On Fri, 2010-02-19, Greg Stein wrote:
> revnum is NOT inheritable.
> I prefer copied/moved-here rather than "added-with-history". The
> latter phrase is somewhat opaque to readers on what operation was
> actually performed by the user.
> The rest seems fine. Thx!
Thanks. Committed with those changes in r912537.
> On Fri, Feb 19, 2010 at 12:30, Philip Martin <philip.martin_at_wandisco.com> wrote:
> > Philip Martin <philip.martin_at_wandisco.com> writes:
> >> Greg Stein <gstein_at_gmail.com> writes:
> >>> revnum is NOT inheritable.
> >> At presnt working_node.parent_relpath says:
> >> /* parent's local_relpath for aggregating children of a given parent.
> >> this will be "" if the parent is the wcroot. NULL if this is the
> >> wcroot node. */
> >> parent_relpath TEXT,
> >> Does it make sense for a working node to be a wcroot?
> Hmm. You're right. I'm not sure how you could possibly get a wcroot
> into such a state. You're not allowed to delete it, nor copy/move
> something over it. ... well, maybe we *could*, but it would be quite
> weird to not do that from a checkout of the parent.
> > Sent too early. I meant to ask is it inheritable?
> Nope. parent_relpath is intended to be a key, so that you can easily
> find all nodes within a given directory. (which is also why we have
> I_PARENT and friends).
> Regarding inheriting repos_* values... I'm beginning to think that may
> be more trouble than its worth. The idea was to avoid having to
> compute them all the time, or to have to update them during a switch,
> etc. But... having to compensate for their absence is now seeming to
> be worse.
Received on 2010-02-22 12:40:18 CET