Re: git history buglet - I'm not the author of /that/...

[Ralf Wildenhues]

[ <http://thread.gmane.org/gmane.comp.gnu.libtool.general/9887> ]

Hi Peter,

* Peter Rosin wrote on Wed, Jan 14, 2009 at 01:40:15PM CET:
> For those who cares, because they implement a cvs->git converter
> or otherwise, I have noticed a minor problem in our git history.

To be more precise, it wasn't the cvs -> git conversion that broke this,
but the 'git filter-branch' script that I used on top of that in order
to "improve" patch author attribution.

> Commits [1] and [2] are recorded we me as author, but I think
> that they in fact have been written by Gary [4]. The conversion
> may have been sidetracked by the fact that Gary made three
> commits but only made one ChangeLog entry (he finished with [3],
> which is the commit that carries the ChangeLog entry).

Indeed, that is what happened.

> I think I got attributed since I was the author of the last
> commit before Gary commited these patches.

Yep.

> So, what I'm really trying to say is: don't blame me when you
> think that libtool.m4 is somewhat mysteriously located inside the
> libltdl tree! ;-)
> (I mean, libtool.m4 isn't exactly local to libltdl...)

Certainly.

> Is it useful to report nitpicks like this, when the cost of
> fixing is so high? I assume that we don't want to tamper with the
> git history for something as minor as this, right?

Yes, it is useful to report such issues, if only for future
filter-branch scripts; thank you for doing so.  And no, we certainly
won't tamper with published git history of Libtool for this issue.  If
you would like, we could either add a note somewhere in the tree, or
the web browser of the code, or we could look into adding a graft that
"fixes" authorship.  Haven't done anything like the latter before, but
AIUI grafts are the hacks to do things like that.

The easiest for us would be if you and Gary (and we) just didn't mind
all that much.  We'll certainly blame Gary on leftover bugs stemming
from those changes; those reorganizations were prominent enough not to
leave any doubts as to their author.  :-)

I would like to apologize for not having been diligent enough back then
to spot these glitches in the filter-branch script.  Should have caught
these instances, and it wouldn't have been all that hard, either.

Cheers,
Ralf