Re: [libvirt PATCH 9/9] tests: Add tests for unknown elements and attributes in cpu defintion

On Wed, Sep 30, 2020 at 13:54:58 +0200, Tim Wiederhake wrote: > On Tue, 2020-09-29 at 11:22 +0200, Peter Krempa wrote: > > On Mon, Sep 21, 2020 at 15:07:32 +0200, Tim Wiederhake wrote: > > > Signed-off-by: Tim Wiederhake <twiederh(a)redhat.com&gt; > > > --- > > > > [...] > > > > > diff --git a/tests/cputestdata/x86_64-bogus-element.xml > > > b/tests/cputestdata/x86_64-bogus-element.xml > > > new file mode 100644 > > > index 0000000000..79f98bad18 > > > --- /dev/null > > > +++ b/tests/cputestdata/x86_64-bogus-element.xml > > > @@ -0,0 +1,3 @@ > > > +<cpu> > > > + <nonExistantElement/> > > > +</cpu> > > > > I'm not persuaded that there's value in such test. > > > > This was my approach to adding negative tests to the "virsh > [hypervisor-]cpu-compare --validate" feature to make sure that > schema > violations are actually caught. If this is not the correct place to > do > so, I would be grateful for a pointer in the right direction. I'd find way more useful to actually test that all the XMLs we use in the tests are validated, which does not seem to happen in this series. This series adds schema validation only for a very synthetic negative case which IMO doesn't make sense. Additionally we do have 'virschematest', which is meant to validate all XML documents in the tests. Based on the fact that you've added a invalid XML example but it's name doesn't end in '-invalid.xml', which is the marker for the 'virschematest' worker to do negative validation, means that the cpu files are not validated. Rather than adding a synthetic case, wire up virschematest which will make sure that we don't have bogus XMLs in the test suite.