On 5/9/19 3:38 PM, Alex Bennée wrote: >> Hm, this makes every iotest print two lines: >> >> $ ./check -T -qcow2 >> [...] >> 001 [20:06:27] -> [20:06:27] >> 001 0s (last 1s) > > Yes - it was a compromise to ensure we printed a start and end > timestamp but I guess we can fix it up with a bit more shell ugliness: > > --8<---------------cut here---------------start------------->8--- > > Subject: [PATCH] fixup! tests/qemu-iotests: re-format output to for make > check-block > > --- > tests/qemu-iotests/check | 8 ++++++-- > 1 file changed, 6 insertions(+), 2 deletions(-) > > diff --git a/tests/qemu-iotests/check b/tests/qemu-iotests/check > index fb239967a32..9f083f06b46 100755 > --- a/tests/qemu-iotests/check > +++ b/tests/qemu-iotests/check > @@ -726,7 +726,11 @@ _report_test_result() > if $pretty; then > echo " TEST iotest: $1 $2" > else > - echo "$1 $2" > + if $timestamp; then > + echo " $2" > + else > + echo "$1 $2" > + fi
Why not just make $1 '' in the case where you've already printed a timestamp, so that this code is still unconditionally echo "$1 $2"? > fi > } > > @@ -793,7 +797,7 @@ do > $run_command >$tmp.out 2>&1) > fi > sts=$? > - $timestamp && echo " [$(date "+%T")]" > + $timestamp && echo -n " [$(date "+%T")]" 'echo -n' is not portable (even in bash, since you can compile a different default for shopt xpg_echo). Better is to use 'printf %s', as is already done in _timestamp(). -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org
signature.asc
Description: OpenPGP digital signature
