[Gnu-arch-users] Re: tla import versus tla commit

[Stefan Monnier]

>> I think we'll be better off following an approach similar to NNTP (where the
>> problem is also that once data is sent, you can't in general trace back all
>> the copies): create a new kind of changeset which modifies a previous one.

> Ick.  Sorting out such a collection of changesets by hand seems
> horridly tedious and, operating globally (rather than in the
> special-case local form I advocate for above) in the arch namespace
> but extended with some kind of undo semantics sounds hopelessly
> intractible (what is to become of mirrors, for example?  When can they
> safely be used in your Brave New World other than when an
> authoritative master server can be contacted?).

Huh?  Mirrors would work just as they do now: if you've updated them since
the `undo', they'll have the undo applied, otherwise they won't.
The undo is a change propagated just like the others.

Of course this change will not be a 100% undo in that it will leave a trace.
E.g. the undo-change itself will have to be kept around so it can be
propagated to further mirrors.

And obviously, we can't allow just arbitrary archive changes via such
special changesets.  Only some specific ones like `undo' or `updatelog'.


        Stefan