{"thread":{"id":"35079","subject":"git cherry could honour \"cherry-picked from commit\" messages?","startedAt":"2013-10-07T20:16:18Z","lastAt":"2013-10-07T20:20:42Z","messageCount":2,"participants":["Adam Spiers"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"228600","messageId":"20131007201617.GC31277@pacific.linksys.moosehall","threadId":"35079","inReplyTo":null,"subject":"git cherry could honour \"cherry-picked from commit\" messages?","fromName":"Adam Spiers","fromEmail":"git@adamspiers.org","sentAt":"2013-10-07T20:16:18Z","receivedAt":"2013-10-07T20:16:18Z","isPatch":false,"sender":{"key":"git@adamspiers.org","avatar":"https://avatars.githubusercontent.com/u/100738?v=4"},"body":"Currently git cherry determines whether a commit has been merged\nupstream by calculating its patch-id and looking for that in the\nupstream branch.  Is there any reason why it couldn't also look for\n\"(cherry picked from commit ...)\" lines which are automatically added\nby cherry-pick's -x switch?  This would combat the blind spot in\ncherry's current approach concerning the case where the patch-id is\nchanged by the cherry-picking process (e.g. when the diff context\nchanges).\n\nMost likely there would be good reasons to make such behaviour\noptional (e.g. via git-config), but it seems like a waste for cherry\nto complete ignore these lines which provide useful information.\n\nThanks,\nAdam\n"},{"id":"228601","messageId":"20131007202042.GD31277@pacific.linksys.moosehall","threadId":"35079","inReplyTo":"20131007201617.GC31277@pacific.linksys.moosehall","subject":"Re: git cherry could honour \"cherry-picked from commit\" messages?","fromName":"Adam Spiers","fromEmail":"git@adamspiers.org","sentAt":"2013-10-07T20:20:42Z","receivedAt":"2013-10-07T20:20:42Z","isPatch":false,"sender":{"key":"git@adamspiers.org","avatar":"https://avatars.githubusercontent.com/u/100738?v=4"},"body":"On Mon, Oct 07, 2013 at 09:16:17PM +0100, Adam Spiers wrote:\n> Currently git cherry determines whether a commit has been merged\n> upstream by calculating its patch-id and looking for that in the\n> upstream branch.  Is there any reason why it couldn't also look for\n> \"(cherry picked from commit ...)\" lines which are automatically added\n> by cherry-pick's -x switch?  This would combat the blind spot in\n> cherry's current approach concerning the case where the patch-id is\n> changed by the cherry-picking process (e.g. when the diff context\n> changes).\n> \n> Most likely there would be good reasons to make such behaviour\n> optional (e.g. via git-config), but it seems like a waste for cherry\n> to complete ignore these lines which provide useful information.\n\nAlso, whether or not this approach works, it would also be useful to\nintroduce a config option for cherry-pick so that it adds the \"(cherry\npicked from commit ...)\" lines by default.  When introducing this\nconfig option, another command-line switch would be required to allow\nusers to negate it.\n"}]}