Re: ancestor path / revision

Branko =?ISO-8859-2?Q?=C8ibej?= <brane@xbc.nu> writes:

> Ben Collins-Sussman wrote:
>
> > Now, it's a separate question about how these absolute coordinates are
> > communicated. As Karl suggested, only the "root" of the working copy
> > may store the absolute path, and when driving a commit editor, the
> > editor callbacks are called with an assumption that all parent
> > directory batons' paths are being implicitly prepended. But that's
> > another topic.
>
> Wouldn't that make operating on a sub-tree of the working copy a pain?
>

Probably. In the status quo, the working copy is storing absolute
repository pathnames for each entry in the `entries' file. Similarly,
when driving a commit-editor, it's supplying absolute pathnames for
each object too. I like this. :)
Received on Sat Oct 21 14:36:17 2006