[prev] [thread] [next] [lurker] [Date index for 2002/09/07]
On Sat, Sep 07, 2002 at 02:44:05PM +0100, Greg McCarroll wrote:
>
> So how are we doing, last time i looked the anchor of storage had ground
> things to a stop, have we burst free of its shackles like an wild eagle now?
>
> Are we moving Simon's stuff to use more methods in Storage::*?
I think it should all be moved over now.
> and maybe have it default to one or the other. i think making it run
> the tests for every data store would be a little bit test fetishistic
As developers (developers, developers, developers), I want make test
to be able to test against all the storage backends, but then I am a
test fanboy.
The way I see it working is at Makefile.PL time, so you can do
perl Makefile.PL STORAGE=DBI PATH=/opt/siesta
make test
PATH then configures where the templates and cgis get installed to.
make test only tests the DBI storage. This can also be used to
specify a different set of dependencies. As a hacker I can do:
for backend in SQLite DBFile Volatile ; do
perl Makefile.PL STORAGE=$backend
make test || exit 0
done
The arguments are then used to generate a Siesta/Config.pm from
Siesta/Config.pm.in populating such things as default storage class,
and where to look for templates post-install.
Running Makefile.PL with no arguments would try to use Siesta::Config,
and either upgrade based on that, or {supply defaults, screetch to a
halt}. I'm not sure whether I want defaults or death, I can see the
benefit of both.
> p.s. are you guys going to dave's birthday bash tonight?
Aye (but then you know that already).
--
Richard Clamp <richardc@xxxxxxxxx.xxx>
Generated at 13:56 on 01 Jul 2004 by mariachi 0.52