The reactorFactory does a pretty good job of caching the
objects that need to internally. By letting it do its job you dont have to
worry about when objects need to be updated into your application scope.
When you ask for a gateway from the factory, the first time it will create it if
it needs to. Each request for it afterwards with get you a cached copy.
In most of my fb apps, I will just shortcut my calls by placing set calls in the xml of my model circuit. In general I consider this bad practice because (without lexicons and such) you cant catch errors and you have limited choices in what you can do from this point. But I often find that making the calls from the xml is "good enough" and that its very simple to move this into an act_ fuse at a later point if I need more control.
In most of my fb apps, I will just shortcut my calls by placing set calls in the xml of my model circuit. In general I consider this bad practice because (without lexicons and such) you cant catch errors and you have limited choices in what you can do from this point. But I often find that making the calls from the xml is "good enough" and that its very simple to move this into an act_ fuse at a later point if I need more control.
--------------------------
So Simeon, if I understand what you are saying, then i
*ONLY* need the ReactorFactory in application scope, and it takes care of
additional efficiencies like storing other things in application scope as
needed?
And, if I understand most of what else has been discussed
here, then my model circuit will be one (or several) xml files for the model,
all of which point to actX files, all of which will contain simply calls to
reactor?
What's troubling me at this point is the number of files I
have to bounce around to, just to write one fuseaction...
1. Create XML block in controller
file
2. Create XML block in model file, create one or more
actX files, write code there
3. Create XML block in view file, create one or more
dspX files, write code there
This amounts to having at least 5 files to work with at all
times... three of which have the same name, but are in different
directories. I've started getting used to this, but it still just feels
like I'm bouncing around a LOT. Has anyone else felt this way, or am I
just a sissy?
Perhaps the fact that Simeon has felt the urge to 'break
from best practice' by putting reactor calls right in the model XML file because
it's "good enough" and really is faster - something I can
identify with - is indicative of a potential problem with having to work with so
many files at once... which for someone new to fusebox could be a source of
confusion and frustration...?
I know, I know, this is the nature of MVC... we're
assembling encapsulated "modules" here... but it just feels a little heavy
writing code nowadays... maybe I'll get used to it.
m
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Reactor for ColdFusion Mailing List
[email protected]
Archives at: http://www.mail-archive.com/reactor%40doughughes.net/
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
