On 07/30/2013 09:29 PM, Eric Blake wrote: > On 07/30/2013 01:10 PM, Michal Novotny wrote: >> Output error message when user provides the invalid machine type >> on the command line. This also saves time to find what issue is >> when you upgrade from one version of qemu to another version that >> doesn't support required machine type any longer. > I know that has historically happened downstream (for example, Fedora 13 > made the mistake of creating a downstream-only fedora-13 machine type > that corresponded roughly to qemu 0.13, it took some further Fedora > patches to libvirt to modernize that name back to an upstream machine > type on an upgrade path before finally being able to drop the downstream > nonsense in Fedora 18 or so). But isn't upstream qemu supposed to be > guaranteeing command-line stability, in that a newer qemu will never > take away a machine type supported by an older qemu? > > But even if the upstream upgrade path never hits this code, your patch > will make it nicer if you _downgrade_ to a version of qemu (still with > this patch, of course) that lacks support for a newer machine. And the > fact that downstream will take advantage of this even if upstream never > causes the problem on upgrades still justifies using it.
You are correct, Eric. I sent a patch v4 to output error message to stderr instead so I fixed the commit message. Thanks, Michal -- Michal Novotny <minov...@redhat.com>, RHCE, Red Hat Virtualization | libvirt-php bindings | php-virt-control.org
