Claiming unlinked commits

[jayvdb]

Before there was GitHub, and git, there were committers who accepted patches and merged them under their own identity, and usually credited the author in the commit message.

It would be nice to have a mechanism for a user to manually list them, perhaps a json file in their <user>.github.io repo.

They could be presented differently, and ensure that the reader can manually check them.

Alternatively we could build a separate repo where these files are placed, and have a verification process where other ghuser users review PRs to check them , and consult mailing list archives or the committers to verify the claims.

[lourot]

Alternatively we could build a separate repo where these files are placed, and have a verification process where other ghuser users review PRs to check them , and consult mailing list archives or the committers to verify the claims.

That sounds great! Do you own such commits?

[jayvdb]

Ya, lots. Finding them is the hard part .. ;-)

[jayvdb]

Probably the easiest one I can think of is quilt -- long history, but small repo, so should be easy to pull out old data.

e.g. https://github.com/yashi/quilt/commits?author=jayvdb are two properly linked commits, and I dont recall how many should be in there. The real repo is on https://savannah.nongnu.org/git/?group=quilt , not sure if there is an official mirror on GitHub yet.

The current NEWS file doesnt retain the old history. There was a file quilt.changes which could be consulted for better evidence. Might even need to find an old copy of the CVS repo.

http://lists.nongnu.org/archive/html/quilt-dev/2006-01/msg00114.html is example of submitting patches the wrong way ;-)

So they are around here somewhere.

https://github.com/yashi/quilt/commits/master?after=0f2a913751c8e37a190b783decfa933f199b48cd+769

e.g. yashi/quilt@34ce343f9d4cf204dbfcbf8