On Nov 10, 2007 7:10 AM, Randy Pearson <[EMAIL PROTECTED]> wrote:
> In our case, we have two
> substantially separate applications but with the same user base and security
> plan, such that they will share common middleware. Obviously we could
> instead just wrap each application with the same middleware combination--and
> we may do this--but there appear to be benefits to combining them in a
> cascade-like scenario. Thus our interest.
What benefits? If the apps are substantially separate and you have
the freedom to mount them on different URL components ("directories"),
what benefit is there to cascading? Are there a small number of
actions or static files they'd share?
> Agree there. One of my colleagues would say this overloading creates
> cognitive friction, by using one thing to mean something else. Maybe a
> pseudo HTTP response (599 Not My Job), or even forget the HTTP part and
> raise a custom NotMyJobException that the cascade manager catches?
The exception would be more Pythonic if status 404 doesn't meet your needs.
> >> Mike: But if you start thinking about "Python apps inside a Python app",
> >> that's something that's just starting to be discussed. The thread
> >> "Sites within sites" two weeks ago has more info. Basically we've
> >> come up with a couple strategies that might work but nobody has tested
> >> them yet.
>
> This is interesting and, oddly, it feels good that the discussion is just
> getting started. Apps within apps, at least in WSGI space, could be
> workable. Maybe a framework could provide general support for an app
> containing sub-apps. When an app gets called, it gives its children (if any)
> each a shot (like Cascade) and if none of them act, the parent takes its
> shot. Just musing.
In my case i've decided to go with a regular Pylons application,
because although a few components could be shared, the majority of
components are too application-specific. Models and templates can go
in component-specific subdirectories. Controllers can't as far as I
can tell, so if a component controller gets too big, I'll have to
think about how to split it. However, I also believe in doing large
things like 'search' in the lib directory, with only the UI-specific
parts and glue in the controller. That'll help keep the controllers
on the small side. However, one component has three levels of
add/modify/delete plus several search/browse modes, so I don't think I
can get by without at least 20 actions.
--
Mike Orr <[EMAIL PROTECTED]>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"pylons-discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/pylons-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---
