On Thu, Dec 15, 2011 at 3:36 PM, David Vaillancourt
<david_v@sympatico.ca> wrote:
> Dave,
>
> Thanks for the quick feedback. I'm really surprised that people who use VCS
> for their 'normal' code, wouldn't see a use for the same when it comes to DB
> Schema/DDL versions.
> In any case, my goal is not to add the functionality directly to PGAdmin,
> but rather create a plugin for PGadmin to use.
Yeah, the current plugin architecture really isn't suited to that I
doubt - its pretty simplistic. I certainly have objections to that
being enhanced though.
> As I mentionned earlier, the extent of the modifications to PGadmin would be
> to extend the plugin framework to allow a tighter coupling between plugins
> and PGadmin.
> I guess maintaining a plugin is different from maintaining yet another
> feature in the existing PGAdmin code, correct me if i'm wrong.
Currently a "plugin" is really just an external application that
pgAdmin can fire up and pass command line options to.
> Finally, since no one wanted to use the VCS feature in PGadmin, what do
> teams of developers use to track versions of Schemas (Tables, Views,
> Procedures...)?
> Do most developers simply not use any? Or do the 'hack' their own? I'd be
> surprised if they used proprietary tools ...
Well, views may have changed now - the original code was probably
written nearly 10 years ago. However, the PostgreSQL infrastructure
team just tend to keep SQL scripts in GIT and update them when
necessary. For the Postgres Enterprise Manager product that I work on
in my day job (which happens to be based on pgAdmin), we maintain one
script which creates a fresh database from scratch (pemserver.sql),
and incremental scripts to upgrade an existing database from one
version to another (eg. 2_0_0-2_0_1.sql, 2_0_1-2_0_2.sql and so one).
All of those are version controlled in GIT as well.
--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake
EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company