[mpich-devel] [mpich-discuss] Use of Signed-off-by in MPICH

[Michael Blocksome]

What you've said is exactly why we add the "Signed-off by" clause. There 
are only two pieces of information and therefore all that gets retained is 
the first (author) and last (committer). 

Furthermore, whether or not a cherry-pick is even needed depends on if the 
"source" branch has been completely rebased  to master or not.  If not, 
then the commit needs to be changed in the process of moving to master 
(which resets the committer), but if it is a strict fast-forward of master 
then the commit is not altered (committer remains the same) and the HEAD 
pointer of master is merely moved to point to the new commit. If you are 
trying to have the commiter != author, then you'd have to mess around with 
the commit in this case.

It seems to me that what you really don't like is the "rebase" style of 
git workflow and would prefer a "merge" style instead? There are 
advantages and disadvantages to both - which are well documented in flame 
wars all across the internet and don't need to be repeated here. Whether 
some commit says "Signed-off" or "Reviewed" or nothing is really not a big 
deal - as long as the convention is documented by the community and 
followed.

Michael Blocksome
Parallel Environment MPI Middleware Team Lead, TCEM
POWER, x86, and Blue Gene HPC Messaging
blocksom at us.ibm.com




From:   Jed Brown <jedbrown at mcs.anl.gov>
To:     Michael Blocksome/Rochester/IBM at IBMUS, 
Cc:     devel at mpich.org
Date:   01/06/2014 02:50 PM
Subject:        Re: [mpich-discuss] Use of Signed-off-by in MPICH



Michael Blocksome <blocksom at us.ibm.com> writes:

> In our workflow, often the commit is pushed to an ibm review branch, 
then 
> the signer (usually me) does a signed-off cherry-pick to an anl review 
> branch and pushes the code to mpich-ibm git repository for ANL to 
> integrate. Using "signed off by" is usually much easier than playing git 

> games to ensure that the author and committer are different.

When you cherry-pick, rebase, or amend, Committer will be automatically
reset; no "games".  The problem with that is there are only two pieces
of information (Author and Committer), so if Pavan (for example)
cherry-picks from your branch, the semantic information that you were a
middle-man would be lost [1].  If that semantic information is about
review rather than IP, I would use Reviewed-by instead of Signed-off-by,
but whatever the convention is/becomes, documentation and consistency is
helpful.


[1] If he merges from your branch instead of cherry-picking, then your
commit does not get rewritten and his name would be on the merge commit.
[attachment "attbohle.dat" deleted by Michael Blocksome/Rochester/IBM] 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/devel/attachments/20140106/b7e3df9b/attachment.html>