On 9/12/25 8:34 PM, Mike Pattrick via dev wrote: > Previously ovs-lib would assume a database is valid if the file exists, > however, it is possible for the database file to exist but not be valid > for ovsdb to open it. > > Now existence checks are augmented with empty file validation. > Empty databases are removed. > > Reported-at: https://issues.redhat.com/browse/FDP-689 > Reported-by: Ihar Hrachyshka > Signed-off-by: Mike Pattrick <[email protected]> > --- > v5: > - Corrected comments > - Moved commands into AT_CHECK > - Stopped supressing mktemp errors > --- > tests/ovsdb-server.at | 63 +++++++++++++++++++++++++++++++++++++++++++ > utilities/ovs-lib.in | 31 ++++++++++++++++++--- > 2 files changed, 91 insertions(+), 3 deletions(-) > > diff --git a/tests/ovsdb-server.at b/tests/ovsdb-server.at > index ca6e931be..9480738b3 100644 > --- a/tests/ovsdb-server.at > +++ b/tests/ovsdb-server.at > @@ -105,6 +105,69 @@ AT_CHECK([uuidfilt output], [0], > ]], []) > AT_CLEANUP > > +AT_SETUP([invalid database is recreated]) > +AT_KEYWORDS([ovsdb unix]) > +AT_SKIP_IF([test "$IS_WIN32" = "yes"]) > +ordinal_schema > schema > + > +. ovs-lib
What happened to the LSB functions patch: https://patchwork.ozlabs.org/project/openvswitch/patch/[email protected]/ ? Somehow I got my system into a state where the test fails 100% of the time due to PATH being reset by sourcing the library. It worked in the morning, I have no idea what changed... One potentially relevant thing that I did was to start a sandbox, then source the ovs-lib inside of it, then test the backup_db function a bit. After exiting the sandbox now the PATH is getting reset while running the test, which wasn't happening before. It's also somehow a system wide change, i.e. the test fails in other terminals too... Weird. But we probably still need the LSB sourcing removal patch. What do you think? Best regards, Ilya Maximets. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
