[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Source Control for Database Aspect

From: Dan Shookowsky <dshookowsky_at_gmail.com>
Date: 2005-07-21 17:22:16 CEST

I see. I think the infrequent commits are the difference between your
environment and my experience. I like to commit things fairly often
so that I can return to a prior state in the event that I completely
bork some code while being "creative" or "efficient". These commits
are not stable at all and there's little reason to tag or include a
database script. You could probably use NANT or something similar to
perform a build/dbscript/commit operation in one fell swoop.

On 7/21/05, Albie Janse van Rensburg <albie.jvr@gmail.com> wrote:
> We're not really looking to subversion as a database recovery tool - I'm
> talking about, say, the "programming" aspect of the database - stored
> procs and "settings" such as metadata - status ids and that sort of
> thing, which are not user editable anyway.
>
> Yes, the scripts do have enforced compiling, but the application code
> might be dependent on a specific set of parameters or expect specific
> outcomes, etc which can change. The new version of code adapts quickly
> and goes into source control, but whenever someone tries to have a look
> at the old source (say 2 versions back, not yet a "stable" release
> gone), they run into problems. I think the need for control in this
> regard is well understood, just thought I'd put it down into writing ;-).
>
> We do very slow commits - say once every two days or so per developer,
> normally once that developer has reached a pretty stable point in his
> work, but scripting everything out before a commit still begs to be
> automated/integrated.
>
> Maybe one day (optimistically, at least 6 months hence) when I have time
> I'll look into building something myself...
>
> I'm checking out that script, thanks.
>
> Thanks for all the help so far.
>
> Dan Shookowsky wrote:
>
> >This guy has a VBScript that automates the scripting and versioning of
> >databases with VSS. It could probably be easily adapted.
> >
> >http://searchsqlserver.techtarget.com/tip/1,289483,sid87_gci1093178,00.html?bucket=ETA
> >
> >Personally, I'd rather only script a known "good" version and rely on
> >database backups for catastrophes involving the database. This is
> >because SQL Scripts, unlike code can't be on the server unless they
> >actually compile.
> >
> >On 7/21/05, Albie Janse van Rensburg <albie.jvr@gmail.com> wrote:
> >
> >
> >>Thanks for the help. Is this the only way? I would obviously prefer a
> >>more integrated approach, but we'll have to make do. Does anybody know
> >>anything about the source control support in SQL server?
> >>
> >>Dan Shookowsky wrote:
> >>
> >>
> >>
> >>>Before you create a tag indicating a stable version that you'd like to
> >>>get back to, script the database into one or more .SQL files and
> >>>commit these to the repository. You end up with all the info that you
> >>>need. If you have specific meta-data that must be in the database,
> >>>you'll probably need to script this in some other way.
> >>>
> >>>On 21 Jul 2005 11:46:50 -0000, users-digest-help@subversion.tigris.org
> >>><users-digest-help@subversion.tigris.org> wrote:
> >>>
> >>>
> >>>
> >>>
> >>>>From: Albie Janse van Rensburg <albie.jvr@gmail.com>
> >>>>To: users@subversion.tigris.org
> >>>>Date: Thu, 21 Jul 2005 11:51:11 +0200
> >>>>Subject: Source Control for Database Aspect
> >>>>Hi all
> >>>>
> >>>>I am implementing Subversion for a .Net project at work, and it's great
> >>>>to be able to use plug-ins like AnkhSVN to get Subversion integrated
> >>>>with Visual Studio. However, like most software project above the
> >>>>"Hello World" level, there is a database aspect to our development. We
> >>>>are using Microsoft SQL Server 2000, and a large part of the project's
> >>>>business logic lies in stored procedures and metadata that populate
> >>>>elements in the project. It is necessary to bring the database aspect
> >>>>into version control, as older versions of the .Net source are often
> >>>>incompatible with different versions of the database.
> >>>>
> >>>>Is there any way to accomplish this level of source control using
> >>>>Subversion? I am also very interested to know whether there are any
> >>>>Subversion plugins available that would work with SQL Server - it has
> >>>>some method by which Visual Source Safe (ew...) can integrate, so it
> >>>>seems natural to assume that third party developers should be able to do
> >>>>something similar.
> >>>>
> >>>>Any hints, suggestions or flames (I did mention VSS. Argh! Again!) are
> >>>>welcome.
> >>>>
> >>>>Albie
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>>
> >>>
> >>>
> >
> >
> >
> >
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Jul 21 17:24:32 2005

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.