Junio C Hamano <gitster@xxxxxxxxx> writes: > In a more perfect world, you would have sent the "--separator" thing > that already supported "--no-separator" from the get-go. You would > also have sent the 4th step that does not need ubsan fix. You do > not have the luxury to do so when your series hit 'next'---you need > to live with your earlier mistakes. > > Until we cross the release boundary, that is. OK, I hope to make the commit more cohesive before they release, but I ignored that the 'next' branch is already in the release period (because I think it might be possible to replace it in the 'next', but actually it shouldn't be done), sorry for the trouble it may cause you. > I'll revert the whole thing out of 'next'. Once the next cycle > opens after the dust from the release settles (i.e. around 10th of > June or so), ... OK. > ... please resubmit the series that does not need "oops, we > probably should have added '--[no-]separator' from the get-go" or > "oops, we had undefined behaviour here, and here is a fix", as if > you are a more perfect developer than you were when the version that > has been queued in 'next' was written (and in fact, you are, after > discussing the design and UBSan issues and stuff during the review). I think the v11 should be a version which is acceptable if there is no incoming review conversitions (The current v11 that separates '--no-separator' as a individual commit OR make 7/7 into 4/7, both approaches seem fine)? Thanks.
