On Fri, 2013-12-20 at 13:02 -0500, Kamil Paral wrote: > > I didn't really mean to preclude the use of a minimal package set, I'm > > just trying to exclude the problems we've been having whereby the test > > cases and criteria are kind of getting 'gamed' with odd choices =) I'll > > see if I can find a happy medium... > > I think a general sentence like this could work fine: > "Please try not to adjust any installation settings that could make a potential bug in the tested area go unnoticed." > > Which means, if you combine too many stuff, maybe it'll blow up and you'll never know why, or maybe it will behave quite differently than in 'expected/usual' case. > > I don't think we need to counter criteria gaming inside test cases. We have the meetings to decide that. The very long and messy meetings... The initial goal of the criteria was to allow for very clear-cut and justifiable blocker decisions, rather than us having to make subjective calls after arguments about every bug. I'd like to restore/preserve/improve on that direction, myself. I like the case where a bug's discussion in the meeting goes: 01:00 topic change 01:01 five +1s 01:02 proposed #agreed 01:03 five acks 01:04 next bug not the case where we argue about how workaroundable it is for fifteen minutes... -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
