As Paul mentioned, there is a missing descriptor 4 in the code. I will fix it up, test it and send a patch.
On Tue, Oct 1, 2013 at 4:17 PM, Paul Ingram <ping...@nicira.com> wrote: > OK, but I don't see the redirect to descriptor 4 in the real code (it's > obviously there in the example). > > :: psi > > On Oct 1, 2013, at 3:58 PM, Duffie Cooley <dcoo...@nicira.com> wrote: > > From the reference. > > Lets examine that step by step. > > ( ( ( ( someprog; #part6 > echo $? >&3 #part5 > ) | filter >&4 #part4 > ) 3>&1 #part3 > ) | stdintoexitstatus #part2 > ) 4>&1 #part1 > > From bottom up: > > A subshell is created with file descriptor 4 redirected to stdout. This > means that whatever is printed to file descriptor 4 in the subshell will end > up as the stdout of the entire construct. > A pipe is created and the commands on the left (#part3) and right > (stdintoexitstatus) are executed. stdintoexitstatus is also the last command > of the pipe and that means the exit status of stdintoexitstatus will be the > exit status of the entire construct. > A subshell is created with file descriptor 3 redirected to stdout. This > means that whatever is printed to file descriptor 3 in this subshell will > end up in stdintoexitstatus and in turn will be the exit status of the > entire construct. > A pipe is created and the commands on the left (#part5 and #part6) and right > (filter >&4) are executed. The output of filter is redirected to file > descriptor 4. In #part1 the file descriptor 4 was redirected to stdout. This > means that the output of filter is the stdout of the entire construct. > Exit status from #part6 is printed to file descriptor 3. In #part3 file > descriptor 3 was redirected tostdintoexitstatus. This means that the exit > status from #part6 will be the final exit status for the entire construct. > someprog is executed. The exit status is taken in #part5. The stdout is > taken by the pipe in#part4 and forwarded to filter. The output from filter > will in turn reach stdout as explained in#part4 > > > > On 10/01/2013 03:34 PM, Paul Ingram wrote: > > Why the redirect from descriptor 4? > > :: psi > > On Oct 1, 2013, at 3:13 PM, Duffie Cooley <dcoo...@nicira.com> wrote: > > What about this? > > referenced from here. > > http://unix.stackexchange.com/questions/14270/get-exit-status-of-process-thats-piped-to-another/70675#70675 > > diff --git a/utilities/ovs-lib.in b/utilities/ovs-lib.in > index 1684ddc..5286e12 100644 > --- a/utilities/ovs-lib.in > +++ b/utilities/ovs-lib.in > @@ -41,6 +41,11 @@ ovs_ctl_log () { > echo "$@" >> "${logdir}/ovs-ctl.log" > } > > +stdintoexitstatus () { > + read exitstatus > + return $exitstatus > +} > + > ovs_ctl () { > case "$@" in > *"=strace"*) > @@ -51,7 +56,7 @@ ovs_ctl () { > ;; > *) > echo "`date -u`:$@" >> "${logdir}/ovs-ctl.log" > - "${datadir}/scripts/ovs-ctl" "$@" 2>&1 | tee -a > "${logdir}/ovs-ctl.log" > + (((("${datadir}/scripts/ovs-ctl" "$@" 2>&1 ; echo $? > &3) > | tee -a "${logdir}/ovs-ctl.log" ) 3>&1) | stdintoexitstatus) 4>&1 > ;; > esac > } > > > Thanks, > > Duffie > > > On 10/01/2013 11:48 AM, Gurucharan Shetty wrote: > > This patch works for me, but it is a little hacky. > > diff --git a/utilities/ovs-lib.in b/utilities/ovs-lib.in > index 1684ddc..b0cdaf2 100644 > --- a/utilities/ovs-lib.in > +++ b/utilities/ovs-lib.in > @@ -51,7 +51,13 @@ ovs_ctl () { > ;; > *) > echo "`date -u`:$@" >> "${logdir}/ovs-ctl.log" > - "${datadir}/scripts/ovs-ctl" "$@" 2>&1 | tee -a > "${logdir}/ovs-ctl.log" > + unique=`(uuidgen) 2>/dev/null` || unique=`date +"%Y %b %d %T"` > + mkfifo "/tmp/${unique}" > + tee -a "${logdir}/ovs-ctl.log" < "/tmp/${unique}" & > + "${datadir}/scripts/ovs-ctl" "$@" > "/tmp/${unique}" 2>&1 > + rc=$? > + rm "/tmp/${unique}" > + return $rc > ;; > esac > } > > On Tue, Oct 1, 2013 at 10:11 AM, Gurucharan Shetty <shet...@nicira.com> > wrote: > > On Mon, Sep 30, 2013 at 7:27 PM, Duffie Cooley <dcoo...@nicira.com> wrote: > > From: Duffie Cooley <dcoo...@nicira.com> > This is a fix for a request to make sure that the openvswitch status command > in rhel based distros gives a useful exit status. That was fixed in > > commit 5e0c05bc058c78a11be6747f62e6ad88e5d06b70 > debian: Fix exit status of openvswitch-switch init script "status" command > > > > Signed-off-by: Duffie Cooley <dcoo...@nicira.com> > --- > rhel/etc_init.d_openvswitch | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/rhel/etc_init.d_openvswitch b/rhel/etc_init.d_openvswitch > index 7e64132..6a53cef 100755 > --- a/rhel/etc_init.d_openvswitch > +++ b/rhel/etc_init.d_openvswitch > @@ -5,7 +5,7 @@ > # chkconfig: 2345 09 91 > # description: Manage Open vSwitch kernel modules and user-space daemons > > -# Copyright (C) 2009, 2010, 2011 Nicira, Inc. > +# Copyright (C) 2009, 2010, 2011, 2013 Nicira, Inc. > # > # Licensed under the Apache License, Version 2.0 (the "License"); > # you may not use this file except in compliance with the License. > @@ -81,6 +81,7 @@ case $1 in > ;; > status) > ovs_ctl status > + exit $? > ;; > version) > ovs_ctl version > -- > 1.8.1.2 > > I don't think this helps. The exit status is always 0. I broke this > with my implementation of 'ovs_ctl' in ovs-lib tee'ing the o/p to > ovs-ctl.log. tee will return an exit status of success even though the > previous command returned an exit status of 1. > > I can use the PIPESTATUS variable. But I don't think it is POSIX > compliant. Any other ideas? > > _______________________________________________ > dev mailing list > dev@openvswitch.org > http://openvswitch.org/mailman/listinfo/dev > > > _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
