xboard-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[XBoard-devel] Auto-instal of engines


From: H.G. Muller
Subject: [XBoard-devel] Auto-instal of engines
Date: Mon, 15 Sep 2014 19:42:43 +0200
User-agent: Mozilla/5.0 (Windows NT 6.2; WOW64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0

I just equipped XBoard (not WinBoard) with an -autoInstall option,
which should still go into the 4.8.0 release. This option uses what I hope
will become the new standard for engine packages in Linux:

All engines should contain a 'plugin-spec' file <enginename>.eng, which they
should install in the directory /usr/share/games/plugins/<protocol>/ , where
<protocol> = uci, xboard, etc. The file will contain a magic sentence "plugin spec 0.0",
the command line needed to start the engine for the given protocol (e.g.
"pos --io-mode xboard") and possibly a working directory (as a preceding cd command), a list of games / variant groups (like chess, chess960, capablanca, xiangqi, shogi, fairychess, loserchess, checkers, go; the recognized game names will still have to be defined).

When -autoInstall is not an empty string, XBoard wil append any newly installed engines for UCI and XBoard protocol that it sees for the first time (as based on the modification date of the .eng file and the user settings file), to the -firstChessProgramNames list. The line will specify the corresponding -fcp, -fd arguments, and in case of UCI engines add a -fUCI flag. In this first implementation the variants list is ignored; in the future it could be matched with the -autoInstall list to decide which engines should be installed, and which should be ignored. (e.g. -autoInstall "shogi" could only install engines that play Shogi).

In the future the plugin-spec files could contain extra lines to give additional information, e.g. about known non-compliancy with the protocol (e.g. UCI engines that do not implement "searchmoves" or do not always repond to "isready"), or other peculiarities (e.g. whether they print the PV as long algebraic rather than SAN). XBoard could then add the applicable work-around options for these engine defects. (Like -fSAN to have the engine's PV converted to SAN, or -firstScoreAbs true for engines that report scores from whithe POV, etc.)



reply via email to

[Prev in Thread] Current Thread [Next in Thread]