On Thu, 06 Apr 2006 10:21:59 +0300, David Smith <
XXXX@XXXXX.COM>
writes:
Quote
>Yes they do. Your local copies are just local copies of **some** stuff
>that is in the repository.
No, they are local copies of some stuff that **was** in the repository
some point in time. That might not be the case any more. So one could
argue that the local copy is *my* master as I do coding and diffs
against it. Only when I see necessary, I can refresh the *local* copy
with possible changes from others and do the merging to my *local* copy.
I'm just using Subversion for my self, as a kind of "backup tool with
history", and an easy way to sync between main box and laptop.
So far I kinda get the impression that in contrast to most other VCS,
where (from my limited experience) everything revolves around the
repository, Subversion's repository is a bit like scientific journals
ala Nature.
You have a bunch of scientists doing research at different locations.
In order to let the other scientists know what results they've found,
they get an article published in a journal. Other scientists can then,
if they care, read about these results. Or they can go to the archive
and check out results published in earlier issues. Later they
themselfs publish something. If you have two different results from
groups probing the same issue, discussion would occur as to what's
really going on etc.
Subversions repository holds a history of all the previous code
written. It also holds the latest commited code, which you are free to
check out if you want to, regardless of what others do (no locking).
You can also commit code whenever you want to, and if this code
conflicts with what's already in the repository, you need to "settle
the argument", ie merge.
So far, Subversion works ok for me. But I am just myself, so I can
treat the repository in a very different way than if I was working in
a team.
- Asbjørn