Yep, we have been indeed using the msg (author name) historically. In cases where author info is there, the guideline is to use git am --signoff.
I don't think we should require git comit --author, as long as there is attribution in the commit msg. But we can do a --author as an option. Enis On Thu, Jan 26, 2017 at 4:29 PM, 张铎(Duo Zhang) <[email protected]> wrote: > See here, our committer guide says that the commit message format should be > HBASE-XXX XXX (the actual author). > > http://hbase.apache.org/book.html#_commit_message_format > > I think the rule was setup when we were still on svn, but we do not change > it. > > I‘m a big +1 on always using 'git am' if possible, and set > author explicitly when using 'git commit'. One more thing is that, do not > forget to add --signoff :) > > Mind opening a issue to modify the above section in hbase book, Appy? > > Thanks. > > 2017-01-27 4:12 GMT+08:00 Chetan Khatri <[email protected]>: > > > Thanks , Appy. For understanding because i am very new to open source > > contribution. > > > > On Fri, Jan 27, 2017 at 1:25 AM, Apekshit Sharma <[email protected]> > > wrote: > > > > > Hi devs, > > > > > > A recent question by new contributor (thread: Git Pull Request to > HBase) > > > and looking at history of commits, i think it is worth re-iterating > that > > > committer should use *git am* while committing patches so that whoever > > did > > > the actual work gets appropriate credit in git history. > > > In case the patch uploaded is unformatted patch (without author tag), > > > please use > > > *git commit --author=<foo>.* > > > I don't think we should be lax about it as a community if the original > > > author doesn't get appropriate credit. > > > > > > Thanks > > > -- Appy > > > > > >
