From: Michael J Gruber Date: Fri, 19 Jun 2015 08:34:23 GMT Subject: Re: Visualizing merge conflicts after the fact (using kdiff3) Message-ID: <5583D40F.7030300@drmicha.warpmail.net> In-Reply-To: Junio C Hamano venit, vidit, dixit 18.06.2015 17:57: > Michael J Gruber writes: > >> This type of request comes up often (for a reason). I'm wondering >> whether we could support it more systematically, either by exposing the >> steps above as a command, or by storing the unresolved merge somewhere >> (leveraging stash or rerere). > > Perhaps 'tr/remerge-diff' (on 'pu') is of interest? > Ingenious! To me, this seems to be the most useful view if you want to understand a merge just from the parents and the merge commit. Since you would use that for individual commits only, the cpu cycles are well spent. As and added benefit, tr/remerge-diff merges to current next with conflicts (oid...) so that you get to test it on its own merge! I haven't reviewed remerge-diff but merged it on top of my own local additions and ran the full test suite successfully. Any big blocker to watch out for? Michael