[Scidvspc-users] Scid vs 960 chess engine issues (Was: Scid vs 960 is not loading default databases)

[Okechukwu I. <af...@ch...>]

Hello, again!

On closer inspection, it seems Phalanx behaves a little strangely, too. 
When first asking it to look at a game in progress, it analyses the 
standard start position instead, until you make a move, and then go 
back. Then, the first move it considers is for the side NOT on the move, 
and often enough, an illegal move (like moving the king into check). 
After the first line offered, the rest of its' analysis of the position 
appears normal.

So, maybe it is more than a UCI/Polyglot problem?

On 9/19/20 6:35 PM, Okechukwu Iwu wrote:
>
> Thanks for the quick response and action!
>
> Using the source for revision 3138 + chess 960 patch + Mr Cramer's 
> TKText patch (applied in that order), the problem with auto-load on 
> Scid vs 960 is now fixed!
>
> Another problem persisted, though. I noticed it after I sent the 
> earlier email. The 960 patch may be breaking something in the UCI 
> (polyglot?) code. Both Stockfish and Toga behave strangely (giving 
> ridiculous evaluations, or just not giving an evaluation), while 
> Phalanx behaves normally. For example, Toga claims an evaluation of ~ 
> +5 in this position:
>
> rnbq1rk1/pp2nppp/4p3/2ppP3/3P2Q1/P1P5/2P2PPP/R1B1KBNR w KQ - 2 8
>
> From personal experience as Black, I know it can be a tough position, 
> but not +5 :-) !
>
> I compiled regular Scid vs PC from r3138 source (using Gregor Cramer's 
> TKText patch, but no chess 960 patch), and both Stockfish 12 and Toga 
> behave normally.
>
> To allow them to coexist, Scid vs 960 is installed in 
> /usr/local/{bin,share}/, while Scid vs PC lives in /usr/{bin,share}/ .
>
> Okey
>
> On 9/19/20 1:29 AM, Steve A wrote:
>>
>> OK, this autoload bug is fixed in subversion patch. Cheers for the 
>> bug report.
>>
>>
>> OK.. But there are some installation "gotchas". The new gui is 
>> "scid-vs-960", but we don't have a new name for the new tkscid.. 
>> Which can cause errors, as the new binary and gui must be used 
>> together. I recommend removing the old tkscid and scid gui before 
>> doing a make clean and configure on the patched code.
>>
>> Ill try to make a windows package sometime, as this is solid code, 
>> apart from our crappy installation.
>>
>>
>> S.
>>
>>
>>
>>
>>
>> On Saturday, September 19, 2020, Okechukwu Iwu <af...@ch... 
>> <mailto:af...@ch...>> wrote:
>>
>>     Hello, Steven A!
>>
>>     I decided to try out the Chess 960 patch, which built "Scid vs
>>     960". I renamed my .s?4 files to .s?960, and can open them
>>     normally from the File menu. However, when I "Save Bases as
>>     Default", close the program and reopen, the default databases are
>>     not opened. The error message at the bottom of the startup window
>>     states:
>>
>>     "Not loading default bases as file /Chess/huge not readble"
>>
>>     (OK, the file path is "slightly" edited. And yes, the spelling
>>     typo is real :-) ).
>>
>>     I'm running Linux. The database files all have read permission,
>>     so I am wondering if Scid vs 960 is still expecting the default
>>     database files to end in .s?4, rather than .s?960 ?
>>
>>     I am sorry if this has already been addressed, but I could not
>>     find it in the mailing list archives.
>>
>>     Thanks for all your work on Scid vs PC!
>>
>>     Okey
>>
>>
>>
>>     _______________________________________________
>>     Scidvspc-users mailing list
>>     Sci...@li...
>>     <mailto:Sci...@li...>
>>     https://lists.sourceforge.net/lists/listinfo/scidvspc-users
>>     <https://lists.sourceforge.net/lists/listinfo/scidvspc-users>
>>