About bidirectional named pipes (FIFO)

[Michael Haubenwallner]

Hi,

I'm encountering problems with some multi-processing bash script
that boils down to:

$ bash -c '
rm -f /tmp/testpipe.$$
trap "rm -f /tmp/testpipe.$$" EXIT
mkfifo /tmp/testpipe.$$
exec {fw}<>/tmp/testpipe.$$
exec {fr}<>/tmp/testpipe.$$
for x in {1..10}
do
(
trap "echo job $x pid \$BASHPID status \$? done >& $fw" EXIT
exit $x
) &
done
for z in {1..10}
do
read -u $fr
echo $REPLY
done
wait
'

Agreed, POSIX does not define behaviour for using '<>' with fifos,
but it works with AIX, *BSD, HP-UX, Interix, Linux, Solaris, ...

Thoughts?
/haubi/

[Corinna Vinschen]

Post by Michael Haubenwallner
Hi,
I'm encountering problems with some multi-processing bash script
$ bash -c '
rm -f /tmp/testpipe.$$
trap "rm -f /tmp/testpipe.$$" EXIT
mkfifo /tmp/testpipe.$$
exec {fw}<>/tmp/testpipe.$$
exec {fr}<>/tmp/testpipe.$$
for x in {1..10}
do
(
trap "echo job $x pid \$BASHPID status \$? done >& $fw" EXIT
exit $x
) &
done
for z in {1..10}
do
read -u $fr
echo $REPLY
done
wait
'
Agreed, POSIX does not define behaviour for using '<>' with fifos,
but it works with AIX, *BSD, HP-UX, Interix, Linux, Solaris, ...

FIFOs in Cygwin have a couple of problems, especially bidirectional
FIFOs since they require a certain order of usage, otherwise they hang.
The implementation is lacking in more than one way and needs a good
old revamp.


Corinna