Re: [Pyst-users] functions for database access

[<jpm...@gm...>]

Matt,

thanks for your feedback. Like I've said, just quick and dirty. At least
to my knowledge (searching the web for a day), so far, there are close
to zero examples for the usage of pyst, so I meant to contribute a
"how-to find back an event associated to an action" (identified by an
action_id), and my next step would indeed have been to implement my own
Action class with, e.g., a method events().

The very same structure, as far as I can deduce from a glimpse on astxx,
is one of the main improvements in the class hierarchy (taken aside the
more robust reduction of threads), so this looks quite promising, and I
guess I prefer to live with my self-knitted solution as is, until you do
better. Anyway, I would like to contribute to the testing, if you are
interested and I find the time.

I share your opinion about the importance of a versatile and robust
low-level functionality, nevertheless I think there should be a medium
level module, too, which reflects the spectrum of Manager actions,
including also database functions. For example, I don't get what makes
mailbox_count or setvar, say, (and even originate) more low-level than
database access?

Matthew Nicholson wrote:
> Jan,
>
> I don't think using self.command directly is the correct way to 
> implement this feature.  Also I don't think automatically handling 
> events for the user is the right way to do it either.
>
> I think pyst should be fairly low level and only provide an convent 
> interface to AGI and the Manager.  Also on the flip side, pyst should 
> not be designed to prevent the user from building a layer of abstraction 
> on top of it either.
>
> Either way, pyst 0.3 will make it easier to do these things in a more 
> straightforward manner.  No ETA on that though, for an idea of what it 
> might look like, check out astxx (http://matt-land.com/astxx).
>
>