Re: [libvirt] [PATCH v3 08/16] conf, schema: add support for memnode elements

On 07/16/2014 02:43 PM, Martin Kletzander wrote: >>> Good point, is this OK to push as trivial (git diff -w): >> >> Count this as my ACK :) >> > > I pushed it then, thank you. Not my day. I was so focused on the 'diff -w' aspect that I completely overlooked another aspect. The patch is wrong: > >>> >>> diff --git i/docs/schemas/domaincommon.rng >>> w/docs/schemas/domaincommon.rng >>> index a0ea300..fb5bdb3 100644 >>> --- i/docs/schemas/domaincommon.rng >>> +++ w/docs/schemas/domaincommon.rng >>> @@ -790,6 +790,7 @@ >>> </element> >>> </optional> >>> <zeroOrMore> >>> + <interleave> <interleave> makes no difference here. As the <zeroOrMore> has only one <element> child, there is nothing to be interleaved. I meant for it to go one level higher, outside the <zeroOrMore>, where it can also interleave with <memory>. >>> <element name="memnode"> >> >> I'm assuming the odd spacing here is due to pasting into the email body, >> not how it actually looked in the diff. That, and diff -w already plays >> games with spacing. >> > > diff -w looked OK when I pasted it in the mail body, but somewhere on > the way it got smudged. As penance, I'm proposing this followup: diff --git i/docs/schemas/domaincommon.rng w/docs/schemas/domaincommon.rng index fb5bdb3..2caeef9 100644 --- i/docs/schemas/domaincommon.rng +++ w/docs/schemas/domaincommon.rng @@ -759,6 +759,7 @@ <!-- All the NUMA related tunables would go in the numatune --> <define name="numatune"> <element name="numatune"> + <interleave> <optional> <element name="memory"> <optional> @@ -790,7 +791,6 @@ </element> </optional> <zeroOrMore> - <interleave> <element name="memnode"> <attribute name="cellid"> <ref name="unsignedInt"/> @@ -806,8 +806,8 @@ <ref name='cpuset'/> </attribute> </element> - </interleave> </zeroOrMore> + </interleave> </element> </define>