Rhialto <rhialto%falu.nl@localhost> writes: > On Tue 14 Jul 2015 at 12:27:02 -0400, D'Arcy J.M. Cain wrote: >> On Tue, 14 Jul 2015 10:04:56 -0400 >> "Hugh Meyer" <hmeyer%meyermat.com@localhost> wrote: >> > One aspect of GIT is that it allows history to be changed and does >> > not track who made the change. >> >> Is that true? Sounds like deal breaker to me. Are you sure that that >> behavior can't be controlled by the configuration? > > No it is definitely not true. Each state in the repository has an ID, > based on a hash, and > > The writer above maybe is confusing "changing history" with "creating a > completely new and recognizably different chain of history based on some > combination of other parts of the history" (git rebase). This only adds > new history, clearly identified by who did it, and it doesn't remove > what was there. People who have the original history still have it. > Operations like this are meant to be used before the original history is > published to any other repository, since this can cause confusion. One can also modify the repository to remove commits or remove files From commits (see "git filter-branch"). But doing this produces new hashes of all commits, and it's pretty obvious. But this is a silly claim, because in any system, you can modify the underlying database and change whatever you want. cvs has "admin -o". In svn, one can surely dump/edit/restore.
Description: PGP signature