[Development] [RFC] more gerrit codereview scores?
oswald.buddenhagen at theqtcompany.com
Fri Mar 6 18:41:13 CET 2015
On Fri, Mar 06, 2015 at 06:10:46PM +0100, Christian Kandeler wrote:
> On 03/06/2015 05:42 PM, Oswald Buddenhagen wrote:
> > -1: "I would prefer this is not merged as is", advisory, non-sticky
> > -2: "This shall not be merged as is", blocking, non-sticky
> > -3: "This shall not be merged [at all]", blocking, sticky
> This makes sense under the assumption that there are patches of which
> you can be almost 100% sure that they are completely unfixable by
> whatever the author could come up with in the new patch set (including
> considerable changes to the concept).
> Is that something that happens reasonably often?
well, that depends on your definition of "reasonably often". i've seen a
bunch of them. it may not be many, but then, adding the additional
negative level also doesn't cost much.
second, if the conclusion was that "it doesn't happen reasonably often",
we still should make -2 non-sticky, even without adding a sticky -3.
but that feels just wrong to me.
additionally, from experience i'd say that most people abandon the
change and create an entirely new one anyway when they change the
approach (which i find unreasonable, but it's still data which makes the
concern less relevant).
lastly, "sticky" doesn't mean "immutable". i can still change my opinion
if somebody redesigns the change from scratch. on the technical side,
there is no regression in that regard.
More information about the Development