On Fri, Aug 23, 2013 at 10:44:03AM -0700, Junio C Hamano wrote:

> My initial reaction was "Why should something as important as 'git
> commit' should be playing a guessing-game?" ;-) and I am kind of
> ashamed to have added 146ea068 (git commit --author=$name: look
> $name up in existing commits, 2008-08-26) and then am embarrased to
> have completely forgotten about it. I never use the feature myself.
> But for that old and established "--author parameter that does not
> use the standard format guesses" feature to be useful, I agree that
> it should honor the mailmap.
> I wonder if it would hurt anybody if we made this unconditional, not
> even with "--no-mailmap" override? Opinions?

I think it would be OK. You can always override by giving the actual
full address you want instead of a partial one. And if somebody is not
up to date in the .mailmap file, maybe this would be a good hint that
you should take care of that. :)

I paused for a second, thinking that such advice might not be good for
people who do not want to make an official change to upstream's
.mailmap (e.g., because they do not want to pollute a long-running fork
that will need to merge from upstream, or do not want to pollute a topic
branch with an unrelated commit). But I forgot that we have
mailmap.file, if they want something custom.

So I think anyone for whom the mailmap lookup does not provide the right
answer will fall into one of two groups:

  1. A one-off, which can be overridden by specifying the address you
     do want.

  2. Somebody you will be mentioning frequently; bother to set up
     a mailmap.file.

As an aside, it seems silly that we do not respect $GIT_DIR/mailmap by
default, even without a config option. But I doubt that anybody cares
too much, if nobody has raised the issue in all of these years.

To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to