On Wed, Jun 14, 2017 at 08:55:25AM -0400, John Ferlan wrote:
On 06/02/2017 11:04 AM, Daniel Liu wrote:
> The option allows someone to run domain-to-native on already existing
> domain without the need of supplying their XML. It is basically
> wrapper around 'virsh dumpxml | virsh domxml-to-native /dev/stdin'.
>
> Resolves:
https://bugzilla.redhat.com/show_bug.cgi?id=835476
> Signed-off-by: Daniel Liu <srwx4096(a)gmail.com>
> ---
> tools/virsh-domain.c | 51 ++++++++++++++++++++++++++++++++++++++-------------
> tools/virsh.pod | 7 ++++---
> 2 files changed, 42 insertions(+), 16 deletions(-)
>
> diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c
> index ccb514ef9..3496359c5 100644
> --- a/tools/virsh-domain.c
> +++ b/tools/virsh-domain.c
> @@ -9848,9 +9848,13 @@ static const vshCmdOptDef opts_domxmltonative[] = {
> .flags = VSH_OFLAG_REQ,
> .help = N_("target config data type format")
> },
> + {.name = "domain",
> + .type = VSH_OT_DATA,
> + .flags = VSH_OFLAG_REQ_OPT,
> + .help = N_("domain name, id or uuid")
> + },
> {.name = "xml",
> .type = VSH_OT_DATA,
> - .flags = VSH_OFLAG_REQ,
> .help = N_("xml data file to export from")
> },
> {.name = NULL}
> @@ -9859,30 +9863,51 @@ static const vshCmdOptDef opts_domxmltonative[] = {
> static bool
> cmdDomXMLToNative(vshControl *ctl, const vshCmd *cmd)
> {
> - bool ret = true;
> + bool ret = false;
> const char *format = NULL;
> - const char *xmlFile = NULL;
> - char *configData;
> - char *xmlData;
> + const char *xml = NULL;
nit: no need to change from xmlFile to xml... The less differences the
better ;-)
There was no change, it is an additional variable, the original one is
below. The number of differences would be the same, I believe.
> + char *xmlData = NULL;
> + char *configData = NULL;
nit: no need to change order from original, just add ' = NULL' is fine.
> unsigned int flags = 0;
> virshControlPtr priv = ctl->privData;
> + virDomainPtr dom = NULL;
>
> if (vshCommandOptStringReq(ctl, cmd, "format", &format) < 0 ||
> - vshCommandOptStringReq(ctl, cmd, "xml", &xmlFile) < 0)
> + vshCommandOptStringReq(ctl, cmd, "xml", &xml) < 0)
> return false;
>
> - if (virFileReadAll(xmlFile, VSH_MAX_XML_FILE, &xmlData) < 0)
> - return false;
> + VSH_EXCLUSIVE_OPTIONS("domain", "xml");
>
> - configData = virConnectDomainXMLToNative(priv->conn, format, xmlData,
flags);
> - if (configData != NULL) {
> - vshPrint(ctl, "%s", configData);
> - VIR_FREE(configData);
> + if (vshCommandOptBool(cmd, "domain") &&
> + (!(dom = virshCommandOptDomain(ctl, cmd, NULL))))
> + return false;
> +
> + if (dom) {
> + xmlData = virDomainGetXMLDesc(dom, flags);
> + } else if (xml) {
> + if (virFileReadAll(xml, VSH_MAX_XML_FILE, &xmlData) < 0)
> + goto cleanup;
> } else {
> - ret = false;
> + vshError(ctl, "%s", _("need either domain or domain
XML"));
> + goto cleanup;
> + }
> +
> + if (!xmlData) {
> + vshError(ctl, "%s", _("failed to retrieve XML"));
> + goto cleanup;
> }
>
> + if (!(configData = virConnectDomainXMLToNative(priv->conn, format, xmlData,
flags))) {
> + goto cleanup;
> + } else {
> + vshPrint(ctl, "%s", configData);
> + ret = true;
> + }
> +
> + cleanup:
> + virshDomainFree(dom);
> VIR_FREE(xmlData);
> + VIR_FREE(configData);
> return ret;
> }
>
> diff --git a/tools/virsh.pod b/tools/virsh.pod
> index aee964689..049c2f3c7 100644
> --- a/tools/virsh.pod
> +++ b/tools/virsh.pod
> @@ -1433,10 +1433,11 @@ the I<format> argument must be B<qemu-argv>. For
Xen hypervisor, the
> I<format> argument may be B<xen-xm>, B<xen-xl>, or
B<xen-sxpr>. For
> LXC hypervisor, the I<format> argument must be B<lxc-tools>.
>
> -=item B<domxml-to-native> I<format> I<xml>
> +=item B<domxml-to-native> I<format>
> +{ I<--domain> I<domain-name-or-id-or-uuid> | [I<--xml>]
I<xml> }
This looks really odd, I think this should be
{ [I<xml> | [I<--domain> I<domain-name-or-id-uuid>] }
Actually, the only thing that Dan can change is the order, but the rest
should remain. His version shows that you must supply either --domain
domain-name-or-id-or-uuid or --xml xml and that in the second case, the
--xml is optional.
Your suggested solution, however, mean that you must supply either the
xml (no mention about the optional --xml) or --domain
domain-name-or-id-or-uuid or nothing (in a really weird way), as the
latter option is in square brackets.
So the only way this could look differently is:
{ [I<--xml>] I<xml> | I<--domain> I<domain-name-or-id-or-uuid>
}
I believe that's the way to describe mutually exclusive options
where
one would be required; otherwise, we have the funky [[ ... ]] which
indicates mutually exclusive optional arguments. Search on "{ [" in
virsh.pod for some examples. This seems to be the first case where a
previously required positional argument can be both a positional and
optional as long as the alternative means of describing is provided.
Keeping the I<xml> first at least indicates that it can be used
*without* the --xml, but is also the "default" when used positionally on
the command line. If not used positionally, then the --xml would be
required, e.g. if someone used "virsh domxml-to-native --xml @file
--format @format".
Having the [--domain] indicates that in order to use it, it must have
that --domain argument because "virsh domxml-to-native qemu-argv
$domainname" would be invalid since the unqualified 2nd argument is
expected to be an <xml> file.
Whether any of this is written down anywhere - I'm not sure - it's what
I've surmised over the few years.
>
> -Convert the file I<xml> in domain XML format to the native guest
> -configuration format named by I<format>. For QEMU/KVM hypervisor,
> +Convert the file I<xml> in domain XML format or existing domain to the
> +native guest configuration format named by I<format>. For QEMU/KVM
hypervisor,
> the I<format> argument must be B<qemu-argv>. For Xen hypervisor, the
> I<format> argument may be B<xen-xm>, B<xen-xl>, or
B<xen-sxpr>. For
> LXC hypervisor, the I<format> argument must be B<lxc-tools>.
>
Convert the file I<xml> into domain XML format or convert an existing
I<--domain> to the native guest configuration format named by I<format>.
The I<xml> and I<--domain> arguments are mutually exclusive.
Yes, this sounds better, there is a slight flaw in the original hunk and
this explains it more nicely.
John
--
libvir-list mailing list
libvir-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list