# Re: An alternate model for preparing partial commits

1 messages from 2008-06-28 to 2008-06-28. Participants: Stephen R. van den Berg.
Thread: https://gitlist.dev/t/14199

## Stephen R. van den Berg, 2008-06-28 09:30

Subject: Re: An alternate model for preparing partial commits
Message-ID: <20080628093028.GA26745@cuci.nl>
URL: https://gitlist.dev/e/20080628093028.GA26745%40cuci.nl

```
David Jeske wrote:
>- I could supercede 2-commits with 1, effectively making the bad-commit
>disappear in the linear history. Users who already have the history, however,
>would be unaffected, because the start/end endpoints are the same.

Superseding does not work in git trees and is a disruptive operation to
*all* commits following the point being superseded.
Also, superseding preserves the startpoint, but *always* alters the
endpoint.

I suggest you take a look at how a git-repository is stored and learn
about how the SHA1-hashes interact with the stored repository.
-- 
Sincerely,
           Stephen R. van den Berg.

Heisenberg might have been here.

```