1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
|
#!/bin/bash
#
# Morgan Deters <mdeters@cs.nyu.edu>
# for the CVC4 project
#
# usage:
#
# run_regression cvc4-binary [ --proof | --dump ] [ benchmark.cvc | benchmark.smt | benchmark.smt2 ]
#
# Runs benchmark and checks for correct exit status and output.
#
# ulimit -t 1 # For detecting long running regressions
ulimit -s 65000 # Needed for some (esp. portfolio and model-building)
prog=`basename "$0"`
if [ $# -lt 2 ]; then
echo "usage: $prog [ --proof | --dump ] [ wrapper ] cvc4-binary [ benchmark.cvc | benchmark.smt | benchmark.smt2 | benchmark.p ]" >&2
exit 1
fi
proof=no
dump=no
if [ x"$1" = x--proof ]; then
proof=yes
shift
elif [ x"$1" = x--dump ]; then
dump=yes
shift
fi
if [ $# -lt 2 ]; then
echo "usage: $prog [ --proof | --dump ] [ wrapper ] cvc4-binary [ benchmark.cvc | benchmark.smt | benchmark.smt2 | benchmark.p ]" >&2
exit 1
fi
wrapper=
while [ $# -gt 2 ]; do
wrapper="$wrapper$1 "
shift
done
[[ "$VALGRIND" = "1" ]] && {
wrapper="libtool --mode=execute valgrind $wrapper"
}
cvc4=$1
benchmark_orig=$2
benchmark="$benchmark_orig"
function error {
echo "$prog: error: $*"
exit 1
}
if ! [ -x "$cvc4" ]; then
error "\`$cvc4' doesn't exist or isn't executable" >&2
fi
if ! [ -r "$benchmark" ]; then
error "\`$benchmark' doesn't exist or isn't readable" >&2
fi
# gettemp() and its associated tempfiles[] array are intended to never
# allow a temporary file to leak---the trap ensures that when this script
# exits, whether via a regular exit or an -INT or other signal, the
# temp files are deleted.
declare -a tempfiles
trap -- 'test ${#tempfiles[@]} -gt 0 && rm -f "${tempfiles[@]}"' EXIT
function gettemp {
local temp="`mktemp -t "$2"`"
tempfiles[${#tempfiles[@]}]="$temp"
eval "$1"="$temp"
}
tmpbenchmark=
if expr "$benchmark" : '.*\.smt$' &>/dev/null; then
lang=smt1
if test -e "$benchmark.expect"; then
expected_output=`grep '^% EXPECT: ' "$benchmark.expect" | sed 's,^% EXPECT: ,,'`
expected_error=`grep '^% EXPECT-ERROR: ' "$benchmark.expect" | sed 's,^% EXPECT-ERROR: ,,'`
expected_exit_status=`grep -m 1 '^% EXIT: ' "$benchmark.expect" | perl -pe 's,^% EXIT: ,,;s,\r,,'`
command_line=`grep '^% COMMAND-LINE: ' "$benchmark.expect" | sed 's,^% COMMAND-LINE: ,,'`
if [ -z "$expected_exit_status" ]; then
expected_exit_status=0
fi
elif grep '^% \(PROOF\|EXPECT\|EXPECT-ERROR\|EXIT\|COMMAND-LINE\): ' "$benchmark" "$benchmark" &>/dev/null; then
expected_output=`grep '^% EXPECT: ' "$benchmark" | sed 's,^% EXPECT: ,,'`
expected_error=`grep '^% EXPECT-ERROR: ' "$benchmark" | sed 's,^% EXPECT-ERROR: ,,'`
expected_exit_status=`grep -m 1 '^% EXIT: ' "$benchmark" | perl -pe 's,^% EXIT: ,,;s,\r,,'`
command_line=`grep '^% COMMAND-LINE: ' "$benchmark" | sed 's,^% COMMAND-LINE: ,,'`
# old mktemp from coreutils 7.x is broken, can't do XXXX in the middle
# this frustrates our auto-language-detection
gettemp tmpbenchmark cvc4_benchmark.smt.$$.XXXXXXXXXX
grep -v '^% \(PROOF\|EXPECT\|EXPECT-ERROR\|EXIT\|COMMAND-LINE\): ' "$benchmark" >"$tmpbenchmark"
if [ -z "$expected_exit_status" ]; then
expected_exit_status=0
fi
benchmark=$tmpbenchmark
elif grep '^ *:status *sat' "$benchmark" &>/dev/null; then
expected_output=sat
expected_exit_status=0
command_line=
elif grep '^ *:status *unsat' "$benchmark" &>/dev/null; then
expected_output=unsat
expected_exit_status=0
command_line=
else
error "cannot determine status of \`$benchmark'"
fi
elif expr "$benchmark" : '.*\.smt2$' &>/dev/null; then
lang=smt2
if test -e "$benchmark.expect"; then
expected_output=`grep '^% EXPECT: ' "$benchmark.expect" | sed 's,^% EXPECT: ,,'`
expected_error=`grep '^% EXPECT-ERROR: ' "$benchmark.expect" | sed 's,^% EXPECT-ERROR: ,,'`
expected_exit_status=`grep -m 1 '^% EXIT: ' "$benchmark.expect" | perl -pe 's,^% EXIT: ,,;s,\r,,'`
command_line=`grep '^% COMMAND-LINE: ' "$benchmark.expect" | sed 's,^% COMMAND-LINE: ,,'`
if [ -z "$expected_exit_status" ]; then
expected_exit_status=0
fi
elif grep '^; \(EXPECT\|EXPECT-ERROR\|EXIT\|COMMAND-LINE\): ' "$benchmark" "$benchmark" &>/dev/null; then
expected_output=`grep '^; EXPECT: ' "$benchmark" | sed 's,^; EXPECT: ,,'`
expected_error=`grep '^; EXPECT-ERROR: ' "$benchmark" | sed 's,^; EXPECT-ERROR: ,,'`
expected_exit_status=`grep -m 1 '^; EXIT: ' "$benchmark" | perl -pe 's,^; EXIT: ,,;s,\r,,'`
command_line=`grep '^; COMMAND-LINE: ' "$benchmark" | sed 's,^; COMMAND-LINE: ,,'`
if [ -z "$expected_exit_status" ]; then
expected_exit_status=0
fi
elif grep '^ *( *set-info *:status *sat' "$benchmark" &>/dev/null; then
expected_output=sat
expected_exit_status=0
command_line=
elif grep '^ *( *set-info *:status *unsat' "$benchmark" &>/dev/null; then
expected_output=unsat
expected_exit_status=0
command_line=
else
error "cannot determine status of \`$benchmark'"
fi
elif expr "$benchmark" : '.*\.cvc$' &>/dev/null; then
lang=cvc4
expected_output=$(grep '^% EXPECT: ' "$benchmark")
expected_error=`grep '^% EXPECT-ERROR: ' "$benchmark" | sed 's,^% EXPECT-ERROR: ,,'`
if [ -z "$expected_output" -a -z "$expected_error" ]; then
error "cannot determine expected output of \`$benchmark': " \
"please use \`% EXPECT:' and/or \`% EXPECT-ERROR:' gestures"
fi
expected_output=$(echo "$expected_output" | perl -pe 's,^% EXPECT: ,,;s,\r,,')
expected_exit_status=`grep -m 1 '^% EXIT: ' "$benchmark" | perl -pe 's,^% EXIT: ,,;s,\r,,'`
if [ -z "$expected_exit_status" ]; then
expected_exit_status=0
fi
command_line=`grep '^% COMMAND-LINE: ' "$benchmark" | sed 's,^% COMMAND-LINE: ,,'`
elif expr "$benchmark" : '.*\.p$' &>/dev/null; then
lang=tptp
command_line=--finite-model-find
expected_output=$(grep '^% EXPECT: ' "$benchmark")
expected_error=`grep '^% EXPECT-ERROR: ' "$benchmark" | sed 's,^% EXPECT-ERROR: ,,'`
if [ -z "$expected_output" -a -z "$expected_error" ]; then
if grep -q '^% Status *: ' "$benchmark"; then
expected_output="$(grep '^% *Status *: ' "$benchmark" | head -1 | awk '{print$NF}')"
case "$expected_output" in
Theorem|Unsatisfiable) expected_exit_status=0 ;;
CounterSatisfiable|Satisfiable) expected_exit_status=0 ;;
GaveUp) expected_exit_status=0 ;;
esac
expected_output="% SZS status $expected_output for $(basename "$benchmark" | sed 's,\.p$,,')"
else
error "cannot determine expected output of \`$benchmark': " \
"please use \`% EXPECT:' and/or \`% EXPECT-ERROR:' gestures"
fi
else
expected_output=$(echo "$expected_output" | perl -pe 's,^% EXPECT: ,,;s,\r,,')
expected_exit_status=`grep -m 1 '^% EXIT: ' "$benchmark" | perl -pe 's,^% EXIT: ,,;s,\r,,'`
fi
if [ -z "$expected_exit_status" ]; then
expected_exit_status=0
fi
if grep -q '^% COMMAND-LINE: ' "$benchmark"; then
command_line=`grep '^% COMMAND-LINE: ' "$benchmark" | sed 's,^% COMMAND-LINE: ,,'`
fi
else
error "benchmark \`$benchmark' must be *.cvc or *.smt or *.smt2 or *.p"
fi
command_line="${command_line:+$command_line }--lang=$lang"
gettemp expoutfile cvc4_expect_stdout.$$.XXXXXXXXXX
gettemp experrfile cvc4_expect_stderr.$$.XXXXXXXXXX
gettemp outfile cvc4_stdout.$$.XXXXXXXXXX
gettemp errfile cvc4_stderr.$$.XXXXXXXXXX
gettemp errfilefix cvc4_stderr.$$.XXXXXXXXXX
gettemp exitstatusfile cvc4_exitstatus.$$.XXXXXXXXXX
if [ -z "$expected_output" ]; then
# in case expected stdout output is empty, make sure we don't differ
# by a newline, which we would if we echo "" >"$expoutfile"
touch "$expoutfile"
else
echo "$expected_output" >"$expoutfile"
fi
check_models=false
if grep '^sat$' "$expoutfile" &>/dev/null || grep '^invalid$' "$expoutfile" &>/dev/null || grep '^unknown$' "$expoptfile" &>/dev/null; then
if ! expr "$CVC4_REGRESSION_ARGS $command_line" : '.*--check-models' &>/dev/null &&
! expr "$CVC4_REGRESSION_ARGS $command_line" : '.*--no-check-models' &>/dev/null; then
# later on, we'll run another test with --check-models on
check_models=true
fi
fi
check_proofs=false
if [ "$proof" = yes ]; then
# proofs not currently supported in incremental mode, turn it off
if grep '^unsat$' "$expoutfile" &>/dev/null || grep '^valid$' "$expoutfile" &>/dev/null &>/dev/null; then
if ! expr "$CVC4_REGRESSION_ARGS $command_line" : '.*--check-proofs' &>/dev/null &&
! expr "$CVC4_REGRESSION_ARGS $command_line" : '.*--no-check-proofs' &>/dev/null &&
! expr "$CVC4_REGRESSION_ARGS $command_line" : '.*--incremental' &>/dev/null &&
! expr " $CVC4_REGRESSION_ARGS $command_line" : '.* -[a-zA-Z]*i' &>/dev/null; then
# later on, we'll run another test with --check-proofs on
check_proofs=true
fi
fi
fi
if [ -z "$expected_error" ]; then
# in case expected stderr output is empty, make sure we don't differ
# by a newline, which we would if we echo "" >"$experrfile"
touch "$experrfile"
else
echo "$expected_error" >"$experrfile"
fi
cvc4dir=`dirname "$cvc4"`
cvc4dirfull=`cd "$cvc4dir" && pwd`
if [ -z "$cvc4dirfull" ]; then
error "getting directory of \`$cvc4 !?"
fi
cvc4base=`basename "$cvc4"`
cvc4full="$cvc4dirfull/$cvc4base"
if [ $dump = no ]; then
echo running $wrapper $cvc4full $CVC4_REGRESSION_ARGS $command_line `basename "$benchmark"` [from working dir `dirname "$benchmark"`]
time ( :; \
( cd `dirname "$benchmark"`;
$wrapper "$cvc4full" $CVC4_REGRESSION_ARGS $command_line `basename "$benchmark"`;
echo $? >"$exitstatusfile"
) > "$outfile" 2> "$errfile" )
else
echo running $wrapper $cvc4full $CVC4_REGRESSION_ARGS $command_line --preprocess-only --dump=clauses --output-lang=smt2 -q `basename "$benchmark"` \| $wrapper $cvc4full $CVC4_REGRESSION_ARGS $command_line --lang=smt2 - [from working dir `dirname "$benchmark"`]
time ( :; \
( cd `dirname "$benchmark"`;
$wrapper "$cvc4full" $CVC4_REGRESSION_ARGS $command_line --preprocess-only --dump=clauses --output-lang=smt2 -q `basename "$benchmark"` | $wrapper "$cvc4full" $CVC4_REGRESSION_ARGS $command_line --lang=smt2 -;
echo $? >"$exitstatusfile"
) > "$outfile" 2> "$errfile" )
fi
# we have to actual error file same treatment as other files. differences in
# versions of echo/bash were causing failure on some platforms and not on others
# (also grep out valgrind output, if 0 errors reported by valgrind)
actual_error=$(cat $errfile)
if [[ "$VALGRIND" = "1" ]]; then
#valgrind_output=$(cat $errfile|grep -E "^==[0-9]+== "|)
valgrind_num_errors=$(cat $errfile|grep -E "^==[0-9]+== "|tail -n1|awk '{print $4}')
echo "valgrind errors (not suppressed): $valgrind_num_errors" 1>&2
((valgrind_num_errors == 0)) && actual_error=$(echo "$actual_error"|grep -vE "^==[0-9]+== ")
fi
if [ -z "$actual_error" ]; then
# in case expected stderr output is empty, make sure we don't differ
# by a newline, which we would if we echo "" >"$experrfile"
touch "$errfilefix"
else
echo "$actual_error" >"$errfilefix"
fi
diffs=`diff -u --strip-trailing-cr "$expoutfile" "$outfile"`
diffexit=$?
diffserr=`diff -u --strip-trailing-cr "$experrfile" "$errfilefix"`
diffexiterr=$?
exit_status=`cat "$exitstatusfile"`
exitcode=0
if [ $diffexit -ne 0 ]; then
echo "$prog: error: differences between expected and actual output on stdout"
echo "$diffs"
exitcode=1
fi
if [ $diffexiterr -ne 0 ]; then
echo "$prog: error: differences between expected and actual output on stderr"
echo "$diffserr"
exitcode=1
fi
if [ "$exit_status" != "$expected_exit_status" ]; then
echo "$prog: error: expected exit status \`$expected_exit_status' but got \`$exit_status'"
exitcode=1
fi
if $check_models || $check_proofs; then
check_cmdline=
if $check_models; then
check_cmdline="$check_cmdline --check-models"
fi
if $check_proofs; then
check_cmdline="$check_cmdline --check-proofs"
fi
# at least one sat/invalid response: run an extra model/proof-checking pass
if ! CVC4_REGRESSION_ARGS="$CVC4_REGRESSION_ARGS$check_cmdline" "$0" $wrapper "$cvc4" "$benchmark_orig"; then
exitcode=1
fi
fi
exit $exitcode
|