Re: plugin loading order

2009-04-06 Thread Musachy Barroso
btw, this provides some very nice reloading of configuration by just re-deploying the bundle. Admittedly , it was not intentional, but it works, so I am kinda surprised. musachy On Sun, Apr 5, 2009 at 8:51 PM, Musachy Barroso wrote: > The changes are in the sanbox and trunk, to make Convention w

Re: plugin loading order

2009-04-05 Thread Musachy Barroso
The changes are in the sanbox and trunk, to make Convention work with OSGi bundles. I still need to implement the plugin load order thing, right now it works because the OSGi plugin gets loaded first. I think we should get the OSGi and OVal plugins into trunk for the next release and label them "e

Re: plugin loading order

2009-04-05 Thread Musachy Barroso
and it finally works :) musachy On Sun, Apr 5, 2009 at 5:00 PM, Musachy Barroso wrote: > The plugin loading order takes 15 mins to implement. Getting OSGi to > work with Convention, 3 days and counting, I am losing my nerve > already :) > > musachy > > On Sun, Apr 5, 2009 at 4:54 PM, Wes Wannema

Re: plugin loading order

2009-04-05 Thread Musachy Barroso
The plugin loading order takes 15 mins to implement. Getting OSGi to work with Convention, 3 days and counting, I am losing my nerve already :) musachy On Sun, Apr 5, 2009 at 4:54 PM, Wes Wannemacher wrote: > On Sunday 05 April 2009 15:17:24 Chris Pratt wrote: >> Sorry, I tried to make it clear

Re: plugin loading order

2009-04-05 Thread Wes Wannemacher
On Sunday 05 April 2009 15:17:24 Chris Pratt wrote: > Sorry, I tried to make it clear that the optional attribute would be > available to intrinsically denote load order without imposing any > all-or-nothing dependency. Basically the point is to tell struts that the > requirement is "if you want C

Re: plugin loading order

2009-04-05 Thread Chris Pratt
Sorry, I tried to make it clear that the optional attribute would be available to intrinsically denote load order without imposing any all-or-nothing dependency. Basically the point is to tell struts that the requirement is "if you want Convention to be able to use OSGi, check to see if it's avail

Re: plugin loading order

2009-04-05 Thread Wes Wannemacher
The other part of that is that Conventions and OSGi do not "depend" on each other. This load order is necessary to make the two play nice, but I'd hate to be forced to use OSGi when using Conventions or vice versa because of a dependency that is unnecessary. For cases like this, there is either

Re: plugin loading order

2009-04-05 Thread Musachy Barroso
yeah I thought about that, but I would rather keep it simple. So far this is the only case where I have needed loading order, and I don't think it will be common at all. musachy On Sun, Apr 5, 2009 at 2:43 AM, Chris Pratt wrote: > Wouldn't it be better to describe the interdependencies of the pl

Re: plugin loading order

2009-04-04 Thread Chris Pratt
Wouldn't it be better to describe the interdependencies of the plug-ins and let struts build a directed graph to automatically load them in the proper order. Otherwise, someone is going to put plugin1 at order=5 and plugin2 at order=6 then plugin3 will come along and need to be in between. There

Re: plugin loading order

2009-04-04 Thread Wes Wannemacher
On Saturday 04 April 2009 21:50:22 Musachy Barroso wrote: > I have been trying a few different ways to get the Convention plugin > to work with the OSGi plugin, and there are some hacks that can be > done but I don't like them. I want to implement (I actually have it > implemented it already somewh