On Mon, Sep 24, 2012 at 6:17 PM, Mark Miller <[email protected]> wrote: > Personally, I think our system is better - use consensus wherever you can > avoid hard rules wherever you can. > > Consensus and Lazy Consensus, 3 +1's for a release. Who needs any more :) >
yeah, i dont want to be the one guy committing to this thing, I don't currently have the time to manage all the fixes, nor do i necessarily know enough about all the pieces involved to do the correct cost-benefit analysis. instead i'm asking that other people do this: maybe the 'correct' fix for some bugs is too risky to go in the release branch and a safe workaround makes more sense (we have 4.x/5.x for bigger changes, thats fine). and please be conservative, just open the issue with Fix version: 4.0 and Blocker, and take your time. let changes bake. add tests to ensure additional regressions are not introduced by fixes (I am concerned about that). run 'ant nightly-smoke' before committing to the release branch (because my jenkins is running this continously, and is configured to automatically revert commits that break the build). of course I want a solid release that isn't full of bugs, I am just of course pushing back on *all changes* because I'm trying to encourage stability. -- lucidworks.com --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
