Am 01.12.2013 20:04, schrieb Dennis Kaarsemaker: > We always ignore anything named .git, but we should also ignore the git > directory if the user overrides it by setting $GIT_DIR > > Reported-By: Ingy döt Net <ingy@xxxxxxxx> > Signed-off-by: Dennis Kaarsemaker <dennis@xxxxxxxxxxxxxxx> > --- > dir.c | 2 +- > t/t7508-status.sh | 7 +++++++ > 2 files changed, 8 insertions(+), 1 deletion(-) > > diff --git a/dir.c b/dir.c > index 23b6de4..884b37d 100644 > --- a/dir.c > +++ b/dir.c > @@ -1198,7 +1198,7 @@ static enum path_treatment treat_path(struct dir_struct *dir, The special case for ".git" is hardcoded in many places in git, including the line immediately above this diff hunk. So I figure that GIT_DIR is not meant to _rename_ the ".git" dir, but to point somewhere _outside_ the worktree (or somewhere within the .git dir). If we want to support the rename case fully, I think there are a few more questions to answer (and a few more places to change), e.g.: - What if GIT_DIR=.foo and someone upstream adds a ".foo" directory? - Should it be possible to track ".git" as a normal file or directory if its not the GIT_DIR? - What about other commands than status, e.g. does 'git clean -df' leave the GIT_DIR alone? If we don't want to support this, though, I think it would be more approrpiate to issue a warning if GIT_DIR points to a worktree location. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html