Yet another mismerge

[ddugovic]

God damn it. https://travis-ci.org/github/ddugovic/Stockfish/builds/702976421

test.fen

r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq - 0 10 moves e2a6 b4c3

build.sh

#!/bin/bash
make clean && git checkout origin/master
git log HEAD | grep "\b[Bb]ench[ :]\+[0-9]\{7\}" | head -n 1 | sed "s/[^0-9]*\([0-9]*\).*/\1/g" > git_sig
export benchref=$(cat git_sig)
echo "Reference bench:" $benchref
rm git_sig
make build ARCH=x86-64-bmi2 debug=yes && mv stockfish stockfish_master
make clean && git checkout master

#export CXXFLAGS="-Werror -D_GLIBCXX_DEBUG"
#if [ -x "$(command -v valgrind )" ]; then make clean && make -j3 ARCH=x86-64-bmi2 debug=yes build > /dev/null && ../tests/instrumented.sh --valgrind; fi
make -j3 build ARCH=x86-64-bmi2 COMP=clang debug=yes #&& ../tests/signature.sh $benchref #&& ../tests/perft.sh

#SYZYGY_PATH='/home/lila/Stockfish/tb' ./get_bench.sh bench.log 'all 128 1 13'

./stockfish_master bench 128 1 6819 test.fen nodes |& sed -u -e 's/info depth //' | sed -u -e 's/seldepth //' | sed -u -e 's/multipv //' | sed -u -e 's/tbhits 0 //' | tee 1.txt
./stockfish bench chess 128 1 6819 test.fen nodes |& sed -u -e 's/info depth //' | sed -u -e 's/seldepth //' | sed -u -e 's/multipv //' | sed -u -e 's/tbhits 0 //' | tee 2.txt
sdiff --width=158 1.txt 2.txt
echo "Reference bench:" $benchref

[ddugovic]

Commenting out most variants seems to help: 64be369

[ddugovic]

Fixed by 6cb44a2 as demonstrated by https://travis-ci.org/github/ddugovic/Stockfish/builds/703120588 (the build no longer breaks after a non-functional upstream merge https://travis-ci.org/github/ddugovic/Stockfish/builds/702976421 ).

@ianfab I may need to rebase http://www.variantfishtest.org:6543/tests/view/5ef9bd336e23db104fb88c81 (maybe it requires 6cb44a2 ) so I reduced its game limit to 5000. Feel free to further reduce that limit and/or stop that test.

[ianfab]

@ddugovic If the test is obsolete because the fix does not work without the other one, then please stop it, and otherwise let it finish. However, for such a technical bugfix LTC maybe anyway does not make sense, since the only purpose of testing it is to make sure that the fix works and does not break anything else (which should be visible at any TC).