Talk:FirefoxSummit/2006/ProposedSessions/Firefox 2 Postmortem
One thing that wasn't mentioned in the meeting.. Patches first landed on trunk and then after a few days of baking they landed on the branch (good). But there was at least once when the trunk was left burning for literally days, and checkins were still landing.
Eventually, the trunk was closed and the problem fixed, but this made finding regression ranges much harder. It also meant there was a bundle of landings on the branch a few days later, which again doesn't help with regression finding.
I would have liked to have seen the trunk closed much sooner (maybe even an auto-close after 24hrs of burning or something) so this situation can't arise in the future.