On 10/07/2015 04:10 PM, Fabrice Desré wrote: > Are you arguing for a gaia-inbound? That was consistently pushed back > by gaia-ers.
I'd be happy with it (especially if we just designated master to be gaia-inbound). I'm not really sure what issue people would have with it, aside from sheriffs/ops having more up-front work. > I disagree. Each app in isolation (except probably the system app) > is not more complex that big websites. We should be able to update > them almost per commit when there's no platform dependency. The platform is a common issue. Unlike most websites, we're using bleeding-edge features that often have performance issues at the Gecko level, not to mention serious bugs that can regress our apps. In fact, one of the more significant recent issues with the Music app (the scanning process stalling partway through) was the result of a Gecko regression. Because of that, I don't think we can treat each app as independent of the platform. Nearly all of our apps rely on some platform bit that doesn't apply to the average website, and those bits are not coincidentally the ones that we typically have the most problems with. In some ways, the bugginess in Gecko has just rolled downhill into Gaia where we don't have the same level of branch management to account for it. > It's different. A "smoketest blocker only" branch is more like an > aurora branch where we limit the backported fixes. I agree that if > it's working well enough we may keep it longer. An Aurora-like branch would be even better. - Jim _______________________________________________ dev-fxos mailing list [email protected] https://lists.mozilla.org/listinfo/dev-fxos
