thoughts on repository/storage/branch/checkout

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Denys Duchier wrote:
> Aaron Bentley <aaron.bentley at utoronto.ca> writes:

> With inheritance, it would just work.  Also there is the problem that the same
> repo may be mounted in different places on different machines, in which case an
> absolute URL would not work.

It does not seem unreasonable to expect a file server's mount point to
be the same on all software development machines.  Certainly, there are
many ways to make it so if it is not.

Inheritance would also mean that you could confuse matters terribly if
you wound up in a situation with nested repositories.

> yes.  I just checked the IRC log and _you_ said to me:
> 
> <abentley> Yes, as I said, the branch uses the containing repository for storage.
> 
> I think that's actually what got me started thinking of inheritance.

A more precise phasing would have been 'when the branch is created, the
location of the enclosing repository is set as the branch's storage
location'.  Sorry if that was confusing.

>>>I am not adding a concept, I am merely making it possible to freely and
>>>uniformely combine 4 primitive concepts, each one very circumscribed and thus
>>>hopefully easy to grasp.
>>
>>As I mentioned earlier, the division is a matter of taste.  Your
>>concepts of STORE, BRANCH, and probably CHECKOUT are also divisible.
> 
> 
> It would be nice to judge my proposal on its own merit rather than dismiss it
> because other divisions are possible.

That was in response to your assertion that you were not adding a
concept.  I still hold that since my model has 3 concepts and yours has
4, that you have added a concept.

I have already, separately, made repeated value judgements about this
division, but once more: I do not believe that the additional
functionality it provides justifies the additional complexity it
entails.  I do not think it is a useful division to make.  I think it
lacks merit.  None of this is news to you, right?

> Also my primitive concepts do no not suffer from multiple
> personalities disorder (the way the word "branch" currently does).

Please, enough with the slams.  The use of "branch" to describe the
result of "bzr init" has been with us since the beginning of the
project.  The terminology I've been proposing would have particular term
for it: a "Standalone Branch".

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFD4j1j0F+nu1YWqI0RAqloAJ9JbcrWyhaefxhnUT6OUTayoJNbmQCeOBNG
XfP2vPNYH8BFtPcZCWLbnV4=
=Wz/s
-----END PGP SIGNATURE-----