[Pixlib] notes online, re. plugin/module structure

[Stefan Vandermeulen]

Btw, have you seen that Francis released a pixIoC teaser? Exciting uh!?

i saw it :) i can't wait to try it. ( and lowra )
<http://www.netdust.be>
Netdust[i]std.
Stefan Vandermeulen



Marcelo de Moraes Serpa wrote:
> Ah... I see :)
>
> Now, we have a PluginHelper for the contract and plugin API, and we 
> can still use the "old" technique of accessing views through 
> MovieClipHelper (by passing the Plugin name, from PluginList). This 
> view would be the main view of the plugin. We could nest other UIs 
> inside it, of course.
>
> However, I think that we really shoudn't access a plugin view from 
> anywhere nor access specific plugins (using the getPluginHelper) 
> UNLESS we consider that other plugins "require" the presence of 
> another plugin to function correctly (this implies coupling). I think 
> that the best way is to update the core models and listen to the core 
> models, right?
>
> Thanks for this great example and for helping! :)
>
> Btw, have you seen that Francis released a pixIoC teaser? Exciting uh!?
>
> Marcelo.
>
> On 1/5/07, * Stefan Vandermeulen* <stefan at netdust.be 
> <mailto:stefan at netdust.be>> wrote:
>
>     Hi Marcelo
>
>     I uploaded a cleaner version now without moviecliphelper, because
>     it wasn't implemented in the right way. sorry for overwriting them
>     each time
>
>     what i did in the previous one was
>     1) load in the plugin swf
>     2) use the newly available classes to make an instance of the plugin
>     3) the plugin would register itself with a PluginHelper that
>     extends a MovieClipHelper, and the necessary models
>     then i thought the circle was round :) mmm...
>
>     the problem here was that the pluginHelper contains core methods
>     so the plugin has the right functionality to work. in order to use
>     the MovieClipHelper in a correct way, the main plugin class should
>     also be UIclass like in the normal pixlib way.
>     this means that the core methods ( more a model type of thing )
>     and the ui methods are all packed into the moviecliphelper. it
>     will work, but it's not good practice i think.
>     I would prefer a separation between core functionality and ui
>     functionality.
>
>     so... the new version i upload should do ( i doesn't yet, because
>     i don't  know if it's ok )
>     1) load in the plugin swf
>     2) use the newly available classes to make an instance of the plugin
>     3) the plugin would register itself with a PluginHelper , and the
>     necessary models
>     4) the plugin initializes ( now it stops here, all functionality
>     is there but no view )
>     5) upon initialization a UIclass that extends moviecliphelper is
>     created and this uiclass uses the loaded swf as it's view.
>
>     // plugin inititialize
>     function inititialize()
>     {
>         var pluginUI = new PluginUI( PluginList.SOME_NAME, mc );
>     }
>
>     public static function main( timeline:MovieClip )
>     {
>         mc = timeline
>     }
>
>     the mc can be accessible as a argument for the static main method
>     like it is now
>
>     again round circle, you have a compiled plugin and a working view
>
>     this also means that:
>      if you want to access plugin methods, you use the PluginHelper or
>     the coreAPI
>      if(for some reason ) the core needs to access the plugin view, it
>     should use the moviecliphelper to get it...
>
>
>     i think this is a nice way, but it would mean a specific way of
>     arranging the plugin class
>
>     did i explain it better?
>
>     stefan
>     <http://www.netdust.be>
>     Netdust[i]std.
>     Stefan Vandermeulen
>
>
>
>
>     _______________________________________________
>     Pixlib mailing list
>     Pixlib at osflash.org <mailto:Pixlib at osflash.org>
>     http://osflash.org/mailman/listinfo/pixlib_osflash.org
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://osflash.org/pipermail/pixlib_osflash.org/attachments/20070106/38de44c8/attachment-0001.htm