uPortal IRC Logs-2007-10-31

Owned by Apereo Infrastructure
Last updated: Oct 31, 2007Version comment
7 min read

[07:40:49 EDT(-0400)] * jayshao (n=jayshao@jayshao.oirt.rutgers.edu) has joined ##uportal
[09:29:18 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[09:42:34 EDT(-0400)] <EricDalquist> morning all
[09:42:35 EDT(-0400)] * athena7 (n=athena@c-24-2-130-130.hsd1.ct.comcast.net) has joined ##uportal
[10:40:59 EDT(-0400)] <EricDalquist> hey there athena7 you got your wiki page: http://www.ja-sig.org/wiki/display/JCON/Winter+2007+Question+Talks
[10:41:16 EDT(-0400)] <athena7> thanks!
[10:41:24 EDT(-0400)] <EricDalquist> yup (smile)
[10:41:29 EDT(-0400)] <athena7> i'll add some stuff to it today
[10:42:18 EDT(-0400)] <EricDalquist> great!
[10:48:24 EDT(-0400)] * pberry (n=pberry@waldorf.CSUChico.EDU) has joined ##uportal
[10:49:56 EDT(-0400)] * pberry yawns
[10:49:59 EDT(-0400)] <pberry> morning...
[10:50:05 EDT(-0400)] <EricDalquist> hello
[10:50:51 EDT(-0400)] <pberry> did you get your channel question answered?
[10:51:21 EDT(-0400)] <EricDalquist> yup
[10:51:31 EDT(-0400)] <EricDalquist> now I'm figuring out how I want to implement this
[10:51:50 EDT(-0400)] <EricDalquist> and trying to figure out what exactly andrew_petro_ubu would have done to convert PortalSessionManager to a Spring Controller
[10:52:47 EDT(-0400)] <pberry> Nice. I'm just making Outlook crash by searching for messages.
[10:53:18 EDT(-0400)] <EricDalquist> mmm windows
[10:56:21 EDT(-0400)] <pberry> yeah...awesome
[10:56:53 EDT(-0400)] <pberry> Although it was interesting to see the huff that got caused when OS X 10.5 didn't ship with Java 6
[10:57:01 EDT(-0400)] <EricDalquist> I realized a few days ago that it has been almost 6 months since I've used windows as a primary OS
[10:57:03 EDT(-0400)] <EricDalquist> uhg
[10:57:04 EDT(-0400)] <EricDalquist> yeah
[10:57:12 EDT(-0400)] <EricDalquist> I'm getting rather annoyed with Apple
[10:57:29 EDT(-0400)] <EricDalquist> I know Jobs has something personal against Java
[10:57:31 EDT(-0400)] <pberry> yeah, but historically the Java releases come almost right after the OS release
[10:57:57 EDT(-0400)] <pberry> most folks who remember this stuff are expecting something to pop in Software Update this week
[10:58:05 EDT(-0400)] <EricDalquist> but they seem to be excluding a large group of people that want to use their OS by being so freaking slow with Java updates
[10:58:08 EDT(-0400)] <EricDalquist> yeah
[10:58:20 EDT(-0400)] <EricDalquist> though it is still inexcusably late
[10:58:28 EDT(-0400)] <pberry> well...lets put "large" into perspective here (wink)
[10:58:32 EDT(-0400)] <EricDalquist> how long has JDK6 been out and final now?
[10:58:47 EDT(-0400)] <EricDalquist> well, large compared to apple's 'huge' userbase
[10:58:52 EDT(-0400)] <pberry> shipping an OS is about priorities
[10:58:56 EDT(-0400)] <EricDalquist> yeah
[10:59:06 EDT(-0400)] <pberry> I don't think anybody can make the argument that Java 6 is a top priority for apple
[10:59:09 EDT(-0400)] <EricDalquist> I still don't see why Apple doesn't let Sun do a standard JVM for their OS
[10:59:12 EDT(-0400)] <pberry> yeah, people want it
[11:00:18 EDT(-0400)] <pberry> Sun said that they didn't because "Apple said they wanted to do it"
[11:00:24 EDT(-0400)] <pberry> which...I dunno...sounds kinda lame
[11:00:39 EDT(-0400)] <pberry> http://blogs.sun.com/jag/entry/solaris_and_os_x_continued
[11:00:49 EDT(-0400)] <pberry> James Gosling weighs in
[11:01:11 EDT(-0400)] <pberry> http://stuffthathappens.com/blog/2007/10/28/os-x-java-definitive-timeline/
[11:02:10 EDT(-0400)] <EricDalquist> though I wouldn't be surprised if Apple did say we'll do it ourselves
[11:02:22 EDT(-0400)] <pberry> I think John Gruber nailed the reasons why Apple shipped Ruby 1.8.6 and not Java 6 http://daringfireball.net/2007/10/shipping_means_prioritizing
[11:02:48 EDT(-0400)] <EricDalquist> I really just want them to stop doing it themselves (smile)
[11:02:52 EDT(-0400)] <pberry> Yeah, but if Sun really cared about Java on OS X, why wouldn't they have their own?
[11:03:07 EDT(-0400)] <EricDalquist> I think because apple asked them not to
[11:03:20 EDT(-0400)] <EricDalquist> since 'they were going to do it'
[11:03:53 EDT(-0400)] <pberry> Yeah, but making a jdk and having an integrated system release are two different things
[11:04:20 EDT(-0400)] <EricDalquist> yup
[11:04:31 EDT(-0400)] <pberry> Look at all the crap jvm offerings on Linux these days. Thank Zeus that Sun still does a linux jdk
[11:04:59 EDT(-0400)] <EricDalquist> oh I do
[11:05:09 EDT(-0400)] <EricDalquist> and I'm very happy that they have a deal with the ubuntu folks
[11:05:28 EDT(-0400)] <EricDalquist> so that I can remove gcj and install Sun's JDK via the package manager
[11:05:44 EDT(-0400)] <pberry> Sun protected Java so long and so fiercely that it's not shocking in the least that they new run into problems like this.
[11:06:06 EDT(-0400)] <pberry> Nobody bows down to kiss the Java ring, and Java developers are the ones that get screwed.
[11:06:16 EDT(-0400)] <EricDalquist> (smile)
[11:10:19 EDT(-0400)] <EricDalquist> hey andrew_petro_ubu are you around?
[11:12:21 EDT(-0400)] <pberry> Ah, looks like you can get RCs of Java 6 from ADC
[11:12:29 EDT(-0400)] <EricDalquist> yup
[11:12:37 EDT(-0400)] <EricDalquist> but removing it is a PITA I hear
[11:13:05 EDT(-0400)] <pberry> ah, good times
[11:13:23 EDT(-0400)] <EricDalquist> (smile)
[13:17:16 EDT(-0400)] * agherna (n=argherna@cites-agherna01.ci.uiuc.edu) has joined ##uportal
[13:17:45 EDT(-0400)] <agherna> anyone got time for a persondirectory.xml question?
[13:20:53 EDT(-0400)] <agherna> Well, let me throw this out. It's lunchtime in the midwest, so I suspect everyone is feeling like taking a break wherever you are...
[13:21:13 EDT(-0400)] <agherna> Right now, we have http://rafb.net/p/xMxeIR88.html in our persondirectory.xml
[13:21:46 EDT(-0400)] <agherna> We need to add an additional datasource. We've added a bean to the persondirectory.xml file but now we need to add it to the personAttributeDao.
[13:21:50 EDT(-0400)] <agherna> how do we do that?
[13:25:53 EDT(-0400)] <agherna> ahh... this is interesting.
[13:26:36 EDT(-0400)] <agherna> right now, the class we have set up is the default for personAttributeDao (org.jasig.portal.services.persondir.support.CashingPersonAttributeDaoImpl)
[13:26:53 EDT(-0400)] <agherna> that particular bean looks like it's set up to handle only 1 datasource at a time.
[13:29:12 EDT(-0400)] * athena7 (n=athena@lumina.its.yale.edu) has joined ##uportal
[13:29:19 EDT(-0400)] <agherna> but, i just saw a class called CascadingPersonAttributeDao in the same package. That one looks like it can take a list of datasources and builds a Composite that acts just like the CachingPersonAttributeDaoImpl class.
[13:30:00 EDT(-0400)] <agherna> So, should I take my personAttributeDao defined at http://rafb.net/p/xMxeIR88.html and change the class to use the CascadingPersonAttributeDao?
[13:30:14 EDT(-0400)] <agherna> have I answered my own question?
[13:33:15 EDT(-0400)] <agherna> wait, looking at the javadoc, it seems that I need to make another bean just like http://rafb.net/p/xMxeIR88.html, and then build a CascadingPersonAttributeDao by passing it the list I want to use to build my composite and also (following the recommendation in the javadoc) put an EchoPersonAttributeDaoImpl at the front of the list, correct?
[13:34:31 EDT(-0400)] <agherna> so in personDirectory, my personAttributeDao bean would be the CascadingPersonAttributeDao and I would pass in a list of refs to the other daos I set up.
[13:36:43 EDT(-0400)] <agherna> I'd want to cache the attributes, so I would need to build a list of CachingPersonAttributeDaoImpls each ref'ing the various datasources I set up (I don't think that I need to query my stores on each request if I don't have to), right?
[13:41:52 EDT(-0400)] <agherna> thoughts anyone?
[13:47:38 EDT(-0400)] <EricDalquist> hey there agherna
[13:47:42 EDT(-0400)] <EricDalquist> just got back from a meeting
[13:48:01 EDT(-0400)] <EricDalquist> I'll read over your questions and see if I can help (smile)
[13:49:56 EDT(-0400)] <EricDalquist> ok
[13:50:06 EDT(-0400)] <EricDalquist> so the cascading DAO probably isn't what you want
[13:51:14 EDT(-0400)] <EricDalquist> it is designed to run through a list of DAOs using the compiled results from all previous DAOs as the seed for each subsiquent DAO
[13:51:14 EDT(-0400)] <agherna> ok, so which one should I do? I need to chain up to n datasources...
[13:51:52 EDT(-0400)] <EricDalquist> do any of the queries of the N datasources depend on the results of any of the other datasources?
[13:52:01 EDT(-0400)] <agherna> no
[13:52:23 EDT(-0400)] <EricDalquist> ok
[13:52:40 EDT(-0400)] <EricDalquist> you want MergingPersonAttributeDaoImpl
[13:53:07 EDT(-0400)] <EricDalquist> so CachingPersonAttributeDaoImpl is your root "personAttributeDao" bean
[13:53:15 EDT(-0400)] <agherna> ok
[13:53:16 EDT(-0400)] <EricDalquist> it points to a MergingPersonAttributeDaoImpl
[13:53:19 EDT(-0400)] <agherna> ok
[13:53:35 EDT(-0400)] <EricDalquist> which points to a List of IPersonAttributeDaos to execute and merge the results from
[13:53:43 EDT(-0400)] <agherna> i see
[13:54:08 EDT(-0400)] <agherna> and the cachingpersonattributdaoimpl caches those attributes after they've been merged, right?
[13:54:12 EDT(-0400)] <EricDalquist> yup
[13:54:44 EDT(-0400)] <agherna> now does it matter if the attributes from one datasource are different from another?
[13:55:10 EDT(-0400)] <EricDalquist> the names?
[13:55:11 EDT(-0400)] <agherna> that is, if i have some in one and different ones in the other, are those merged
[13:55:23 EDT(-0400)] <EricDalquist> yup
[13:55:40 EDT(-0400)] <EricDalquist> the merging dao even allows you to configure the merge strategy
[13:55:43 EDT(-0400)] <EricDalquist> there is a IAttributeMerger interface
[13:55:59 EDT(-0400)] <EricDalquist> the default configuration is to use MultivaluedAttributeMerger
[13:56:25 EDT(-0400)] <EricDalquist> so if two datasources return the same attribute name for one query that merger interface will make it a multivalued attribute
[13:56:27 EDT(-0400)] <agherna> hmm, well if we have to, we'll configure it. but i want to learn a little more about the default first. that might do it.
[13:57:03 EDT(-0400)] <agherna> i don't think that will be an issue (m datasources returning the same attribute)
[13:57:20 EDT(-0400)] <EricDalquist> (smile)
[13:58:03 EDT(-0400)] <agherna> so to summarize, the "root" bean is a cachingpersonattributedao that points to a mergingpersonattributedaoimple
[13:58:31 EDT(-0400)] <EricDalquist> yup
[13:58:51 EDT(-0400)] <agherna> the mergingpersonattributedaoimpl will take a list of the datasources i'll be querying to get the attribute values and merge them in to one set of attrs.
[13:59:06 EDT(-0400)] <agherna> that's good stuff, man
[13:59:09 EDT(-0400)] <EricDalquist> yup
[13:59:11 EDT(-0400)] <EricDalquist> what the cascading dao would have given you is similar to the merging dao (it runs a list of child DAOs and merges the results) the difference is it provides each subsequent child dao with all previous results, allowing you to have DAO B depend on results from DAO A
[13:59:12 EDT(-0400)] <agherna> now while i have you here...
[13:59:14 EDT(-0400)] <agherna> (smile)
[13:59:23 EDT(-0400)] <agherna> I have another quesiton about rdbmservices
[13:59:23 EDT(-0400)] <EricDalquist> heh, I do have a phone call I have to get on to in just a minute
[13:59:25 EDT(-0400)] <EricDalquist> ok
[14:00:06 EDT(-0400)] <agherna> if i pass a jndi name to it like jdbc/someotherdatasource, it'll look for that in the uportal.xml file (the context file) and use that instead of jndi/PersonDb, right?
[14:00:28 EDT(-0400)] <EricDalquist> yeah
[14:00:31 EDT(-0400)] <EricDalquist> well it looks for it in JNDI
[14:00:35 EDT(-0400)] <agherna> poifect
[14:00:41 EDT(-0400)] <EricDalquist> which is configured by that .xml file
[14:00:50 EDT(-0400)] <agherna> which xml file?
[14:00:52 EDT(-0400)] <agherna> uportal.xml?
[14:00:57 EDT(-0400)] <EricDalquist> yeah
[14:00:59 EDT(-0400)] <EricDalquist> you may need to double check the prefxing
[14:01:09 EDT(-0400)] <EricDalquist> you may only need to pass "foodb" not "jdbc/foodb"
[14:01:12 EDT(-0400)] <agherna> yeah, we've been through that one already
[14:01:21 EDT(-0400)] <agherna> right
[14:01:34 EDT(-0400)] <EricDalquist> ok
[14:01:35 EDT(-0400)] <EricDalquist> meeting time
[14:01:41 EDT(-0400)] <agherna> buh bye and thanks
[14:44:23 EDT(-0400)] * ac_chan (n=alex@tempoutsidepix.pratt.edu) has joined ##uportal
[14:44:26 EDT(-0400)] <ac_chan> hello
[15:20:48 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[15:21:00 EDT(-0400)] <EricDalquist> hey there andrew_petro_ubu
[15:42:20 EDT(-0400)] * apetro_win_hom1 (n=andrew_p@ip68-99-92-218.ph.ph.cox.net) has joined ##uportal
[16:22:58 EDT(-0400)] <andrew_petro_ubu> howdy
[17:03:29 EDT(-0400)] * michelled (n=team@142.150.154.114) has left ##uportal
[17:20:34 EDT(-0400)] * athena7 (n=athena@c-24-2-130-130.hsd1.ct.comcast.net) has joined ##uportal
[17:31:35 EDT(-0400)] * colinclark (n=atrcwrk2@142.150.154.101) has joined ##uportal
[18:08:36 EDT(-0400)] * deuce (n=deuce@uni1.unicon.net) has joined ##uportal
[18:23:30 EDT(-0400)] * colinclark (n=atrcwrk2@142.150.154.101) has left ##uportal
[18:26:07 EDT(-0400)] * colinclar1 (n=atrcwrk2@142.150.154.101) has joined ##uportal
[18:26:17 EDT(-0400)] * colinclar1 (n=atrcwrk2@142.150.154.101) has left ##uportal
[18:37:40 EDT(-0400)] * awills (n=awills@uni1.unicon.net) has joined ##uportal
[19:56:55 EDT(-0400)] * awills (n=awills@uni1.unicon.net) has left ##uportal
[20:18:20 EDT(-0400)] * pberry (n=pberry@waldorf.CSUChico.EDU) has joined ##uportal
[21:19:23 EDT(-0400)] * jayshao (n=jayshao@pool-72-79-105-68.nwrknj.east.verizon.net) has joined ##uportal
[22:50:25 EDT(-0400)] * jayshao (n=jayshao@pool-72-79-105-68.nwrknj.east.verizon.net) has joined ##uportal