On Thu, 2018-01-11 at 13:45 +0000, Jason Bailey wrote: > Couldn't this also be phrased as being that the import be squashed or > rebased into a single commit prior to the import occurring?
Yes, it can be phrased like that. I am not sure whether attaching a source code file with a checksum is a hard requirement or just the most convenient way. Robert > > -----Original Message----- > From: Robert Munteanu [mailto:[email protected]] > Sent: Wednesday, January 10, 2018 6:39 AM > To: [email protected] > Subject: Re: Where do we put new git modules? > > EXTERNAL > > On Tue, 2018-01-09 at 16:55 +0100, Bertrand Delacretaz wrote: > > On Tue, Jan 9, 2018 at 4:48 PM, Stefan Seifert <sseifert@pro-vision > > .d > > e> wrote: > > > ...it makes it more difficult if the contribution was already > > > developed in a standalone git repo outside the ASF and perhaps > > > there > > > is already a commit history we want preserve... > > > > I don't think we are interested in the history of those donations > > in > > general. > > > > If a special case needs that we can always handle it in a > > different > > way. > > This got me thinking for a little bit. Donations were always done as > as source code dumps, and that's what was voted on. We did not vote > on histories, and a case can be made that source code histories are > less safe to import, as we don't review the whole history, but > instead the latest state - the patch. > > Unless strictly needed, I would suggest that we import just the > latest state, not the repository history. > > Robert
