Re: Git terminology

From: Junio C Hamano <junkio@cox.net>
Date: Mar 29, 2006, 15:59 UTC
Message-ID: <7vodzp9r8r.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <20060329052113.6a773480.sebastien@xprima.com>

Sébastien Pierre <sebastien@xprima.com> writes:

> In the process of learning git, and browsing the docs, I
> realized that many particular terms are used, but I had
> difficulties finding their definitions.

I had difficulties reading your message.  Please fold your lines.

> Among those I found a bit arcane, there are (in no particular
> order) "refspec", "rebase", "rewound", "pickaxe", "refnames",
> "objectnames" ,"refnames", "fast-forward update".

I see some are missing from our glossary but some others are found there.

refspec is used by fetch and push to describe the mapping between remote ref and local ref. E.g. "git fetch $URL refs/heads/master:refs/heads/origin" means "grab the master branch head from there and store it as my origin branch head". "git push $URL refs/heads/master:refs/heads/to-upstream" for "publish my master branch head as to-upstream master head at $URL".

Probably we can just say "ref" where we say "refname".

Probably we can just say "ref" where we say "refname".

A fast-forward is where you have a revision and you are "merging" other's change that happen to be a descendant what you have. In such a case, you do not make a new merge commit but just update to his revision. That is:

          o---o---o---o his               o---o---o---o his
         /                      ->       /             \
    o---o---o---o yours             o---o---o---o-------* yours

updating to his needs a merge like above, but in this case:

          o---o---o---o his               o---o---o---o his = yours
         /                      ->       /
    o---o yours                     o---o

you just update to his.

A related concept is "being already up-to-date". In the latter picture, if he attempts to "merge" your changes and notices he is simply ahead of you (i.e. his revision is a descendant of yours), nothing happens.

> If some of you are willing to detail/explain/define those
> terms (and some others not mentioned here), I would be glad to
> collect the definitions and make a "git terminology" document.

Thanks, and what J Bruce Fields already said.