Thanks! I did not know about that file. > On 23. Oct 2021, at 08:16, Đoàn Trần Công Danh wrote: > > On 2021-10-23 05:58:37+0000, "Schanzenbach, Martin" wrote: >> Hello, >> >> there is some odd behaviour in "git log" which affects me in various ways: >> For some reason, it "canonicalizes" the author/commiter information in our repository: https://git.gnunet.org/git/gnunet.git >> 5 years ago or so I was using another email address for commits. I have been using a different email address since that. >> >> But if you clone this repo, and take a look at a recent commit such >> as be9b998b3369056c7c8fb0ae84fb4e2a064b496c using "git log", you >> will see that my email address is actually the old one and different >> from the _real_ information. > > By old one, do you mean: > > Martin Schanzenbach > > It's because you have those entries in .mailmap: > > ---- 8< ---- > Martin Schanzenbach Schanzenbach, Martin > Martin Schanzenbach Schanzenbach, Martin > Martin Schanzenbach Martin Schanzenbach > ---- >8 ----- > > Which will map your: > > Martin Schanzenbach > > into it. > > >> You can see that I correctly set the author/commiter email here >> using either cgit >> (https://git.gnunet.org/gnunet.git/commit/?id=be9b998b3369056c7c8fb0ae84fb4e2a064b496c) >> or a specific format such as >> >> $ git log --pretty=format:"[%h] %cd - Committer: %cn (%ce), Author: %an (%ae) >> >> This is driving me crazy as the information is just plain wrong and >> when our CI tries to collect a blamelist when sending out >> notifications, it also uses the info from "git log" instead of the >> actual author/commiter emails, promptly sending the reports to the >> wrong account. > > The original value is recorded in the commit. You can retrieve raw > value with git-cat-file > >> >> Any idea what to do? > > Change your mailmap entries to: > > ---- 8< --- > Martin Schanzenbach Schanzenbach, Martin > Martin Schanzenbach Schanzenbach, Martin > Martin Schanzenbach Martin Schanzenbach > ---- >8 ---- > > >> I really do not want to rebase and rewrite >> thousands of commits with questionable success. >> Doing that locally also does not change the output from "git log", in case you are wondering. >> This problem will reemerge eventually whenever an author wants to >> change to another email address it seems so is this a bug? >> >> Any help/pointers appreciated. >> >> Martin > > > > -- > Danh