Jiang Xin <worldhello.net@xxxxxxxxx> writes: > From: Jiang Xin <zhiyou.jx@xxxxxxxxxxxxxxx> > > ## Changes since v18 > > 1. This series is based on "next" branch, in order to resolve a conflict > with commit 95e7c38539 (refspec: make sure stack refspec_item > variables are zeroed, 2020-08-14). See patch 9/10. We cannot ever merge such a topic that depends on 'next' down to 'master' without dragging all the other topics that present in 'next', plus the merge commits that bring these topics to 'next'. In this particular case, I think the conflict resolution is trivial and more importantly, it is not a *new* conflict v19 introduces but did not exist with v18. IOW, what is in 'seen' has already resolved the same conflict between 95e7c38539 and this topic. I've applied these directly on 'next' (call it c0), rebased the result on the same base as v18 is queued on, and then merged the rebased v19 (call it c1) with 'next' to make sure that the result (call it c2) matches, i.e. tree of c0 and tree of c2 are identical. I'll use c1 to replace what is queued in 'seen'. A workable alternative would be to base these on top of the topic that contains 95e7c38539 (i.e. jk/refspecs-cleanup). Then, this topic is still taken hostage by that 'cleanup' topic, but at least as long as that topic graduates, this topic can graduate to 'master' without having to drag any other cruft with it. Thanks.
