[openstack-dev] [metrics] Another issue when analyzing SSH Gerrit (as young as August 2014)

Jesus M. Gonzalez-Barahona jgb at bitergia.com
Fri Sep 19 12:38:58 UTC 2014


On Fri, 2014-09-19 at 11:39 +0100, Daniel P. Berrange wrote:
> On Fri, Sep 19, 2014 at 12:30:50PM +0200, Daniel Izquierdo wrote:
> > Hi!
> > 
> > After having a more in depth review of data coming from Gerrit, some other
> > issues appear when using the SSH Gerrit API.
> > 
> > Some of the dates when reviewing the code (Code-Review, Verified, etc) seem
> > to be inconsistent with the data found in the comments section.
> > 
> > Example: issue https://review.openstack.org/#/c/98688/  (Heat project).
> > 
> > Let's take the PatchSet 44.
> > 
> > $ ssh -p 29418 review.openstack.org gerrit query 98688 --format=JSON
> > --files --comments --patch-sets --all-approvals --commit-message
> > --submit-records
> > 
> > After parsing a bit the JSON file...
> > 
> > Revision (patchset) 44 (cf7819ccdbf8999fc963e9c8d400c932ad780674)
> >   Date: 2014-08-14 03:18:48
> >   Code-Review: 1
> >     Date: 2014-07-23 07:01:31
> >   Code-Review: 1
> >     Date: 2014-07-31 11:16:21
> >   Code-Review: 1
> >     Date: 2014-08-01 10:53:28
> >   Code-Review: 2
> >     Date: 2014-08-12 20:15:14
> >   Code-Review: -1
> >     Date: 2014-08-14 10:23:55
> >   Workflow: -1
> >     Date: 2014-08-18 00:36:32
> >   Verified: 1
> >     Date: 2014-09-01 12:39:05
> > 
> > 
> > 
> > As you can see, the date of some of the reviews are older than the upload
> > date of the patchset 44 (Aug 14 3:18).
> 
> It looks like Patchset 44 is a trivial update of previous patchsets. When
> gerrit detects this, the votes from previous patchsets are copied across
> to the new patchset, and the time of those votes is preserved, since they
> are not actual new votes. So in that way, you can get votes which are
> older than the patchset itself.

Thanks a lot for the clarification. In that case, I understand that the
"safe" way of knowing to which patchset corresponds a given approval is
to look at the dates (of the approval and of the patchset), not at the
patchset number. Do you think that's right?

	Jesus.

> Regards,
> Daniel

-- 
Bitergia: http://bitergia.com
/me at Twitter: https://twitter.com/jgbarah




More information about the OpenStack-dev mailing list