Am 16.07.2012 18:37, schrieb Eric Blake: > On 07/16/2012 02:12 AM, Pavel Hrdina wrote: >> On 07/12/2012 07:59 PM, Eric Blake wrote: >>> On 07/12/2012 10:55 AM, Pavel Hrdina wrote: >>>> Signed-off-by: Pavel Hrdina<phrd...@redhat.com> >>>> --- >>>> hmp-commands.hx | 2 +- >>>> hmp.c | 10 ++++++++++ >>>> hmp.h | 1 + >>>> qapi-schema.json | 17 +++++++++++++++++ >>>> qmp-commands.hx | 24 ++++++++++++++++++++++++ >>>> savevm.c | 21 +++++++++++---------- >>>> sysemu.h | 1 - >>>> 7 files changed, 64 insertions(+), 12 deletions(-) >>>> + >>>> +## >>>> +# @delvm: >>> This name is worse than 'loadvm' or 'savevm', in that we are NOT >>> deleting the entire vm, but a named snapshot stored within the vm. >>> Would naming it 'delete-vm-snapshot' make more sense (in which case the >>> others might make more sense as 'save-vm-snapshot' and >>> 'load-vm-snapshot')? >>> >> This naming looks nice. I definitely agree that it could be >> save-vm-snapshot, load-vm-snapshot, delete-vm-snapshot and >> query-vm-snapshots. > > On seeing that spelled out, I wonder if the '-vm' is just noise; where > we could use 'query-snapshots' instead of 'query-vm-snapshots'. Then > again, we already have 'blockdev-snapshot-sync', which is an entirely > different snapshot (just a block device, rather than the entire VM), so > dropping -vm is probably a bad idea. Anyone else want to chime in on > the bikeshed painting discussion of the best QMP name?
Sure. ;-) If there is a blockdev-snapshot-sync, then the counterparts for the whole VM shouldn't be called load/save/del-vm-snapshot, but vm-snapshot-load/save/del. And possibly add a -sync as well. Kevin