Re: [libvirt] [PATCH 2/8] tests: Use DO_TEST_CAPS_*() for Intel IOMMU

On Tue, 2019-05-21 at 16:15 +0200, Ján Tomko wrote: > On Mon, May 20, 2019 at 01:37:47PM +0200, Andrea Bolognani wrote: > > diff --git a/tests/qemuxml2argvdata/intel-iommu-machine.xml b/tests/qemuxml2argvdata/intel-iommu-machine.xml > > index 90aba16156..85027cd938 100644 > > --- a/tests/qemuxml2argvdata/intel-iommu-machine.xml > > +++ b/tests/qemuxml2argvdata/intel-iommu-machine.xml > > @@ -5,7 +5,7 @@ > > <currentMemory unit='KiB'>219100</currentMemory> > > <vcpu placement='static'>1</vcpu> > > <os> > > - <type arch='x86_64' machine='q35'>hvm</type> > > + <type arch='x86_64' machine='pc-q35-2.6'>hvm</type> > > Given that this is the only difference between intel-iommu-machine.xml > and intel-iommu.xml, could you delete intel-iommu-machine.xml > and use DO_TEST_CAPS_VER with 2.6.0? > > We won't be able to reuse the XML file due to only striping the machine > aliases for latest XML files, but it will show we use the same input > file. I'm not sure I fully understand what you're suggesting... Do you mean squashing in something like the diff below?

Personally I like the idea of using the same input file for different DO_TEST*() calls, highlighting how the environment is the only thing causing differences in the output. That said, in the past I've been told (I think by Peter?) doing so is not a good idea, so I've avoided it since.

We'd also lose, as you mention yourself, the nice property of the output file being a symlink to the input file. tl;dr I'm perfectly happy using either the patch as-is or with the diff below squashed in; you guys tell me which one I should go for.