Subversion user with non network questions

[Timothy Weaver <tw...@eclypse.org>]

I am a software engineer for a specialized programming language called ELITE
that controls an automated wire tester.  We as a company currently use 2
version control programs.

 

We use Code Co-op for our programs written in ELITE and Subversion for our
support programs that are written in Python and C++.

 

The reason we use both is currently we can't find a way to do a check in
when you don't have internet connection.  Much of our time is spent at the
customer's site where we may not have internet access and we have to do
debug of the program to meet customer specs.  This debug has lasted for as
much as 3 weeks without internet access.  3 weeks without ever doing a check
in can be scary since anything can happen in that much time period.  

 

Code Co-Op allows non-networked check-ins with an email server that will
sync up when the whole project is done.  I am not a huge fan of the email
server sync script method, but am happy that I have a method of version
control on site.  

 

My question is this:  Is there any planning to be able to do local check-ins
when internet connection is not found and then be able to sync up with one
master update when network connection does exist?

 

I have found SVK from Kwiki, but have been unable to get that to work for
me.  I may be doing it wrong or it may be not what I am looking for.  I do
know that I will have to set it up to work through windows as some of the
people that are programmers for ELITE know nothing of any other programming
language and don't know how to use DOS or LINUX.

 

Just as a note, our entire network is windows based as well as our
programming host language.

 

Timothy M. O. Weaver

Eclypse International

Application Engineer

Email - tweaver@eclypse.org

Phone - 951-371-8008 (x34)

Cell - 951-300-8405

 

Re: Subversion user with non network questions

[Stephane Bortzmeyer <bo...@nic.fr>]

On Mon, Aug 29, 2005 at 03:10:30PM -0700,
 Timothy Weaver <tw...@eclypse.org> wrote 
 a message of 258 lines which said:

> The reason we use both is currently we can't find a way to do a
> check in when you don't have internet connection.

This is a very common requirment and a motivation for decentralized
VCS.

You basically have two solutions:

1) Stay in the Subversion world and use SVK
(http://svk.elixus.org/). Good things: you keep your Subversion
environment and habits.

2) Switch to a real decentralized SCM like darcs
(http://www.darcs.net/) or Mercurial
(http://www.selenic.com/mercurial/). Both seem to work fine on
MS-Windows (warning: I use only Unix).

> I have found SVK from Kwiki, but have been unable to get that to
> work for me.

There is a SVK-specific list,
http://rt.openfoundry.org/Foundry/Project/Forum/List.html/wws/arc/svk-dev?Queue=82,
so I believe you should ask them.

> I may be doing it wrong or it may be not what I am looking for.

It seems a good match for your requirments (warning: I never used SVK
for real.)

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Subversion user with non network questions

["Todd A. Jacobs" <no...@codegnome.org>]

On Mon, Aug 29, 2005 at 03:10:30PM -0700, Timothy Weaver wrote:

>The reason we use both is currently we can't find a way to do a check 
>in when you don't have internet connection.

This is an inherent issue to using *any* centralized repository. 
Projects like monotone use a completely different model that allows 
private commits to a local repository with push/pull to a 
more-centralized repository on demand.

Having said that, I handle things differently. I use Subversion as a 
central repository, but if I need to do disconnected operations, I use 
RCS or CVS on specific files to track my offline history, then merge 
them back into Subversion the next time I'm online.

I'm not aware of any way to directly import the version history back and 
forth, but it's not a great deal of trouble to treat my Subversion files 
as a vendor branch in RCS, and just use the output of rlog as the input 
to a Subversion commit (e.g. svn commit -m"$(rlog foo)" foo) when 
needed.

It's not pretty, but it works. All VCSs have warts; the question is, 
which warts are the ones you can live with? :)

-- 
Re-Interpreting Historic Miracles with SED #141: %s/water/wine/g

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org