Scid 4.1 released

[pgeorges]

Hi,

I released an updated version of Scid at http://scid.sourceforge.net/

Scid is a chess database and training application. For those not knowing it, there are some screenshots here : http://scid.sourceforge.net/screenshots.html

Below is the complete changelog since Scid 4.0 :

- Features :
- Added the display of arrows when several variations are available. The display can be toggled using the menu Options -> Moves -> Show arrows for variations (Thanks to Fulvio)
- When playing a move already existing in a variation, the variation is automatically entered into. To add a new variation starting with the same move, the menu Edit-> Add variation must be used (Thanks to Fulvio)
- Added undo feature (10 undos max). Shortcut is Ctrl-z.
For coders : rollback points are set by calling {sc_game undoPoint} in Tcl or sc_game_save_for_undo() in C++.
- Updated Scid Pocket to new base format (version 4). Version 4.1 includes Robbolito chess engine, which is by far the best of all ! Really impressive. It outweighs all other engines on Pocket PC.
- Fixes :
- entering comments for last move of a line was broken
- setting the blunder flag from maintenance window was broken ("?" was considered part of a regular expression)
- when compacting a game file, custom flags description was lost
- custom flags description are now commited to index file as soon as they are modified
- Informant values were overwritten by default values at startup
- UCI engines : send "go infinite" instead of "go infinite ponder" because it is not handled correctly by some engines and "go infinite" seems sufficient
- fixed a bug when USCF piece set was selected and the library tkimg not installed (under Linux)
- UI may be unresponsive when an engine outputs lots of garbage when nearing a mate. Fixed for UCI engines only as I don't have an example of such misbehaviour of an xboard engine.
- Fixed a possible buffer overflow when parsing a malformed PGN (Thanks to Fulvio)
- Modified slightly UCI parser in order to get Prodeo to work (there's no garantee that Prodeo will work as it does not comply to UCI protocol) (Thanks to Shiv Shivaji)
- Fixed an error popping up when pressing right arrow very quickly with the variations window displayed and engines taking all available CPU
- Fixed impossibility of stripping comments and variations when display of short headers was turned on.
- Rewrite of "Review game" code : moves are no longer in comments, engine analysis is updated more often

Pascal Georges

[BeatMachine]

hi pascal,
thank you for all this work , but what about palm ? i know that you are the programmer imported ARM engines to openchess ,what about stockfish and robolito ?

[garybelton]

Thanks for this. I'd like to use SCID, I gave your new version a try, but like all the previous ones, when the *.sg4 database file grows to 2,147,483,366 bytes I can no longer import PGNs to the SCID database, ,message "error saving game to database".

Question: does anyone have a windows x64 compile of SCID? This problem doesn't happen on Linux x64, there the SCID database can grow as big as it likes, but I prefer to use windows.

Best.

[pgeorges]

hi pascal,
thank you for all this work , but what about palm ? i know that you are the programmer imported ARM engines to openchess ,what about stockfish and robolito ?

I no longer use any Palm since I stopped understanding their (lack of) strategy. I think Palm Os is dead.

Pascal

[pgeorges]

Thanks for this. I'd like to use SCID, I gave your new version a try, but like all the previous ones, when the *.sg4 database file grows to 2,147,483,366 bytes I can no longer import PGNs to the SCID database, ,message "error saving game to database".

Question: does anyone have a windows x64 compile of SCID? This problem doesn't happen on Linux x64, there the SCID database can grow as big as it likes, but I prefer to use windows.

Best.

I think it is more a file system limitation than a Scid one (its sole limit is that one base can only contain 16 M games). If you use NTFS you should not have any trouble. The only time I got this problem is when trying to generate a PGN file over 2GB under Windows (which I nearly never use).

Note that I have a DB of 4M games with a size of 800 MB. So I guess that to need more than 2GB, your base is over 10 M games : is it the case ? If so be aware of the hard limit of 16 M games I mentioned above.

Pascal Georges

[garybelton]

If you use NTFS you should not have any trouble.

Yes I am using NTFS on Windows Server 2008 but still get this problem, yes I get the 2GB file limit for PGN files too when exporting from ChessBase, but other files seem to be ok with being bigger than 2GB, for example I have virtual minidisk files of over 40gb, so I don't think it's a filesystem problem. I note that ChessBase is a 32-bit windows program as well, although I recall that I may have had the same issue with the zapchess CB 64-bit gui but my memory is shaky on that.

Note that I have a DB of 4M games with a size of 800 MB. So I guess that to need more than 2GB, your base is over 10 M games : is it the case?

Depends, if I leave the game comments in then around 5M games blow the 2GB limit, without comments around 9M games blow it. I have a 13.5M game db but this could grow a lot, I didn't know about the 16M hard limit but am still below that.

[pgeorges]

There is indeed a limit on 32 bits system, where the game size cannot grow over 2 GB. I will try to find where this comes from.

Pascal Georges

[garybelton]

Thanks, please note that I am using Win 2008 Server x64, so the problem is not just limited to 32-bit systems. A while ago I did actually attempt to make an x64 windows compile but gave up, too many compile errors related to my TCL iirc, both with MS and Intel C++ compilers.