Finding a fork point

[Halbert, Daniel - 0662 - MITLL]

From: mercurial-bounces at selenic.com [mailto:mercurial-bounces at selenic.com]
On Behalf Of Paul Sargent
> I have a repository (in this case it's the linux kernel) with a very
> branchy history. I also have a tar-ball where somebody has forked at
> some point, done some changes, and then released the whole source
> tree with all the version control information stripped out (You have to
> love how some companies follow the letter of licences, but not the
> spirit of them).

Is there any kind of a version file somewhere in the kernel tree? Or if not
a version for the whole thing, maybe for various components that are under a
lot of development?

My general thought would be to take checksums of files in the tarball, and
then compare them with checksums of the same set of files at various branch
points. You want to to pick the point with the maximal number of matches. I
think you could automate this with bisect, but you'll need to bisect
starting at a bunch of the branch tips. There is not necessarily going to be
a single point that is clearly the branch point, the same files can change
on a number of different branches.

Dan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 5285 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20100629/f8fd9ae4/attachment-0002.bin>