Re: pl./sql deployment stragegies?

From: Mark C. Stock <mcstockX_at_Xenquery>
Date: Thu, 30 Sep 2004 09:47:24 -0400
Message-ID: <CcGdnZdUkfQPk8HcRVn-sQ@comcast.com>

"Galen Boyer" <galenboyer_at_hotpop.com> wrote in message news:uoejosqmk.fsf_at_standardandpoors.com...
| On Wed, 29 Sep 2004, mcstockX_at_Xenquery .com wrote:
| >
| > "Terry Dykstra" <tddykstra_at_forestoil.ca> wrote in message
| > news:OKC6d.3197$Du2.608_at_edtnps89...
| >| PLSQL Developer also supports Source Control.
| >|
| >| --
| >| Terry Dykstra
| >| Canadian Forest Oil Ltd.
| >|
| >|
| >| "Sybrand Bakker" <gooiditweg_at_sybrandb.verwijderdit.demon.nl>
| >| wrote in message
| >| news:7lsll05ad0mj47to6k8m3egicok1dvd54c_at_4ax.com...
| >| > On Wed, 29 Sep 2004 09:25:02 -0400, "Mark C. Stock"
| >| > <mcstockX_at_Xenquery .com> wrote:
| >| >
| >| > >On a multi-developer project, what's your standard/typical
| >| > >way of
| >| handling
| >| > >the deployment of PL/SQL packages from developer schemas to
| >| > >a
| > centralized
| >| > >schema (version control schema), test schemas, and finally
| >| > >production schema(s)?
| >| > >
| >| > >Since PL/SQL code dependencies are determined at compile
| >| > >time, this
| > poses
| >| > >some interesting challenges.
| >| > >
| >| > >++ mcs
| >| > >
| >| >
| >| >
| >| > Toad supports Sourcesafe
| >| >
| >| >
| >| > --
| >| > Sybrand Bakker, Senior Oracle DBA
| >|
| >|
| >
| > my key interest is deploying from/to and synchronizing master,
| > developer, test, and production schemas
| >
| > i'm looking for feedback on how many developer schemas (one per
| > developer or shared?), how do you keep schemas in sync (objects
| > and tables), how do you support unit testing and integration
| > testing

|

| We have an instance per developer, and at any point, the database
| code can be checked out and created by typing a command.

|

| This strategy is used to deploy to the more controlled
| environments, but this time from particular tagged versions of
| the sourcebase.

|


| > ie, developer is working on a revision for package022 which

| > depends on 5 other packages, and 12 tables, and is called via

| > package037 -- typically, are people duplicating the entire

| > application schema for each developer, sharing one development
| > schema, or using some other means?

|

| I also like the suite of schemas per developer. The main point
| is isolation for accuracy.

|

| --
| Galen Boyer

thanks galen, et al...

multiple instances, each with a suite of schemas sounds ideal -- anybody got any feedback on how they handle multiple developers on a PL/SQL project where it is not possible (politically or otherwise) for each developer to have their own instance?

also, in playing around with TOAD built-in Team Coding, it appears to just work via TOAD (it did not prevent me from overwriting the test function via SQL*Plus, and when i checked in the function in TOAD, the SQL*Plus version persisted) -- is SQL Navigator's implementation essential the same?

are any of you aware of any VCS products that manage check-in check-out at the database level so that once checked out, nobody can overwrite the object unless connected and verified as the developer that checked it out? i.e., true clobber-protection enforced at the database level, not just at the tool level?

++ mcs Received on Thu Sep 30 2004 - 08:47:24 CDT