Re: [jgit-dev] What do you think are the chances of a --changeid flag?

[james.moger@xxxxxxxxxxx]

I've posted an RFE to the Git mailing list if anyone would care to
participate.
http://permalink.gmane.org/gmane.comp.version-control.git/236429

-J


On Sun, Oct 20, 2013, at 09:27 PM, Shawn Pearce wrote:
> On Sun, Oct 20, 2013 at 5:33 PM,  <james.moger@xxxxxxxxxxx> wrote:
> > Clearly this is not the native git forum, but there is no doubt a lot of
> > cross-pollination. And perhaps this has already been posited/considered
> > by JGit team members
> >
> > What do you think are the chances of the _native_ git team offering a
> > --changeid flag for the commit command to generate a change-id and
> > replace an inserted "Change-id: I0000..." template value just like
> > EGit/JGit can do?  I know this is currently handled by a shell commit
> > hook, like the one Gerrit offers, but a hook for this is less than
> > ideal, IMO.  JGit definitely has improved on git in this scenario.
> 
> git-core might balk, but might surprise you and accept it. :-)
> 
> Linux kernel team balked and doesn't like the Change-Id field in their
> messages, as it has no "long lasting value and no relationship to a
> public database". I proposed using the Message-Id from the LKML email
> that kicked off a discussion on a patch, but they also were not in
> favor of this idea as it polluted the history visible in git log.
> *sigh*
> 
> Might be worth raising on git mailing list. I keep getting annoyed in
> Gerrit when I downloaded a new plugin repository and then need to add
> the hook.