bash and sh difference in break this while do loop

[jamesps]

This is an extract from /opt/VRTS/bin/vxrootmir

tfound=no
tmpdg=
mdm=
mda=

vxdisk -q list | \
while read da type dm grp state rest
do
if [ "$dm" = "$1" ]; then
tmpdg=$grp
mdm=$1
tfound=yes
break
elif [ "$da" = "$1" ]; then
if [ "$dm" != "-" ]; then
tmpdg=$grp
mdm=$dm
fi
tfound=yes
break
fi
done

if [ "$tfound" = "no" ]; then
abort "Disk $1 not found in configuration" vxvmshm:0
fi


Here's what I would like to explain: under bash (3.00.14) the tfound variable is set to "no" again after it is set to "yes" and breaks after the first if. Because of that the disk not found message is displayed and the script exits unsuccessfully.

This behavior does not happen under sh. It looks so simple but... why?

Thanks!
james

[jamesps]

I am using B.11.11, latest security patches.

[James R. Ferguson]

Hi James:

I'd use 'set -x' and perhaps add 'set -v' and trace the execution and variable setting.

Regards!

...JRF...

[A. Clay Stephenson]

I suspect that your read is the problem.

Try this very simple test:

echo "One Two Three" | read A B C
echo "A = \"${A}\""
echo "B = \"${B}\""
echo "C = \"${C}\""

I think you will find different results in the POSIX shell and bash -- at least some implementations of bash. It's one of the reasons I'm not a huge fan of bash. My shell of choice under Linix is zsh.

[Kent Ostby]

Perhaps it should be == instead of = in

if [ "$tfound" = "no" ]; then