[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] hacking: document format for git commit messages

From: Eric Blake
Subject: Re: [PATCH] hacking: document format for git commit messages
Date: Mon, 27 Feb 2012 08:48:44 -0700
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:10.0.1) Gecko/20120216 Thunderbird/10.0.1

On 02/26/2012 02:02 AM, Stefano Lattarini wrote:
>> Do you really want to start requiring a Signed-off-by line, now?

I don't care either way, but it's slightly less work without.

>> I explicitly avoid such lines as redundant when they merely repeat
>> what's on the Author: line.
> I'm not truly sure about this; but:
>   - many other projects (linux, git itself) seems to use them, and I
>     believe there's a reason for this (even if I've failed to find it
>     so far);

git explicitly documents that 'Signed-off-by' has no inherent meaning to
git itself, but rather, should be assigned a per-project meaning.  Git
and Linux have assigned it the per-project meaning of 'I am in a legal
position to submit the portions of this commit that I touched'.  But FSF
projects already have stricter copyright assignments, so we are already
guaranteed that any commits in an FSF project are in correct legal
standing without an S-o-B line.

Libvirt is an interesting middle ground - it doesn't require strict
copyright assignments (and it's Red Hat's risk if this causes problems
down the road), nor does it require a Signed-off-by line on commits; but
at the same time, no one deletes a line from any commit pushed, even if
the line is "redundant" with the author.

>   - Ralf Wildenhues used the "Signed-off-by" as well (bit I never
>     asked him why); Eric Blake uses them too (Eric, if you are reading,
>     care to tell us why?);

Habit, because I've submitted to git before.  But I can drop them, just
as easily, especially if we make it project policy (as was done on
coreutils - where you will notice that my commits there lack the line).

>   - last but not least, I'd like to start using the various "Acked-by",
>     "Reviewed-by", "Tested-by" etc. lines in the future as well, so having
>     also a "Signed-off-by" line seems more consistent.

I don't care whether you require it, require it to be omitted, or (like
libvirt) leave it optional; as long as we document _this_ project's
policy, then people can live by the policy.

Eric Blake   address@hidden    +1-919-301-3266
Libvirt virtualization library

Attachment: signature.asc
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]