Hadn't seen anything here http://fedoraproject.org/wiki/FUDCon/FUDConF10 about testing related sessions at FUDCon and wondered if there was interest in having a few? (forgive me if someone has posted something like this already).
I think a lot of good things came out of the anaconda/snake sessions in RDU in January 2008 so would a couple more topics be good for Boston at the end of this week?
As a starting point here are some of the points that came out of the F9 postmortem (some of which I think we are already working on):
1) Document and reconsider installation testing approach 2) Do we need more RHT provided hardware? 3) Evaluate "rawhide" as a viable and worthwhile testing target 4) Better utilizing Anaconda's ability to install different repos of packages with a known good stage2 5) Test plans for other areas besides installation testing 6) Expanding testing--larger number number of consistent community testers 7) Exposing Fedora 10 Blocker Bugs a) identify earlier and more often b) make blocker bugs more visible to community 8) Increase effectiveness of Triage group at identifying blocker bugs a) Provide Training b) Create clearer documentation and guidelines around what types of bugs should be added to blocker bug --make policies broad so as to capture more bugs (err on the side of capturing too many bugs--can always remove) 9) Formal tracking of test plans and their execution with greater overall visibility into the process 10) Reward contributions from testers 11) More involvement of RHEL QE group--testing resources and expertise 12) Test/re-test installer issues earlier and more often during release. This depends on: a) test automation b) installable trees
On Mon, 2008-06-16 at 11:24 -0400, John Poelstra wrote:
Hadn't seen anything here http://fedoraproject.org/wiki/FUDCon/FUDConF10 about testing related sessions at FUDCon and wondered if there was interest in having a few? (forgive me if someone has posted something like this already).
I think a lot of good things came out of the anaconda/snake sessions in RDU in January 2008 so would a couple more topics be good for Boston at the end of this week?
Good timing! See below ...
As a starting point here are some of the points that came out of the F9 postmortem (some of which I think we are already working on):
- Document and reconsider installation testing approach
- Do we need more RHT provided hardware?
- Evaluate "rawhide" as a viable and worthwhile testing target
- Better utilizing Anaconda's ability to install different repos of
packages with a known good stage2 5) Test plans for other areas besides installation testing 6) Expanding testing--larger number number of consistent community testers 7) Exposing Fedora 10 Blocker Bugs a) identify earlier and more often b) make blocker bugs more visible to community 8) Increase effectiveness of Triage group at identifying blocker bugs a) Provide Training b) Create clearer documentation and guidelines around what types of bugs should be added to blocker bug --make policies broad so as to capture more bugs (err on the side of capturing too many bugs--can always remove) 9) Formal tracking of test plans and their execution with greater overall visibility into the process 10) Reward contributions from testers 11) More involvement of RHEL QE group--testing resources and expertise 12) Test/re-test installer issues earlier and more often during release. This depends on: a) test automation b) installable trees
Anyone interested in discussing points 1, 5, 6 and 9, I encourage you to join a hackfest session I'll be hosting @ FUDCON later this week. I'm anxious for more feedback/ideas!
Thanks, James