[Gluster-devel] [Gluster-Maintainers] 'Reviewd-by' tag for commits

[Michael Adam]

On 2016-09-30 at 17:52 +0200, Niels de Vos wrote:
> On Fri, Sep 30, 2016 at 08:50:12PM +0530, Ravishankar N wrote:
> > On 09/30/2016 06:38 PM, Niels de Vos wrote:
> > > On Fri, Sep 30, 2016 at 07:11:51AM +0530, Pranith Kumar Karampuri wrote:
> ...
> > > Maybe we can add an additional tag that mentions all the people that
> > > did do reviews of older versions of the patch. Not sure what the tag
> > > would be, maybe just CC?
> > It depends on what tags would be processed to obtain statistics on review
> > contributions.
> 
> Real statistics would come from Gerrit, not from the 'git log' output.
> We do have a ./extras/who-wrote-glusterfs/ in the sources, but that is
> only to get an idea about the changes that were made and should not be
> used for serious statistics.
> 
> It is possible to feed the Gerrit comment-stream into things like
> Elasticsearch and get an accurate impression how many reviews people do
> (and much more). I hope we can get some contribution diagrams from
> someting like this at one point.
> 
> Would some kind of Gave-feedback tag for people that left a comment on
> earlier versions of the patch be appreciated by others? It will show in
> the 'git log' who was involved in some way or form.

I think this would be fair.

Reviewed-by tags should imho be reserved for the final
incarnation of the patch. Those mean that the person named
in the tag has aproved this version of the patch for getting
into the official tree. A previous version of the patch can
have been entirely different, so a reviewed-by for that
previous version may not actually apply to the new version at all
and hence create a false impression!

It is also difficult to track all activities by tags,
and anyone who wants to measure performance and contributions
only by looking at git commit tags will not be doing several
people justice. We could add 'discussed-with' or 'designed-by'
tags, etc ... ;-)

On a serious note, in Samba we use 'Pair-programmed-with' tags,
because we do pair-programming a lot, but only one person can
be an author of a git commit ...

The 'Gave-feedback' tag I do like. even though it does
not quite match with the foobar-by pattern of other tags.

Michael
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 163 bytes
Desc: not available
URL: <http://www.gluster.org/pipermail/gluster-devel/attachments/20161005/b91d3e07/attachment.sig>