[OpenAFS-devel] Gerrit reviews and the rate of acceptance

[Jeffrey Altman]

This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--------------enigD89D20CE128AC7D065C9A0C4
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

This week I have received private mail from several individuals which
are concerned about the rate of OpenAFS development.  There were several
topics raised across the e-mails and I'm going to pick out themes and
try to address them in separate mails to this list and openafs-info as
appropriate.

The first theme is that it takes too long for patches to be accepted
into OpenAFS.  As evidence of the overwhelming delays is the nearly 100
patchsets that are sitting in Gerrit for more than two weeks.  Many of
which have been sitting there for nearly two years.

Many of these patchsets are ideas that were sent to Gerrit either so
they wouldn't be lost or so others could use them as a starting point.
Some of the patchsets are items that have been reviewed but which have
outstanding issues which were never addressed.  But the vast majority of
the items sitting in Gerrit for an extended period of time are patches
that have simply not received sufficient reviews to decide what to do
with them.

The fact is that the number of active reviewers is exceedingly small.
Here are the statistics for the 4180 patches contributed to the 'master'
branch since Gerrit was deployed.

   3178     Reviewed-by: Derrick Brashear
   1355     Reviewed-by: Jeffrey Altman
    217     Reviewed-by: Simon Wilkinson
    155     Reviewed-by: Russ Allbery
    110     Reviewed-by: Andrew Deason
     68     Reviewed-by: Marc Dionne
     51     Reviewed-by: Rod Widdowson
     49     Reviewed-by: Asanka Herath
     43     Reviewed-by: Alistair Ferguson
     24     Reviewed-by: Peter Scott
     21     Reviewed-by: Matt Benjamin
     15     Reviewed-by: Tom Keiser
     12     Reviewed-by: Dan Hyde
     11     Reviewed-by: Michael Meffie
      9     Reviewed-by: Chaz Chandler
      6     Reviewed-by: Hartmut Reuter
      6     Reviewed-by: Benjamin Kaduk
      5     Reviewed-by: Stefan Kueng
      5     Reviewed-by: Garrett Wollman
      3     Reviewed-by: Jacob Thebault-Spieker
      2     Reviewed-by: Thomas L. Kula
      2     Reviewed-by: Steve Simmons
      2     Reviewed-by: Phillip Moore
      2     Reviewed-by: Mickey Lane
      2     Reviewed-by: Jonathan A. Kollasch
      2     Reviewed-by: Jeffrey Hutzelman
      2     Reviewed-by: Claudio Bisegni
      2     Reviewed-by: Adam Megacz
      2     Reviewed-by: Ken Dreyer
      1     Reviewed-by: sanket
      1     Reviewed-by: Tharidu Fernando
      1     Reviewed-by: Steven Jenkins
      1     Reviewed-by: Stephan Wiesand
      1     Reviewed-by: Rainer Toebbicke
      1     Reviewed-by: Jason Edgecombe
      1     Reviewed-by: Christof Hanke
      1     Reviewed-by: Chaskiel Grundman
      1     Reviewed-by: Chas Williams
      1     Reviewed-by: Alexander Ivan Redinger
      1     Reviewed-by: Alex Chernyakhovsky

All reviews are performed by volunteers.   It is nobody's day job to
perform Gerrit reviews.  Obviously Derrick, Russ and I as the
gatekeepers that review patches as part of the approval process have
performed the largest number of reviews.  A special note of thanks to
Simon, Andrew, Marc and Rod for being at the top of the non-gatekeepers
list.

In the last year alone there have been 44 contributors whose work has
been committed to openafs master.  Yet there are fewer than that number
of individuals who have reviewed someone else's work since Gerrit was
implemented.

No one is being paid to review code.  Not even the gatekeepers.  The
nearly 100 patchsets that are sitting in Gerrit represent approximately
2% of all of the patches that have been submitted.  The role of the
gatekeepers in the review process is important but they should not be
the sole review mechanism for OpenAFS submissions.

OpenAFS needs a much more active reviewer community.

If we were to impose a requirement that every patchset must receive a
total vote of +4 or +5 before it could be submitted, would that help to
obtain reviews?

The floor is open for discussion.

Jeffrey Altman






--------------enigD89D20CE128AC7D065C9A0C4
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)

iQEcBAEBAgAGBQJPhbidAAoJENxm1CNJffh4Zv4IAKFryhnTRMIZ7LLK8ao+4fRf
kDQnLAJXFPmB4kZO7DCxyQZmsDzzTemxrAVuCdufsTS6pf63tFof4ZSycZu1SST6
XC0mwyNENczE0D6eipdPOLYVjVsIb6x/QSc1SxjhWuMkaO2Ou8UPDYFm9Hmn3lpt
MxRNiSJTFOfajbsK9ALNUVH8AfYXbQ9jEp7E3s8WJnU1mDjD3rqRRw5Fw31EeAd4
2/gH27HvOzU57vF7nhkvOS7UgOwB68djDaxvbgqjWmoPW5wNKM5S5bsLiUTZAjZf
mQWT0FCklgCZzXlv3es6V8Yh8GHcBgquGygkYxMbYU5Od09M2/WtZHX53JjFBoA=
=oGrL
-----END PGP SIGNATURE-----

--------------enigD89D20CE128AC7D065C9A0C4--