Hi, I think we would need some way to make the commit more clear.
Currently committers are committed the patch in any format he/she want. E.g. using reviews.apache.org 's ticket number as subject, add some "Contributed-by:" or "Sent-by:" or "Wrote-by:" in the context to identify the contributor(but committer themselves remained "Author"), which is inconsistent across the committers, and hard for statistics. Also many commits lacks of proper explanation and subjects to be easily identified. So the best way to me, is let contributor listed as Author in the commit, and committer can "Signed-off" on it. I suggest we can follow the Linux Kernel community case on this issue. 1. the contributor should provide git-format-patch generated patch, which is able to indicate who is the author and what's to fix. 2. Committer should use git-am to apply the patch, add Signed-off-by: xxxx([email protected]) at end of it, to indicate who agreed and committed this patch(in fact there are more potential rules here, but let's stick with the simplest one first). Then author need to write the subject and more detail explaining of the patch if necessary, and it would be easier for others to track which patch is from who on what issue, and what's the solution. Another advantage/disadvantage of git-am is, it's very strict. You have provided up-to-date patch, otherwise it would fail. No fuzz is allowed. One problem currently is reviews.apache.org doesn't accept the git formatted patch, it would only accept diff. So if we want to get patch from reviews.apache.org, it would result in committer have to find subject and detail explanation for the patch, which may varies from author's, and also increase committer's workload. I think probably we can let contributor send out two mail for each patch? One attached latest patch, another one which is sent by reviews.apache.org, provided a convenient way for reviewing? The alternative way of doing is let contributor send out git-format-patch-ed patch(which would be based on the latest code) after committer reviewed patch and think it's ready to be applied. --Sheng
