Stéphane Nicolet
Stéphane Nicolet
This stackoverflow post seems to have some related info: https://stackoverflow.com/questions/30829364/open-utf8-encoded-filename-in-c-windows Not sure if we should care much about it in Stockfish
Since Stefano has a very clear idea of the kind of unit tests he/she would like, but has not proposed any unit code in the last six months, I'll close...
In case the PGN files with the time losses disappears from the fishtest database, I have copied them here: http://cassio.free.fr/divers/6346c3d94bc7650f075471ce-3.pgn http://cassio.free.fr/divers/6346c3d94bc7650f075471ce-26.pgn
yes, stopping the run with adjudication=OFF and the time losses, then starting a new one with adjudication=ON is probably a good idea to get a comparison
Question: why don't we use a clamp in line 476 of search, to ensure that the variable `complexPosition` stay positive? ``` double complexPosition = std::min(1.0 + (complexity - 261) /...
So, comparison run of 40.000 games against SF13 and adjudication=ON ended with time loses also, both for the patch and for SF13: https://tests.stockfishchess.org/tests/view/6346fbfc4bc7650f075479f2
@peregrineshahin Indeed I think we could add some clamping and/or asserts in that part of the code, just to be on the safe side about the underlying hypotheses of our...
Can we instead modify line 66 to read like that? ``` #if defined(_WIN64) && defined(_MSC_VER) && !defined(__clang__) // No Makefile used ```
@maximmasiutin what would you think about the above solution? :-)
Can you confirm that this would fix the warning for you? :-)