deleting a named branch

[Martin Geisler]

Douglas Philips <dgou at mac.com> writes:

> On 2008 Jun 13, at 5:14 AM, Martin Geisler indited:
>> I like this idea. It seems lightweight and I think it would be perfect
>> if there was a good way to move bookmarks back and forth between
>> repositories.
>>
>> This includes deleting them in one repository and having the deletion
>> propagate to other repositories. But this of course goes against the
>> nice append-only design of Mercurial.
>
> How is that any different than "deleting" a tag?
> (Isn't that what Adrian's "closed" marker means:
> http://selenic.com/pipermail/mercurial-devel/2008-March/005607.html
> and "closed" seems to be more of a social/human 'marker' than the mere  
> absence of a marker on a downstream changeset...)

The difference would be that the bookmark is deleted from the meta
data repository (exactly like how a normal tag is deleted today), but
the effect of that is seen on the real repository.

So the bookmarks leave no permanent trace in the real repository, they
only exist in the meta data repository.

If it is only you who can pull the meta data repository, then the
bookmarks is effectively private to you, but if others can pull it as
well, then it becomes public.

> This whole issue/discussion is fascinating because of the tension
> between things (tags, branch names, bookmarks etc.) being local vs.
> global or in the history vs not in the history...

Yes, I also find this very interesting.

I just thought of another use for separating the file history from the
meta data: it might be nice to put the .hg/hgrc file in a meta data
repository too. I keep a number of path aliases in there, and whenever
I clone I have to copy the hgrc file manually. If it were cloned as
part of a meta data clone, then I wouldn't have to remember this.

-- 
Martin Geisler

VIFF (Virtual Ideal Functionality Framework) brings easy and efficient
SMPC (Secure Multi-Party Computation) to Python. See: http://viff.dk/.