On Sat, Mar 30, 2019 at 12:16 AM Christian Couder <christian.couder@xxxxxxxxx> wrote: > > 6. `[Bonus]` If everything goes well and time permits, discuss with the mentor(s) > > the possibility of deprecating the am backend of rebase. This point is last to > > work on as it provides no "cosmetic" difference on the user side. Elijah > > [mentioned][12] the possibility of a "social" problem that might occur which shall > > be discussed then. > > Maybe it would be interesting to see if there is a performance > difference between the 2 backends. Yes, it may make sense to repeat https://public-inbox.org/git/nycvar.QRO.7.76.6.1901312310280.41@xxxxxxxxxxxxxxxxx/ to verify we're still in a good state, and maybe even try to figure out why split-index is weird. However, more important is the handling of empty commits (both start-empty and become-empty) and trying to bring some consistency there or at least implementing more options for the interactive backend and figuring out which empty-handling options should be the default for which sets of flags. It's kind of a mess.
