----- Original Message ----- > On Aug 16, 2013, at 5:01 PM, Mark Finkle < [email protected] > wrote:
> > > Wait. For migration purposes, don't we need to keep new and old code > > > fully > > > functional? That was my belief when I authored this: > > > https://bugzilla.mozilla.org/show_bug.cgi?id=906011 > > > > > We can use preferences/flags, if needed, in the final shippable form, but > > we > > shouldn't use them to protect half-finished code from the rest of the code. > > That's the purpose of using a branch. > > So to be landable in m-c, the pref is a requirement. I guess you're just > saying that it's easier to add the pref at the end in the branch right > before landing? If we need prefs, then add them as the code lands in elm. Don't wait for the end. > Is this correct? If so, it blows up my iterative testing by jbonacci to > ensure no mainline regression thought. Mainline regression testing can happen in elm, since we are merging m-c into elm every few days. We should find regressions in elm before we land the code on m-c.
_______________________________________________ Sync-dev mailing list [email protected] https://mail.mozilla.org/listinfo/sync-dev
