HashTable init complete with 5592405 entries for 128 Mo
id name Zangdar 0.34
id author Philippe Chevalier
uciok
quit
Entrez une commande : [help]
When I put the files on github, they didn't seen as a new version, but as new files.
Can someone explain how to use git, so I can put the files from my pc ?
I think that this topic is a little bit too extensive to just explain in a few lines of text. It would be best to follow some Git and Github tutorials if you have no experience with this.
I put this new version on github. At first, I didn't thought to do it, but after some matches, it shows an interesting improvement.
Surely I removed a bug, almost without thinking. How strange !!
But there are still bugs. For example, in this position : 8/8/8/8/1q4k1/8/6P1/2R4K b - - 0 131
Zangard played Qb1 ????. But if I put the position and ask it to play, it plays Kg3. I don't understand.
Well, This is new, but I don't go the XXRL list every day.
Good news then, so 2100+ perhaps ? I fear to be too optimistic.
How do the guys at CCRL knows that a new version is there ?