FICS

Discussion of chess software programming and technical issues.

Moderator: Ras

adams161
Posts: 626
Joined: Sun May 13, 2007 9:55 pm
Location: Bay Area, CA USA
Full name: Mike Adams

Re: FICS

Post by adams161 »

when someone sends an abort and adjournment on move 1 to an engine that often means the engine didnt move. If it sent its first move and winboard didn't accept it, could it be a bad book move?
adams161
Posts: 626
Joined: Sun May 13, 2007 9:55 pm
Location: Bay Area, CA USA
Full name: Mike Adams

Re: FICS

Post by adams161 »

also another thing you want to watch out for is nobody managed to start a game that was not standard chess. winboard normally guards against this but i dont know that every case of wild chess on fics would be stopped by zippy. seeing the full debug of that game would be helpful.
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: FICS

Post by sje »

adams161 wrote:when someone sends an abort and adjournment on move 1 to an engine that often means the engine didnt move. If it sent its first move and winboard didn't accept it, could it be a bad book move?
I think 1 e4 is a legal first move. :)

It must be an xboard screw-up. The only time an "undo" should be sent by xboard to a local engine is a takeback from the ICS (but "notakeback" is in force) or a local menu pick (but the machine was running unattended).
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: FICS

Post by sje »

adams161 wrote:also another thing you want to watch out for is nobody managed to start a game that was not standard chess. winboard normally guards against this but i dont know that every case of wild chess on fics would be stopped by zippy. seeing the full debug of that game would be helpful.
Symbolic's xboard interface explicitly outputs "normal" during the feature sequence as the only supported variant. Also, if the "variant" command is received and the operand is not "normal", Symbolic will flag an error.
MattieShoes
Posts: 718
Joined: Fri Mar 20, 2009 8:59 pm

Re: FICS

Post by MattieShoes »

sje wrote:
MattieShoes wrote:
sje wrote:
hgm wrote:You would have to know if this undo was originating from XBoard or from the ICS. XBoard also does send undo commands when it receives an out-of-turn move from an engine, for instance.
I saw the mystery "undo" happen again on FICS. It occurred when Symbolic was playing White and sent its first move. The opponent requested an adjournment, and then requested an abort. (Symbolic was running unattended and ignored both requests.)
Before your first move on fics, aborting aborts the game, it doesn't request an abort.

Anywhere from 2-5, disconnects are aborted rather than adjourned.

Is this perhaps it?

1) noob matches symbolic, xboard accepts, informs symbolic
2) noob aborts
3) symbolic moves
Symbolic plays nearly instantly in the opening, so as White it most likely sent its first move before the opponent could do anything.

Anyway, xboard should in all cases of opponent initiated aborts, adjournments, and disconnects by sending a "result", not an "undo".

And remember, the FICS notakeback variable is set.

Maybe there's a mystery xboard command line option that also needs to be set to avoid an "undo" emission.
Perhaps post the relevant portion of the log file? :-)
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: FICS

Post by sje »

MattieShoes wrote:Perhaps post the relevant portion of the log file? :-)
Symbolic's log tracks all the engine/xboard data exchange, but none of the xboard/ICC exchange as it has no access. I don't have an xboard debug output. Symbolic's log basically says:

Code: Select all

< move e4
> undo
adams161
Posts: 626
Joined: Sun May 13, 2007 9:55 pm
Location: Bay Area, CA USA
Full name: Mike Adams

Re: FICS

Post by adams161 »

i'm confused on one point. i always wrote my moves like e2e4, maybe this is entirely irrelevent but i thought it said in the winboard doc that xboard expected moves that way. but i've never been bold enough to try differently. without the xboard log though thats all i can add.
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: FICS

Post by sje »

adams161 wrote:i'm confused on one point. i always wrote my moves like e2e4, maybe this is entirely irrelevent but i thought it said in the winboard doc that xboard expected moves that way. but i've never been bold enough to try differently. without the xboard log though thats all i can add.
Symbolic has always used SAN. There's an xboard feature option in protocol version two that supports this.

I can't recall an "undo" ever being received during ICS play when using earlier versions of xboard.
MattieShoes
Posts: 718
Joined: Fri Mar 20, 2009 8:59 pm

Re: FICS

Post by MattieShoes »

sje wrote:
MattieShoes wrote:Perhaps post the relevant portion of the log file? :-)
Symbolic's log tracks all the engine/xboard data exchange, but none of the xboard/ICC exchange as it has no access. I don't have an xboard debug output. Symbolic's log basically says:

Code: Select all

< move e4
> undo
Perhaps turn on debugging in xboard and wait for it to happen again? :-)
bnemias
Posts: 373
Joined: Thu Aug 14, 2008 3:21 am
Location: Albuquerque, NM

Re: FICS

Post by bnemias »

sje wrote:
MattieShoes wrote:Perhaps post the relevant portion of the log file? :-)
Symbolic's log tracks all the engine/xboard data exchange, but none of the xboard/ICC exchange as it has no access. I don't have an xboard debug output. Symbolic's log basically says:

Code: Select all

< move e4
> undo
Does your engine support ping? If so, can you include the ping/pong data?