[mythtv] More ideas on PostgreSQL support, comments welcome
Ronald van Eijck
ronald at vaneijck.com
Tue Aug 31 00:21:12 EDT 2004
> On Monday 30 August 2004 04:59 pm, David Härdeman wrote:
> > Fine, so don't update the psql file, just update the mysql one and I'm
> > sure that someone that cares for postgresql can update the corresponding
> > pgsql soon thereafter. Worst case scenario is that PostgreSQL support is
> > broken for a period of time. Would this be an acceptable compromise?
> No, I wouldn't really consider that's acceptable at all. What happens
> want to make a release and the postgresql support isn't updated?
To support multiple databases it is better to have some sort of common
definition for tables and queries and a 'driver' for the database that
translates this to the specific sql dialect.
I do al lot of java projects and we use the hibernate framework in
some projects. It is a generic object-relational wrapper and supports
lots of different databases. There must be something like that for
QT / C++ out there. If not the first step should be a set of support
methods to perform queries, updates & deletes in an sql independent
> > >I will not add support for other databases if it makes the current
> > > internal database usage any harder to use. I see absolutely no
> > > in doing that.
> > I fail to see how my suggestion would make internal database usage
> > harder? Is updating a separate file with SQL queries harder than
> > a .cpp file with SQL queries?
> It's not harder, but it's a large step backwards, IMO. It's one more
> keep track of, and one more thing that can possibly break.
> If multiple databases can't be supported without extra work for me down
> road, then there won't be support for multiple databases. I don't see the
> benefit in it.
There will always be extra work but this should NOT be in the form of
writing every sql statement multiple times.
More information about the mythtv-dev