Thanks for the feedback and positive support. If any of the 22 Core
Contributors to the Tools CG object to this selection, please contact
me before 1800 Pacific tomorrow, and we'll set up a vote. (I have
received no objections so far.) Otherwise, the proposal below should
be considered adopted.
- Stephen
----
OpenSolaris
DEFECT TRACKING SYSTEM SELECTION
Stephen Hahn
1. Summary
Based on the requirements, evaluations and experience gained with
various defect tracking systems (DTS), we, the Tools Community Group
(CG) Core Contributors, propose the adoption of Bugzilla, version
3.0 and higher, for OpenSolaris projects that require defect
tracking. It is expected that the Website project will deploy one
or more instances to be shared among projects and community groups
requiring such tracking.
2. Discussion
Since early summer 2007, we have been evaluating a number of defect
tracking systems (DTS) to select one to use for OpenSolaris issue
tracking.
The basis for our evaluation was the requirements document:
http://opensolaris.org/sc/src/website/spec/dts-requirements/d-dts-requirements.txt
As was suggested shortly after the requirements were developed, E2
"selective differentiated access" along with E0 "open source" proved
to be the primary filter that eliminated candidates. Of the set
considered--Bugzilla, Eventum, Mantis BT, and RoundUp--only Bugzilla
has default support for group access controls, meeting E2, and only
RoundUp appear to be easily customizable to add such support.
It seems risky to assume that the Tools CG would invest
time and effort to assume the competency to develop and maintain a
highly customized DTS derived from RoundUp, even though that might
result in a DTS candidate that most closely matches the
requirements. As with the distributed source code management (DSCM)
selection [1], we believe it more appropriate to accept a reasonable
fit with an active effort, and make modifications to terminology and
processes to reach an intermediate outcome.
The DTS to be used for OpenSolaris projects has to potentially
support the high defect update rates that have historically been
associated with the proprietary Solaris releases, which have
resulted in thousands of reports across hundreds of categories. We
anticipate similar activity around the OpenSolaris code as it
becomes used in a larger set of distributions. Because of its use
for Mozilla Corporation projects, we believe that Bugzilla is the
most suited system amongst the candidates for these circumstances.
It has also been important that its developers have been very
responsive, both during the evaluation and with projects elsewhere
using Bugzilla.
Therefore, we have decided to select Bugzilla 3.x as the DTS for
OpenSolaris. In the coming time, we hope to work with the Bugzilla
community to address any issues we may find while integrating
Bugzilla into the OpenSolaris infrastructure or while using it for
tracking our many, many bugs.
3. Acknowledgments
The initial requirements were derived in part from a draft developed
by Keith Wesolowski. Useful feedback on the evaluation came from
many community members, particularly Jim Carlson, Richard Lowe, and
Shawn Walker.
We also are very appreciative of Max Kanat-Alexander and timeless's,
well, timeless efforts in supporting our sandbox deployment of
Bugzilla, as well as their responses to the requirements document.
4. References
[1] B. Corwin, S. Hahn, M. Kupfer, and F. van der Linden, OpenSolaris
distributed source code management selection, 2006.
http://mail.opensolaris.org/pipermail/tools-discuss/2006-April/004484.html
--
[EMAIL PROTECTED] http://blogs.sun.com/sch/
_______________________________________________
tools-discuss mailing list
tools-discuss@opensolaris.org
