On 12/08/2016 09:00 AM, Nikolay Shirokovskiy wrote:
On 08.12.2016 16:21, John Ferlan wrote:
>
> Perhaps a commit message would answer my question below...
>
> On 11/22/2016 06:09 AM, Nikolay Shirokovskiy wrote:
>> ---
>> src/conf/domain_conf.c | 5 ++++-
>> 1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
>> index 6e008e2..fb6ff0b 100644
>> --- a/src/conf/domain_conf.c
>> +++ b/src/conf/domain_conf.c
>> @@ -22534,7 +22534,10 @@ virDomainGraphicsDefFormat(virBufferPtr buf,
>> virBufferAsprintf(buf, " autoport='%s'",
>> def->data.vnc.autoport ? "yes" :
"no");
>>
>> - if (def->data.vnc.websocket)
>> + if (def->data.vnc.websocketGenerated &&
>
> Wouldn't websocketGenerated imply an active domain? And a change of the
> websocket in the active xml to be some value?
>
> So is the purpose of this because if you have an active domain you don't
> want to display the websocket that was generated?
>
> And why is that?
>
> IOW: What are you trying to ensure with this patch?
>
AFAIU this combination - active domain with inactive dump flag is used
to serialize config in situations described in cover letter - migration
or saving of a domain. So instead of saving port we save the fact it is autogenerated
and regenerate on migration destination/restore. One can infer this from
socket port logic in near by code.
Oh - didn't read the cover....
I can move that explanation in here before pushing. Although I'll wait
for a response to my recent question in 2/3 before doing so.
John