[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: I, too, miss tags.

From: Saulius Grazulis <grazulis_at_ibt.lt>
Date: 2006-02-28 07:57:49 CET

On Monday 27 February 2006 21:45, Rob van Oostrum wrote:

> The one that says tags only exist by convention as copies of a ...

This is a point -- tags are copies *by convention*.

Other people might have other conventions. Why do you want to dismiss them
(people, conventions)?

> It's a free web, it can be my religion if I want it to be! ;-)

You are free, for sure. So are others.

Can we find a consensus and a point of (mutually useful) cooperation that
suites both sides?

The proposed 'tags as copies' convention is unacceptable to those who need
labels. "Tags as labels' are unacceptable to developers since they allegedly
fear too much implementation and support burden, and for some of those, who
prefer them to labels (for no clear reason).

OK, so is there a golden middle?

Wrappers? They are bad for Win users. I personally do not use Win, but a
feature must be usable on all platforms, so that users have choice, IMHO.

Client side hooks? Can this be done? Does client support hooks? If yes, are
they sufficiently powerful to implement label functionality? I don't know.
Can anybody give information on this?

If client does not support hooks, are they planned? They can be useful for
other purpose, say for long-requested binary file diff utilities. Thus, such
feature might have broader application and meet less resistance. So maybe one
can design this mechanism sufficiently flexible so that labels can be
implemented in it.

Lets stop teaching each other "do that, don't do this", and instead
concentrate on the mentioned technical points. This will save time, nerves
and bandwidth.

> > >Using the "tags" directory method, what stops a completely unrelated
> > >source tree from being copied to the tags directory of some project?
> Vigilance, competence, and a credible threat of bodily harm ;)

I would prefer that important repository invariants are ensured by means more
reliable that social engineering...

> Seriously though, what harm would it do if a mistake like this occurred?
> It's nothing you can't undo.

You mean, 'svn obliterate' is already implemented? Without it, virtually
nothing can be undone after a commit.

Dr. Saulius Gra┼żulis
Institute of Biotechnology
Graiciuno 8
LT-02241 Vilnius
Lietuva (Lithuania)
fax:          (+370-5)-2602116
tel.: office: (+370-5)-2602556
      mobile: (+370-684)-49802, (+370-614)-36366
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue Feb 28 07:59:29 2006

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.