Hi all,
Whilst fixing the git-format-patch "Signed-off-by identity bug", I decided to have a look at the other calls to setup_ident(), to check for similar problems. I noticed a couple of things that didn't look quite right (see also next email).
In particular, http-push accesses git_default_email directly, using it as part of a (DAV?) LOCK_REQUEST. However, since http-push does not call git_config(), only the generic <user@localhost.localdomain> determined by setup_ident() will be used.
Now this may not be a problem. I don't really know what the email address is used for in the lock request; it could have no effect on the transaction at all, apart from perhaps some message text in the web-server logs.
Also, there may be a reason for not reading the config file. I can't think of one, but hey ...
Unfortunately, since my config has NO_CURL, I can't even compile http-push.c, let alone test the change given below.
comments anyone?
Ramsay
diff --git a/http-push.c b/http-push.c index e281f70..4de62e4 100644 --- a/http-push.c +++ b/http-push.c @@ -2340,6 +2340,7 @@ int main(int argc, char **argv) setup_git_directory(); setup_ident(); + git_config(git_default_config); remote = xcalloc(sizeof(*remote), 1);