"Constantine A. Murenin" [off-list ref] writes:
I have two distinct trees that were not managed by any RCS, and I'd
like to import them into a single repository into two separate orphan
branches, then make sense of what's in there, merge, and unify into
'master'.
(To give some context, it's /etc/nginx config files from nginx/1.0.12
on Debian 6 and nginx/1.2.2 on OpenBSD 5.2.)
As these come from two totally separate sources, I'd find it more
natural to do two repositories, deb-nginx-conf and obsd-nginx-conf,
each with one commit and then pull one into the other (or pull both
to master-nginx-conf if you really wanted to), to me.
On 5 February 2013 14:29, Junio C Hamano [off-list ref] wrote:
"Constantine A. Murenin" [off-list ref] writes:
quoted
I have two distinct trees that were not managed by any RCS, and I'd
like to import them into a single repository into two separate orphan
branches, then make sense of what's in there, merge, and unify into
'master'.
(To give some context, it's /etc/nginx config files from nginx/1.0.12
on Debian 6 and nginx/1.2.2 on OpenBSD 5.2.)
As these come from two totally separate sources, I'd find it more
natural to do two repositories, deb-nginx-conf and obsd-nginx-conf,
each with one commit and then pull one into the other (or pull both
to master-nginx-conf if you really wanted to), to me.
Yeah, I guess it might be more of a git-style to have two/three
separate repositories here. (The sources are just a couple of files,
so I think my specific example still calls for merely two orphan
branches.)
Still, is it really expected that you can't create an orphan branch in
an empty repository? On the outside, this sounds like a rather benign
bug.
C.