Messages in this thread | | | From | Martin Fick <> | Subject | Re: Re* Regulator updates for 3.3 | Date | Mon, 16 Jan 2012 16:43:22 -0700 |
| |
On Monday, January 16, 2012 04:33:00 pm Junio C Hamano wrote: > With your suggestion, they need to export > "GIT_MERGE_EDIT=0" today, and they will need to update > again to export "GIT_MERGE_SOMETHINGELSE=0" when such an > incompatible change comes. > > With a single "GIT_MERGE_LEGACY=YesPlease", they can be > future-proofed today and will not be affected when we > make another incompatible change. > > So I am not sure why separating the big-red-switch into > smaller pieces would be an improvement, especially wnen > the scripts that want to specify finer-grained control > of features can use "--[no-]edit" options to explicitly > ask for it.
Then, what would I do if I write a script which uses the new edit functionality (without even being aware that there was an old way) and you introduce a new incompatibility? I can't turn on GIT_MERGE_LEGACY then since it would revert to behavior which my script would not expect (since it was written after the current incompatibility, but before the new one)!
-Martin
| |