Am Thu, 08 Jan 2015 16:42:14 +0100 schrieb Gilles <[email protected]>:
> On Thu, 8 Jan 2015 10:24:49 -0500, Gary Gregory wrote: > > On Thu, Jan 8, 2015 at 9:24 AM, Luc Maisonobe <[email protected]> > > wrote: > > > >> Le 08/01/2015 07:59, Thomas Neidhart a écrit : > >> > On 01/08/2015 02:40 AM, Phil Steitz wrote: > >> >> > >> >> > >> >> > >> >> > >> >>> On Jan 6, 2015, at 3:10 PM, Gilles > >> <[email protected]> > >> wrote: > >> >>> > >> >>> Hi. > >> >>> > >> >>> Do we head towards 4.0, starting to implement the dreaded > >> >>> breakings? ;-) > >> >> > >> >> How about starting a new 4.0 branch and keeping head set up for > >> 3.x so > >> we can cut 3.4.1 when the bugs start coming in against 3.4. > >> > > >> > +1 > >> > >> Of course a branch is needed, but why consider master should be > >> 3.x and > >> 4.0 should be a side branch. > >> > >> If we consider 4.0 is the future and 3.x is only for bug fixes, > >> then 4.0 > >> should be master and 3.x a side branch. This side branch could be > >> created later on, if needed. It is straightforward to create new > >> branches in git even from very old commits, so the branch creation > >> can > >> be delayed until needed (and only if needed). > > What would happen if we then need to merge (backport) a modification > performed under the "o.a.c.math4" hierarchy into the "o.a.c.math3" > hierarchy of that branch? > Would two files "o/a/c/math4/SomeAlgo.java" and > "o/a/c/math3/SomeAlgo.java" > be considered the same w.r.t. to "diff"? Or would we need to > manually copy > from "math4"? Git does detect rename-and-modify, however it is an heuristic so you cannot rely on it. But at least interactively you can diff aribritary file/revisions. (internally of course Git does not store diffs anyway, so it is just a representation of a commit - which is also used in some situations for merging). Gruss Bernd --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
