error when using SCID about alexandria HOT 7 CLOSED

Hi, testing locally using Scid vs pc version 4.24 works just fine, this is probably a specific problem of Scid v5.02. I'll look for a 5.02 binary to confirm this
Edit: just noticed this is strictly about Scid and not Scid vs pc, i'll look into it

from alexandria.

Update: Found the issue, a fix should be published soon

from alexandria.

indeed it's about Scid, not the vs-PC version.
compiling went well and now it runs in Scid !

What did you find (in general terms) ? Did you find Scid not handling the UCI protocol properly ? Is it a timeout issue ? Maybe i can use your finding for other engine authors to be aware and correct their code (many engines fail in Scid here..)

from alexandria.

Technically speaking it's not an error on Scid's end, at least it wasn't for me, there are 2 ways to encode startpos with a uci legal command
position startpos and position startpos moves with no actual moves being appended, Alex assumed that if you were appending the moves token there were moves to parse, so it tried to parse moves that didn't exist and ended up crashing.
Position startpos moves is a legal uci command though, even if unusual.

from alexandria.

OK, thanks.
I know position startpos moves is used in polyglot .epd files to set a starting position, it's not unusual, it's an alternative to position fen [..] which not all engines understand.

from alexandria.

The command itself isn't rare at all, and it's basically a requirement to play any game with any uci GUI. What's unusual is sending the moves token with no moves attached, instead of simply sending position startpos, when there are no moves to pass on to the engine.

from alexandria.

now i got you!
a simple fix yes .. nice :-)

from alexandria.