New shell scripts for testing, used for travis CI (#957)

Perform more complex verification and validation.

- signature.sh : extract and optionally compare Bench/Signature/Node count.
- perft.sh : verify perft counts for a number of positions.
- instrumented.sh : run a few commands or uci sequences through valgrind/sanitizer instrumented binaries.
- reprosearch.sh : verify reproducibility of search.

These script can be used from directly from the command line in the src directory.

Update travis script to use these shell scripts.

No functional change.

diff --git a/.travis.yml b/.travis.yml
index 64cc2b6342c3b304288b2318fcfbd3a61f8a6a1c..a0673b85c6dae3d15270376079061a035059e564 100644 (file)
--- a/.travis.yml
+++ b/.travis.yml
@@ -44,20 +44,29 @@ before_script:
   - cd src
 
 script:
-  - make clean && make build ARCH=x86-64 && ./stockfish bench 2>&1 >/dev/null | grep 'Nodes searched' | tee bench1
-  - make clean && make build ARCH=x86-32 && ./stockfish bench 2>&1 >/dev/null | grep 'Nodes searched' | tee bench2
-  - echo "Checking for same bench numbers..."
-  - diff bench1 bench2 > result
-  - test ! -s result
-  # verify perft numbers (positions from https://chessprogramming.wikispaces.com/Perft+Results)
-  - printf ' set timeout 10\n lassign $argv pos depth result\n spawn ./stockfish\n send "position $pos\\n perft $depth\\n"\n expect "Nodes searched  ? $result" {} timeout {exit 1} \n send "quit\\n"\n expect eof\n' > perft.exp
-  - expect perft.exp startpos 5 4865609 > /dev/null
-  - expect perft.exp "fen r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq -" 5 193690690 > /dev/null
-  - expect perft.exp "fen 8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - -" 6 11030083 > /dev/null
-  - expect perft.exp "fen r3k2r/Pppp1ppp/1b3nbN/nP6/BBP1P3/q4N2/Pp1P2PP/R2Q1RK1 w kq - 0 1" 5 15833292 > /dev/null
-  - expect perft.exp "fen rnbq1k1r/pp1Pbppp/2p5/8/2B5/8/PPP1NnPP/RNBQK2R w KQ - 1 8" 5 89941194 > /dev/null
-  - expect perft.exp "fen r4rk1/1pp1qppp/p1np1n2/2b1p1B1/2B1P1b1/P1NP1N2/1PP1QPPP/R4RK1 w - - 0 10" 5 164075551 > /dev/null
-  # if valgrind is available check the build is without error, reduce depth to speedup testing, but not too shallow to catch more cases.
-  - if [ -x "$(command -v valgrind )" ] ; then make clean && make ARCH=x86-64 debug=yes build && valgrind --error-exitcode=42 ./stockfish bench 128 1 10 default depth 1>/dev/null ; fi
-  # use g++-6 as a proxy for having sanitizers ... might need revision as they become available for more recent versions of clang/gcc than trusty provides
-  - if [[ "$COMPILER" == "g++-6" ]]; then make clean && make ARCH=x86-64 sanitize=yes build && ! ./stockfish bench 2>&1 | grep "runtime error:" ; fi
+  #
+  # checking bench for various build types
+  #
+  # obtain reference
+  - make clean && make ARCH=x86-64 optimize=no debug=yes build > /dev/null && export benchref=$(../tests/signature.sh)
+  - echo "Reference bench:" $benchref
+  # verify against reference
+  - make clean && make ARCH=x86-64 build > /dev/null && ../tests/signature.sh $benchref
+  - make clean && make ARCH=x86-32 build > /dev/null && ../tests/signature.sh $benchref
+  #
+  # perft
+  #
+  - make clean && make ARCH=x86-64 build > /dev/null && ../tests/perft.sh
+  #
+  # reproducible search
+  #
+  - make clean && make ARCH=x86-64 build > /dev/null && ../tests/reprosearch.sh
+  #
+  # valgrind
+  #
+  - if [ -x "$(command -v valgrind )" ]; then make clean && make ARCH=x86-64 debug=yes optimize=no build > /dev/null && ../tests/instrumented.sh --valgrind; fi
+  #
+  # sanitizer
+  #
+  # use g++-6 as a proxy for having sanitizers, might need revision as they become available for more recent versions of clang/gcc
+  - if [[ "$COMPILER" == "g++-6" ]]; then make clean && make ARCH=x86-64 sanitize=yes build > /dev/null && ../tests/instrumented.sh --sanitizer; fi

diff --git a/tests/instrumented.sh b/tests/instrumented.sh
new file mode 100755 (executable)
index 0000000..a6950e1
--- /dev/null
+++ b/tests/instrumented.sh
@@ -0,0 +1,82 @@
+#!/bin/bash
+# check for errors under valgrind or sanitizers.
+
+error()
+{
+  echo "instrumented testing failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+# define suitable post and prefixes for testing options
+case $1 in
+  --valgrind)
+    echo "valgrind testing started"
+    prefix=''
+    exeprefix='valgrind --error-exitcode=42'
+    postfix='1>/dev/null'
+  ;;
+  --sanitizer)
+    echo "sanitizer testing started"
+    prefix='!'
+    exeprefix=''
+    postfix='2>&1 | grep "runtime error:"'
+  ;;
+  *)
+    echo "unknown testing started"
+    prefix=''
+    exeprefix=''
+    postfix=''
+  ;;
+esac
+
+# simple command line testing
+for args in "eval" \
+            "go nodes 1000" \
+            "go depth 10" \
+            "go movetime 1000" \
+            "go wtime 8000 btime 8000 winc 500 binc 500" \
+            "bench 128 1 10 default depth"
+do
+
+   echo "$prefix $exeprefix ./stockfish $args $postfix"
+   eval "$prefix $exeprefix ./stockfish $args $postfix"
+
+done
+
+# more general testing, following an uci protocol exchange
+cat << EOF > game.exp
+ set timeout 10
+ spawn $exeprefix ./stockfish
+
+ send "uci\n"
+ expect "uciok"
+
+ send "ucinewgame\n"
+ send "position startpos\n"
+ send "go nodes 1000\n"
+ expect "bestmove"
+
+ send "position startpos moves e2e4 e7e6\n"
+ send "go nodes 1000\n"
+ expect "bestmove"
+
+ send "quit\n"
+ expect eof
+
+ # return error code of the spawned program, useful for valgrind
+ lassign [wait] pid spawnid os_error_flag value
+ exit \$value
+EOF
+
+for exps in game.exp
+do
+
+  echo "$prefix expect game.exp $postfix"
+  eval "$prefix expect game.exp $postfix"
+
+done
+
+rm game.exp
+
+echo "instrumented testing OK"

diff --git a/tests/perft.sh b/tests/perft.sh
new file mode 100755 (executable)
index 0000000..f3d5b88
--- /dev/null
+++ b/tests/perft.sh
@@ -0,0 +1,32 @@
+#!/bin/bash
+# verify perft numbers (positions from https://chessprogramming.wikispaces.com/Perft+Results)
+
+error()
+{
+  echo "perft testing failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+echo "perft testing started"
+
+cat << EOF > perft.exp
+   set timeout 10
+   lassign \$argv pos depth result
+   spawn ./stockfish
+   send "position \$pos\\n perft \$depth\\n"
+   expect "Nodes searched  ? \$result" {} timeout {exit 1}
+   send "quit\\n"
+   expect eof
+EOF
+
+expect perft.exp startpos 5 4865609 > /dev/null
+expect perft.exp "fen r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq -" 5 193690690 > /dev/null
+expect perft.exp "fen 8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - -" 6 11030083 > /dev/null
+expect perft.exp "fen r3k2r/Pppp1ppp/1b3nbN/nP6/BBP1P3/q4N2/Pp1P2PP/R2Q1RK1 w kq - 0 1" 5 15833292 > /dev/null
+expect perft.exp "fen rnbq1k1r/pp1Pbppp/2p5/8/2B5/8/PPP1NnPP/RNBQK2R w KQ - 1 8" 5 89941194 > /dev/null
+expect perft.exp "fen r4rk1/1pp1qppp/p1np1n2/2b1p1B1/2B1P1b1/P1NP1N2/1PP1QPPP/R4RK1 w - - 0 10" 5 164075551 > /dev/null
+
+rm perft.exp
+
+echo "perft testing OK"

diff --git a/tests/reprosearch.sh b/tests/reprosearch.sh
new file mode 100755 (executable)
index 0000000..9fd847f
--- /dev/null
+++ b/tests/reprosearch.sh
@@ -0,0 +1,61 @@
+#!/bin/bash
+# verify reproducible search
+
+error()
+{
+  echo "reprosearch testing failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+echo "reprosearch testing started"
+
+# repeat two short games, separated by ucinewgame. 
+# with go nodes $nodes they should result in exactly
+# the same node count for each iteration.
+cat << EOF > repeat.exp
+ set timeout 10
+ spawn ./stockfish
+ lassign \$argv nodes
+
+ send "uci\n"
+ expect "uciok"
+
+ send "ucinewgame\n"
+ send "position startpos\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "position startpos moves e2e4 e7e6\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "ucinewgame\n"
+ send "position startpos\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "position startpos moves e2e4 e7e6\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "quit\n"
+ expect eof
+EOF
+
+# to increase the likelyhood of finding a non-reproducible case,
+# the allowed number of nodes are varied systematically
+for i in `seq 1 20`
+do
+
+  nodes=$((100*3**i/2**i))
+  echo "reprosearch testing with $nodes nodes"
+
+  # each line should appear exactly an even number of times
+  expect repeat.exp $nodes 2>&1 | grep -o "nodes [0-9]*" | sort | uniq -c | awk '{if ($1%2!=0) exit(1)}'
+
+done
+
+rm repeat.exp
+
+echo "reprosearch testing OK"

diff --git a/tests/signature.sh b/tests/signature.sh
new file mode 100755 (executable)
index 0000000..2eb52c8
--- /dev/null
+++ b/tests/signature.sh
@@ -0,0 +1,26 @@
+#!/bin/bash
+# obtain and optionally verify Bench / signature
+# if no reference is given, the output is deliberately limited to just the signature
+
+error()
+{
+  echo "running bench for signature failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+# obtain
+
+signature=`./stockfish bench 2>&1 | grep "Nodes searched  : " | awk '{print $4}'`
+
+if [ $# -gt 0 ]; then
+   # compare to given reference
+   if [ "$1" != "$signature" ]; then
+      echo "signature mismatch: reference $1 obtained $signature"
+   else
+      echo "signature OK: $signature"
+   fi
+else
+   # just report signature
+   echo $signature
+fi