[Gnu-arch-users] Re: tag --seal - where does version-0 go?

[Matthew Palmer]

On Wed, Sep 29, 2004 at 07:07:01AM +0200, Christian von Kietzell wrote:
> On Mon, 27 Sep 2004 23:18:51 +1000, Matthew Palmer <address@hidden> wrote:
> > 
> > On Mon, Sep 27, 2004 at 03:13:37PM +0200, Christian von Kietzell wrote:
> > > But in case of base-0 no version-0 will be created, base-0 will be
> > > sealed directly.
> > > Something I find rather confusing...
> > 
> > I would find that incredibly confusing too -- in fact, it might explain why
> > I thought a test tag --seal I'd tried didn't appear to work, which (in part)
> > triggered my initial question...
> 
> Ok, guess what. I tried exactly this and when I tried to tag something
> afterwards (with --fix) tla complained that I couldn't use --fix
> before using --seal. So obviously, base-0 doesn't get sealed at all
> and --seal doesn't work when the continuation revision is base-0.

I think (and this is kinda confirmed by what I've just been told on IRC) it
helps to think of tag --seal and --fix are a sort of "commit into other
tree" type thing.  You can do a tla commit --seal on your local tree, and
it'll seal the branch underlying that tree, or you can tla tag --seal src
dest to seal the dest branch.  But you can't seal a branch that doesn't
exist -- hence the problem of getting a base-0 when you asked for --seal on
a new branch.

Good sigmonster.  Maybe Arch is just mathematics.

- Matt

-- 
"After years of studying math and encountering surprising and
counterintuitive results, I came to accept that math is always reasonable,
by my intuition of what is reasonably is not always reasonable."
                -- Steve VanDevender, ASR

signature.asc
Description: Digital signature