This is the mail archive of the
newlib@sourceware.org
mailing list for the newlib project.
Re: CVS or git now?
- From: "Cary R." <cygcary at yahoo dot com>
- To: Peter Rosin <peda at lysator dot liu dot se>, Joel Sherrill <joel dot sherrill at oarcorp dot com>, "newlib at sourceware dot org" <newlib at sourceware dot org>
- Date: Tue, 12 Nov 2013 09:49:03 -0800 (PST)
- Subject: Re: CVS or git now?
- Authentication-results: sourceware.org; auth=none
- References: <52812F85 dot 2030205 at oarcorp dot com> <5282300E dot 3040101 at lysator dot liu dot se>
- Reply-to: "Cary R." <cygcary at yahoo dot com>
On Tuesday, November 12, 2013 5:42 AM, Peter Rosin wrote:
> On 2013-11-11 20:27, Joel Sherrill wrote:
>
>> With all the recent conversion to git of the old src/
>> tree, I am curious where newlib stands.
> I had a brief browse of https://sourceware.org/git/?p=newlib.git
> and that "mirror" looks poor. It has not retained authorship of
> the commits, as the committer is always listed as author. Here
> is one recent example:
If the authorship is not being maintained than git is not being
used correctly.
If the patch is being applied the normal git way then git am will
preserve the authorship. The problem with this method is dealing
with conflicts can get complicated.
If the patch is being applied using patch and then being committed
once it has been reviewed and any conflicts dealt with then the
--author flag can be used with git commit to preserve the original
authorship information.
git pulls from remote branches are another way to get patches into
the master repository, but that's more advanced.
I'm guessing the authorship problem mentioned above is because
more traditional methods (patch, etc.) are being used to apply
and review the patches and the original author information is
not being passed to git commit. This is easy to forget when you
are applying a bunch of patches since you need to remember an
extra step to preserve the original author information.
Cary