On 21 April 2014 17:31, Juan Quintela <quint...@redhat.com> wrote: > Patches are easy to review in sequence, any of them is very simple, and > the few ones that are long (minimum_version_id_old) review is just > looking that the previous line is minimum_version_id = <same number>.
But there are simply _far too many_ of them in this series. For instance "remove minimum_version_id_old" should be a series by itself: it is self contained and can be reviewed and committed much more easily that way. It seems likely that there are three or four other things also going on in this set of patches. "Improve the vmstate tests" should also be a separate patchset. I'm not sure about merging the field versioning with the field test function stuff, but I'm not about to try to fish the relevant patches out of this enormous mess so I'll wait until they appear in a series of their own before I comment on them. You mention at least one bugfix of some kind: that ought to be its own patch or patchset. Even 40 patches or so in a set is pushing the boundaries of what is reasonably reviewable in my opinion. thanks -- PMM
