uPortal IRC Logs-2006-10-25

Owned by Apereo Infrastructure
Last updated: Oct 09, 2007Version comment
2 min read

[00:22:48 CDT(-0500)] * edalquist (n=dalquist@76.210.71.87) has joined ##uportal
[09:15:49 CDT(-0500)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[11:06:35 CDT(-0500)] * dmccallum (n=dmccallu@uni1.unicon.net) has joined ##uportal
[11:25:19 CDT(-0500)] <dmccallum> Another question for you Eric re this spring context file re-org
[11:26:40 CDT(-0500)] <dmccallum> I'm working on the cache manager portlet context file right now. It had been passing the /caches beanfactory to controllers so that the controllers could retrieve ICache instances by calling getBeanNamesForType()
[11:28:27 CDT(-0500)] <dmccallum> now, I can retain this behavior and implementation by making the controllers BeanFactoryAware, but now they're potentially searching through a very large app context for all ICache instances, and getBeanNamesForType() isn't especially performant to begin with. plus the BeanFactory dependency is something of a bad smell, IMHO. would you object to sacrificing some level of dynamic behavior in favor of injecting a collection of enumerable ICache instan
[14:19:19 CDT(-0500)] * peterk (i=[U2FsdGV@66.226.77.81) has joined ##uportal
[14:26:27 CDT(-0500)] <EricDalquist> dmccallum sorry for the slow response, been running through meetings today
[14:26:43 CDT(-0500)] <EricDalquist> Lets not worry too much about the performance of the cache manager portlet
[14:27:05 CDT(-0500)] <EricDalquist> in reality folks are much more likely to use something like jconsole to monitor those caches
[14:27:55 CDT(-0500)] <EricDalquist> if injecting the enumerable ICache instances is easy enough thats not a problem I just don't want that little proof of conecpt portlet to take too much time
[14:42:09 CDT(-0500)] <dmccallum> already done, actually
[14:42:20 CDT(-0500)] <EricDalquist> (smile)
[14:42:42 CDT(-0500)] <dmccallum> thanks for the feedback, though.
[15:16:36 CDT(-0500)] <dmccallum> peterk, nick said you may have had some specific ideas around re-organizing spring bean definition files.
[15:16:45 CDT(-0500)] <dmccallum> don't know if you checked the irc logs from yesterday... i'm working on flattening the context heirarchy to avoid deadlocks and am reorganizing the file system layout somewhat as i go.
[15:18:04 CDT(-0500)] <dmccallum> so i wanted to check with you before i progressed too far in a direction you'd be uncomfortable with
[15:23:36 CDT(-0500)] <dmccallum> the two most significant changes, i think, aside from eliminating the ContextResolvingXxxApplicationContexts and ContextResolvingXxxBeanFactories, are 1) to split bean def files into several categories represented by file system directories, all below WEB-INF/beans: gap, portal, portal/contexts, portlets and wsrp and 2) switch to .-based bean namespacing rather than /-based
[15:27:13 CDT(-0500)] <EricDalquist> dmccallum & all ... just to let you know I'll be out of town Thursday & Friday this week
[15:27:30 CDT(-0500)] <dmccallum> for situations where a beanfactory had been wrapped up in a proxy, i have been planning on using name-based dynamic proxying for the beans which had previously 'belonged' to those factories, but i haven't gotten to the point of actually implementing such definitions.
[15:27:44 CDT(-0500)] <dmccallum> thanks eric, i'll be sure to let charise know
[15:27:59 CDT(-0500)] <EricDalquist> thank you
[19:45:56 CDT(-0500)] * dmccallum (n=dmccallu@uni1.unicon.net) has left ##uportal
[20:40:11 CDT(-0500)] * dmccallum (n=dmccallu@ip70-190-4-78.ph.ph.cox.net) has joined ##uportal
[21:23:40 CDT(-0500)] * jackflack (n=xor@ip70-171-225-9.tc.ph.cox.net) has joined ##uportal
[21:24:26 CDT(-0500)] <jackflack> Always this quiet around here?
[21:24:31 CDT(-0500)] <jackflack> (smile)
[21:48:03 CDT(-0500)] <jackflack> exit
[22:35:47 CDT(-0500)] * dmccallum (n=dmccallu@ip70-190-4-78.ph.ph.cox.net) has left ##uportal