On 05/31/2018 12:16 AM, Peter Xu wrote:
This event is not really necessary. After OOB series it might affect
the timing of the script so this event may or may not be there comparing
to the old *.out results. Let's just filter it out.
This is worrying. Are you stating that the SHUTDOWN event can occur in a
different order than it used to, or is it even worse that the SHUTDOWN
event disappears altogether? If enabling OOB makes the SHUTDOWN event
sometimes disappear, that's a regression that we should fix. If it just
makes things occur in a different order, we need an explanation why that
is okay.
Since some of the scripts are using qmp-pretty, we need some trick in
the filtering script to make sure sed works for multiple lines to
explicitly mask out this event.
CC: John Snow <js...@redhat.com>
CC: Max Reitz <mre...@redhat.com>
CC: Kevin Wolf <kw...@redhat.com>
Signed-off-by: Peter Xu <pet...@redhat.com>
---
+++ b/tests/qemu-iotests/067.out
@@ -70,6 +70,7 @@ Testing: -drive
file=TEST_DIR/t.IMGFMT,format=IMGFMT,if=none,id=disk -device vir
}
}
+
=== -drive/device_add and device_del ===
Why is this blank line being added (multiple times in this file)? Is
there something about the new filter that isn't quite stripping all the
newlines when encountering the SHUTDOWN event in pretty form?
Aha - it's because test 067 is already doing a very similar filtering of
events. Can we reduce the code duplication by promoting
_filter_qmp_events from there into common.filter (as a separate patch)?
+++ b/tests/qemu-iotests/common.filter
@@ -88,7 +88,10 @@ _filter_qmp()
sed -e 's#\("\(micro\)\?seconds": \)[0-9]\+#\1 TIMESTAMP#g' \
-e 's#^{"QMP":.*}$#QMP_VERSION#' \
-e '/^ "QMP": {\s*$/, /^ }\s*$/ c\' \
- -e ' QMP_VERSION'
+ -e ' QMP_VERSION' | \
+ tr '\n' '\r' | \
+ sed -e
's/{\s*"timestamp":\s*{\s*"seconds":\s*TIMESTAMP,\s*"microseconds":\s*TIMESTAMP\s*},\s*"event":\s*"SHUTDOWN",\s*"data":\s*{\s*"guest":\s*false\s*}\s*}\s//'
| \
Really long line. This should do the same:
sed -e 's/\r{\(\r[^}]\|[^\r]\)*SHUTDOWN\(\r[^}]\|[^\r]\)*\r}//'
where the \(\r[^}]\|[^\r]\)* subpattern picks up all line breaks that do
not end the current top-level {}, as well as any non-line breaks.
In fact, if you like my suggestion about promoting the filter from 67
into common.filter, we have two use cases: filter a single
pretty-printed filter, and filter ANY pretty-printed filter. Maybe we
do that as follows:
# $1 is a regex of event names to filter; leave empty to filter all
_filter_qmp_events()
{
fluff='\(\r[^}]\|[^\r]\)*'
tr \\n \\r | \
sed -e "s/$fluff"'"event": "'"$1$fluff\\r}//" \
| tr \\r \\n
}
at which point '_filter_qmp_events SHUTDOWN' works in this patch, and
'_filter_qmp_events' works for 067. [Untested, but hopefully that gives
you some ideas to play with]
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
