{"thread":{"id":"28934","subject":"mergetool's test for rerere seems to be incorrect","startedAt":"2011-11-14T20:26:18Z","lastAt":"2011-11-14T20:26:18Z","messageCount":1,"participants":["Erik Carstensen"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"179441","messageId":"CALeBojRuZP3dbEs=rPX1E5p=oD43_yJO_Ec=Pe2SRpjG-rH2Vg@mail.gmail.com","threadId":"28934","inReplyTo":null,"subject":"mergetool's test for rerere seems to be incorrect","fromName":"Erik Carstensen","fromEmail":"mandolaerik@gmail.com","sentAt":"2011-11-14T20:26:18Z","receivedAt":"2011-11-14T20:26:18Z","isPatch":false,"sender":{"key":"mandolaerik@gmail.com","avatar":null},"body":"Hi,\n\nOccasionally when I use mergetool in a conflicted state ('git status'\nsays it's a conflict), nothing happens.\n\nBy bisection I found that bb0a484e985ef8d9bbbbeb172b1fcf4982634bef is\nthe offending commit.\n\nI have rerere.enabled unset, but I did find an empty file MERGE_RR in\n.git/, which explains my problem: mergetool chooses to use rerere\nbecause it finds the MERGE_RR file, but rerere just exits silently\nbecause it's not enabled.\n\nWouldn't it make sense to restrict mergetool's use of rerere to the\ncase when rerere.enabled is set?\n\nI'm also a little confused about why the MERGE_RR file appeared there\nin the first place. I can see that empty MERGE_RR files, with\ndifferent timestamps, have ended up in various other of my git\nrepositories as well. I never used rerere in these repos, and the only\nnon-git tools I have used for version control are stgit and kdiff3.\n\nErik\n"}]}