Simon Laws wrote:
Following on from Jean-Sebastien's SPI point. There are still open
issues,
e.g. [1] related to SPIs that we choose to put where. I'm only now
getting
to grips with what has been done in the reorganization effort
(apologies for
not keeping up) and am spotting some things that, on the face of it,
don't
seem obvious, e.g. I was looking at one of the extension processors
and it
extends BaseArtifactProcessor which is in o.a.t.assembly.xml and not
obviously identified as something I should extend.
That's because BaseArtifactProcessor is not an SPI. It's a base
implementation class, which should not be used or extended directly by
Tuscany extensions. I think that the few processors that use it should
be fixed to not extend BaseArtifactProcessor anymore.
Now I fully expect that
when I write an extension I'll just copy one that already exists and not
think any more about it. However what we are hoping to do next is have
people try this SPI and implement new extensions. In my experience this
means they they will find problems and we will have to consider
changing it.
So my question now is (you knew I would get there eventually) how
stable is
stable?
It's too early to speculate. Let's see if people run into any problems,
address any concrete problems then and try to do it without breaking our
SPIs.
Simon
[1] http://issues.apache.org/jira/browse/TUSCANY-1281
TUSCANY-1281 issue can be fixed with changes in the implementation
packages, without changing the SPIs themselves.
--
Jean-Sebastien
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
