Author: Will Singleton
Date: 13:38:48 02/06/03
Go up one level in this thread
On February 06, 2003 at 14:26:56, Daniel Mehrmannn wrote:
>Hi :-)
>I send version to Leo :-))
>
>changes:
>
>Alle versions are allow from Tim:
>---------------------------------
>
>06.02.2003 winboard-4.2.6-dm-beta2 release
>
>* bugfix a lot of errors in the GUI checks:
> -engine restart if user makes settings and engine running
> appDatas was not correct set
>
> -Windows handling: More then 1 Errormessage
> works now. Only 1 errormessage if user make a mistake
>
>* bugfix Analysis Window:
> -DisplayAnalysis() was call but never works.
> Add my appData to ANALYSIS_TIMER_ID
>
> -Analysis Windows freezing during a long run
> Create a new IcsAnalyzeWindowUp () to do the right steps.
>
> -Call Windows down every engine analyse setting
> Safe now state and compare and restart engine only if we must do that
>
>* bugfix Engine: -Check if engine support analyze. I forgot this :((
> Build in IcsAnalyze()
>
>
>* bufgix observe If User abort observing the engine don't stop
> -Check now state
>
>* bugfix WindowState EnableWindow(GetDlgItem(hDlg...
> Set disable all analyze feature if we have no engine found
>
>
>* bugfix -Split depth 0 and 1
> For book depth we are now using the struct from tim.
>
>* bugfix engine state -If observe a game and (no analyze): Engine want play
>(this bug is in 4.2.6 too)
> SendToEngine force
>
>
>* new redesign Engine handling:
>
> Now, we don't reset engine very move. We do that only if we start a new
>analyze (Enable
> Analyze Engine)
> So engines won't be stopped using HT or somthing else and don't need a new
>move list.
> It's faster, it's better. I make some checks and if we have a problem we
>start the
> engine new. Normaly we only push rhe move
>
>
>* new takeback moves If a takeback was made Engine reject this move (ICS
>Observing) (this bug is in 4.2.6 too)
> Add a check with folling steps:
>
> /* If we found takebacks during IcsAnalyze try send to engine and check
>later!!*/
> if (!newGame && appData.icsAnalyze && first.analyzing &&
> moveNum < forwardMostMove) {
> if (appData.debugMode) fprintf(debugFP, "take back move");
> takeback = forwardMostMove - moveNum;
> /* safty first */
> SendToProgram("force \n", &first);
> for (i = 0; i <= takeback; i++) {
> SendToProgram("undo \n", &first);
> }
> /* Some engine need analyze and stat again */
> SendToProgram("analyze \n", &first);
> SendToProgram(". \n", &first);
>
>
>
>* new Message: call DisplayMessage() if we do something
>
>
>
>05.02.2003
>* First version Winboard-4.2.6-dm-beta1 release *
I noticed that when an engine is observing a game and "telling" to a ICC
channel, that it sends analysis of the same ply repeatedly. Is this because the
analyze function of winboard requests the data every few seconds?
This kind of noise should be avoided. It makes it difficult to read other
channel text, and it's useless and annoying. It would be helpful to set some
noise parameters, like only one tell per ply, and suppress analysis until ply 10
or 11. Note also that ICC puts a limit on the frequency of channel tells, so
that if you do 3 or 4 in the span of a few seconds, it cuts you off for awhile.
Engine authors who want to kib observed games might also consider making their
score plus for white, negative for black. Otherwise, it gets really strange
when your scores flip back and forth.
Will
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.