HUP signal handling?

I have built Jun 20 snapshot on my FreeBSD box. I have tried to run
test/bench and got following result.

ERROR:  destroydb: database bench does not exist.

This is normal. Problem is the make command is aborted by a signal. My
guesss is elog(ERROR) raises HUP signal, but for some reason nobody
catches it.  Here is the place the problem occurs in runwisc.sh:

echo "drop database bench" | postgres -D${1} template1 > /dev/null

Comments?
--
Tatsuo Ishii
t-ishii@sra.co.jp

[srapc451.sra.co.jp]t-ishii{107} gmake all runtest
if [ -z "$USER" ]; then USER=$LOGNAME; fi; \
if [ -z "$USER" ]; then USER=`whoami`; fi; \
if [ -z "$USER" ]; then echo 'Cannot deduce $USER.'; exit 1; fi; \
rm -f create.sql; \
C=`pwd`; \
sed -e "s:_CWD_:$C:g" \
    -e "s:_OBJWD_:$C:g" \
    -e "s:_SLSUFF_::g" \
    -e "s/_USER_/$USER/g" < create.source > create.sql
x=1; \
for i in `ls query[0-9][0-9]`; do \
  echo "select $x as x" >> bench.sql; \
  cat $i >> bench.sql; \
  x=`expr $x + 1`; \
done
rm -f bench.out bench.out.perquery
/bin/sh ./create.sh $PGDATA && \
/bin/sh ./runwisc.sh $PGDATA >bench.out 2>&1
=============== destroying old bench database... =================
ERROR:  destroydb: database bench does not exist.
ERROR:  destroydb: database bench does not exist.
gmake: *** [bench.out] Hangup
Hangup