info-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: CVS-SQLServer Query


From: Matthew Herrmann
Subject: RE: CVS-SQLServer Query
Date: Thu, 23 Aug 2001 16:06:15 +1000

Hi Tim,

I just checked out my copy of SQL Server (ie looked at i mean) and you can
generate scripts as one object per file. If you do this into a cvs working
directory, then you can commit that by looking at it as files.

so you would have on the dev server:
a) enterprise manager and the mdf holding the work in progress
b) the working directory holding file snapshots of the stored procedures

c:\scripts\
...\updateCustomer.spo
...\getCustomer.spo
(etc.)

As for deployment, you can say:
- we need to deploy changes to stored procs x, y & z -- you find this out by
doing a log or diff of sorts

- you then run those *.SPO files on the server and it should drop existing
and create new stored procedures.

the nice thing about this is that you have your valuable scripts in a text
format and not locked away in a mdf file and your developers use the
development server as a sort of sandbox, editing in enterprise manager with
a sort of 'commit' back to the sandbox. Of course, new stored procedures
create new files (so a cvs update will give ?'s that you can pick up and add
them to the repository)...

i think that would be a reasonably clean solution -- and you could give
change management something saying "we are 100% certain that with this
change we will only change _these_ lines of code".

If you could automate the generate script function in some way, that would
be the ideal.

There might be some minor issues you need to sort out like handling merges
back from old versions --> how do we then resolve conflicts between
enterprise manager and your sandbox. You'd probably need to make two
sandboxes and ... mmm... well like i said that's something that would need
to be worked out.

this might lead nicely into some sort of deployment management system --
almost like an automated build system but for SQL (...?) but that's getting
carried away.

Hope this is helpful,

Regards,

Matthew Herrmann
Far Edge Technology




reply via email to

[Prev in Thread] Current Thread [Next in Thread]