Conceptual question about several users using the same "local repository"

[Carlos Henrique Da Costa Cano <hc...@pucrs.br>]

Hello All,

 

In my company we have a discussion about the usaga of Subversion.

After I set the sever (Linux + Apache + auth_ldap(active directory).

The developers started using it...

After some time I checked and the developers was using it in some weird
way for me.

 

The developers make a checkout in a windows network drive.

And the others developers use this networkdrive and in the end
"somebody" (or maybe the last one in the day) do a commit.

 

It strike me, maybe I am wrong but I don't feel like this is the best
way.

Today this is working but I have the feeling that we are missing
something or in the future something bad will happen.

 

By my understanding the developers must do a checkout in a each local
drives, work on it and after make the commit.

 

Who is right? What are the problems of this "network" several users
approach?

 

Thanks

 

Carlos Henrique da Costa Cano

 

Re: Conceptual question about several users using the same "local repository"

[Garance A Drosihn <dr...@rpi.edu>]

At 10:08 AM -0300 10/10/08, Carlos Henrique Da Costa Cano wrote:
>
>The developers make a checkout in a windows network drive.
>And the others developers use this network drive and in the
>end "somebody" (or maybe the last one in the day) do a commit.
>
>It strikes me, maybe I am wrong but I don't feel like this is
>the best way.  Today this is working but I have the feeling
>that we are missing something or in the future something bad
>will happen.

It isn't the best way, especially if multiple developers are
actively changing files in the same checkout-area.  There might be
some reason they feel like they need to do things this way (such
as:  they are all working on a set of tightly-related changes,
and this is the only sensible way to test everyone's changes at
the same time).  But in general, this sounds like a bad idea to
me.  If they don't have some reason which forces them to work in
that manner, then they should split up into separate work-areas.

-- 
Garance Alistair Drosehn            =   gad@gilead.netel.rpi.edu
Senior Systems Programmer           or  gad@freebsd.org
Rensselaer Polytechnic Institute    or  drosih@rpi.edu

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

Re: Conceptual question about several users using the same "local repository"

[Andy Levy <an...@gmail.com>]

On Fri, Oct 10, 2008 at 09:08, Carlos Henrique Da Costa Cano
<hc...@pucrs.br> wrote:
>
> In my company we have a discussion about the usaga of Subversion.
>
> After I set the sever (Linux + Apache + auth_ldap(active directory).
>
> The developers started using it…
>
> After some time I checked and the developers was using it in some weird way
> for me.
>
>
>
> The developers make a checkout in a windows network drive.
>
> And the others developers use this networkdrive and in the end "somebody"
> (or maybe the last one in the day) do a commit.
>
> It strike me, maybe I am wrong but I don't feel like this is the best way.

Nope, you're completely right. Here are two examples of what's wrong
with this usage:

How do you prevent 2 users from overwriting each others' changes?
How do you know who was responsible for each change that was committed?

In your current usage, those two questions cannot be answered
definitively. If you're subject to any kind of auditing of your source
code (ISO compliance, Sarbanes Oxley) this *will* send up a HUGE red
flag for the auditors.

> Today this is working but I have the feeling that we are missing something
> or in the future something bad will happen.

Well, you're not guaranteed that "something bad" will happen but that
depends upon your definition of "bad".

You will be losing or negating several of the benefits of/reasons for
using Subversion or any other VCS. And losing portions of your audit
trail/change history

> By my understanding the developers must do a checkout in a each local
> drives, work on it and after make the commit.

That is a much more appropriate usage model, yes.

> Who is right? What are the problems of this "network" several users
> approach?

You are right, IMO. For the problems, see above.

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

Re: Conceptual question about several users using the same "local repository"

[Toby Thain <to...@telegraphics.com.au>]

On 10-Oct-08, at 9:08 AM, Carlos Henrique Da Costa Cano wrote:

> Hello All,
>
> In my company we have a discussion about the usaga of Subversion.
> After I set the sever (Linux + Apache + auth_ldap(active directory).
> The developers started using it…
> After some time I checked and the developers was using it in some  
> weird way for me.
>
> The developers make a checkout in a windows network drive.
> And the others developers use this networkdrive and in the end  
> “somebody” (or maybe the last one in the day) do a commit.
>
> It strike me, maybe I am wrong but I don’t feel like this is the  
> best way.
> Today this is working but I have the feeling that we are missing  
> something or in the future something bad will happen.

This is completely the wrong way to use Subversion. Evidently they  
don't understand how it works.

>
> By my understanding the developers must do a checkout in a each  
> local drives, work on it and after make the commit.

Yes.

--Toby

>
> Who is right? What are the problems of this “network” several users  
> approach?
>
> Thanks
>
> Carlos Henrique da Costa Cano
>