Message-ID: <41D4DCA5.3090204@arbash-meinel.com>
Date: Thu, 30 Dec 2004 22:59:17 -0600
From: John A Meinel <john@arbash-meinel.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
MIME-Version: 1.0
To: Zenaan Harkness <zen@CashPoolComps.com>
Subject: Re: [Gnu-arch-users] tla archive-setup/init-tree/import
References: <1104468209.8958.167.camel@localhost.localdomain>
In-Reply-To: <1104468209.8958.167.camel@localhost.localdomain>
Cc: arch <gnu-arch-users@gnu.org>
Precedence: list
Content-Type: multipart/mixed; boundary="===============1447400436=="
Sender: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Errors-To: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Status: RO
Content-Length: 3373
Lines: 101
This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--===============1447400436==
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature";
	boundary="------------enigEA1725EABF016A51292801A6"

This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--------------enigEA1725EABF016A51292801A6
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Zenaan Harkness wrote:

>Why do we have init-tree and import?
>
>As in, from a user perspective, surely these could happen transparently,
>behind the scenes, so that after archive-setup, simply do:
>
>tla get $version
>cd $version
># cp/ edit initial files into tree
>tla add $initial_files
>tla commit
>
>???
>
>So get will know to do a mkdir and archive-setup, since there's no
>base-0 revision, and commit would automatically know to do whatever
>import is supposed to do.
>
>Sorry if this has been discussed before, I've been cramming work lately
>and therefore been offlist for a while.
>
>tia
>zenaan
>  
>

I think it makes sense to have tla commit do the work of tla import, but 
I would still want a tla init-tree. To me, 'tla get' says get something 
out of the repository, and fail if it doesn't exist. 'tla init-tree' 
(could be shortened to 'tla init'?) says I want to create a new entry.

I think the reason for import/commit is that they do different things 
(commit always commits a changeset, import creates a complete tarball). 
However, since the only time you use import is on the *first* time, and 
it doesn't do anything worthwhile any other time (that I know of), I 
think it would be reasonable to have tla commit check if there is a 
base-0, and if not, do an 'import'.

The one flag I use from import is 'tla import -S' so that it will create 
the category/branch/version for me if it doesn't exist. I've actually 
gone back on my earlier stance, and think -S should be the default. Even 
if there was a typo in my init-tree, there really isn't any way to catch 
it. (Unless everyone uses tla make-category make-branch make-version, 
etc. *Does* anyone do that?)

I suppose I would be more okay with a 'tla commit --import' for the 
first commit. Just to make sure that you realize it is going to create 
new branches if it has to.

There is also the idea of making 'tla tag -S' the default, but since you 
can tag into existing trees, I think it's probably better to leave it 
the way it is.

John
=:->

--------------enigEA1725EABF016A51292801A6
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (Cygwin)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFB1NypJdeBCYSNAAMRAp4KAKCJk+TgB8mnHU8YRafAzhh4ueXTrACfUai0
/kAuX/XO8ZicX6LlPae5b6M=
=6Asj
-----END PGP SIGNATURE-----

--------------enigEA1725EABF016A51292801A6--


--===============1447400436==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Gnu-arch-users mailing list
Gnu-arch-users@gnu.org
http://lists.gnu.org/mailman/listinfo/gnu-arch-users

GNU arch home page:
http://savannah.gnu.org/projects/gnu-arch/
--===============1447400436==--