Re: [libvirt] [PATCHv2 7/8] tests: Add tests for virtio-rng device handling
On 02/23/13 01:29, Eric Blake wrote:
On 02/21/2013 07:47 AM, Peter Krempa wrote:
> Adds XML parsing and qemu commandline tests for the VirtIO RNG device
> support.
> ---
>
> Notes:
> Version 2:
> - ACKed
>
> .../qemuxml2argv-virtio-rng-egd.args | 1 +
> .../qemuxml2argv-virtio-rng-egd.xml | 26 ++++++++++++++++++++++
> .../qemuxml2argv-virtio-rng-random.args | 1 +
> .../qemuxml2argv-virtio-rng-random.xml | 23 +++++++++++++++++++
> tests/qemuxml2argvtest.c | 5 +++++
> tests/qemuxml2xmltest.c | 3 +++
> 6 files changed, 59 insertions(+)
> create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-virtio-rng-egd.args
> create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-virtio-rng-egd.xml
> create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-virtio-rng-random.args
> create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-virtio-rng-random.xml
Is it worth testing that a filename containing an XML-special character
is properly escaped? Other than that, this one is still good to go.
> +++ b/tests/qemuxml2argvdata/qemuxml2argv-virtio-rng-random.xml
> @@ -0,0 +1,23 @@
> + <rng model='virtio'>
> + <backend model='random'>/test/phile</backend>
That is, should this use something like /test/<phile as the XML
encoded file name?
Uh, I'm not following you on this one. You mean that if the user
specifies some characters that are invalid from the perspective of XML
as a source path?
Anyways, I fixed the issues you pointed out in 1-6 and provided
explanation for the other stuff. I'm pushing patches 1-7 (the test suite
can be improved at any time) now and will follow up later with a
improved version of 8 as well as with a patch that will allow multiple
RNG devices. That should be better to review as slicing apart the
existing patches.
Peter