uPortal IRC Logs-2007-05-29

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

[07:25:56 EDT(-0400)] * jayshao (n=jayshao@ool-44c59034.dyn.optonline.net) has joined ##uportal
[10:36:20 EDT(-0400)] * jayshao (n=jayshao@jshao.rutgers.edu) has joined ##uportal
[10:44:54 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[10:48:51 EDT(-0400)] * lescour (n=JBouncer@adsl-38-10-98.tulsaconnect.com) has joined ##uportal
[10:56:52 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[11:24:54 EDT(-0400)] <apetro_andreasde> good morning world
[11:25:05 EDT(-0400)] <EricDalquist> good morning andrew
[11:25:16 EDT(-0400)] <apetro_andreasde> So Drew wills and I have dreamed up a subtly improved approach to handling dependencies in uP 2.6
[11:25:28 EDT(-0400)] <apetro_andreasde> getting to Maven would be great but is too much to bite off
[11:25:42 EDT(-0400)] <apetro_andreasde> in fact, I moved the mvn stuff out of the way last night since it wasn't being kept up to date
[11:25:46 EDT(-0400)] <apetro_andreasde> however
[11:25:55 EDT(-0400)] <apetro_andreasde> build.properties enumerating every dependency
[11:26:05 EDT(-0400)] <apetro_andreasde> and then having this sprinkled throughout the build.xml
[11:26:05 EDT(-0400)] <apetro_andreasde> is tiresome
[11:26:08 EDT(-0400)] <EricDalquist> yes
[11:26:09 EDT(-0400)] <EricDalquist> very
[11:26:14 EDT(-0400)] <apetro_andreasde> it's not actually true that anyone doesn't have to edit build.xml
[11:26:20 EDT(-0400)] <apetro_andreasde> I bet most schools introduce add'l dependencies
[11:26:24 EDT(-0400)] <apetro_andreasde> and so everyone hacks it
[11:26:26 EDT(-0400)] <apetro_andreasde> so
[11:26:31 EDT(-0400)] <apetro_andreasde> Maven has this lovely concept of scopes
[11:26:36 EDT(-0400)] <apetro_andreasde> here's what I've done
[11:26:43 EDT(-0400)] <apetro_andreasde> created subfolders of /lib for the Maven scopes
[11:26:53 EDT(-0400)] <apetro_andreasde> and invented a couple new scopes for special behavior
[11:27:00 EDT(-0400)] <apetro_andreasde> the build auto-discovers libs in those directories
[11:27:12 EDT(-0400)] <EricDalquist> sounds great
[11:27:17 EDT(-0400)] <apetro_andreasde> want to change .jar version? Stick the updated version in the directory corresponding to what you want the build to do with it
[11:27:23 EDT(-0400)] <EricDalquist> that is what most of my ant based projects do
[11:27:26 EDT(-0400)] <apetro_andreasde> want to add a dependency? Great, stick it into the directory
[11:27:27 EDT(-0400)] <apetro_andreasde> k
[11:27:36 EDT(-0400)] <apetro_andreasde> I'm probably going to finish testing this this morning and check it in
[11:27:41 EDT(-0400)] <EricDalquist> I end up with a /lib and /testlib usually
[11:27:55 EDT(-0400)] <EricDalquist> but the lib/compile lib/test lib/provided etc sound good
[11:28:23 EDT(-0400)] <apetro_andreasde> right now it's /lib/compile /lib/test /lib/provided /lib/runtime /lib/container-common /lib/container-shared /lib/common-endorsed
[11:28:23 EDT(-0400)] <EricDalquist> and sounds like a much saner way of dealing with jars in up2
[11:28:32 EDT(-0400)] <EricDalquist> wonderful
[11:28:32 EDT(-0400)] <apetro_andreasde> it's incrementally saner
[11:28:47 EDT(-0400)] <apetro_andreasde> I fought with the Maven for a short while and concluded it's not worth doing for 2.6
[11:29:04 EDT(-0400)] <apetro_andreasde> for a 2.7 maybe, especially if we go after my dream of an integrated build including channels
[11:29:44 EDT(-0400)] <EricDalquist> yeah
[11:29:48 EDT(-0400)] <EricDalquist> this sounds like a good, fast step
[11:30:41 EDT(-0400)] <apetro_andreasde> cool.
[11:31:26 EDT(-0400)] <apetro_andreasde> rev'ing the spring.jar version was the last straw for me. Updating build.properties is a pita
[11:31:35 EDT(-0400)] <EricDalquist> yeah
[11:31:46 EDT(-0400)] <EricDalquist> and if you miss one of the like 5 places to update the version
[11:31:57 EDT(-0400)] <EricDalquist> it all goes down in flames
[11:38:55 EDT(-0400)] * andrew_petro_ubu (n=apetro@uni1.unicon.net) has joined ##uportal
[13:25:44 EDT(-0400)] * esm (n=esm@asdf.dkc.jhu.edu) has joined ##uportal
[13:27:36 EDT(-0400)] <jayshao> wow andrew_petro_ubu you really are trying to do the ubuntu kick
[13:29:18 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has left ##uportal
[13:29:25 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[13:47:29 EDT(-0400)] <esm> marco andretti's wreck at indy this weekend: http://www.youtube.com/watch?v=6DE_J1iGK84&amp;mode=related&amp;search=
[13:47:47 EDT(-0400)] <esm> and gordon's wreck: http://www.youtube.com/watch?v=2YqoJy9oaPM&amp;mode=related&amp;search=
[13:48:59 EDT(-0400)] <andrew_petro_ubu> dangerous sport
[13:49:28 EDT(-0400)] <esm> i'd love to try
[13:49:31 EDT(-0400)] <esm> (smile)
[14:03:33 EDT(-0400)] * esm (n=esm@asdf.dkc.jhu.edu) has joined ##uportal
[16:13:57 EDT(-0400)] * esm (n=esm@asdf.dkc.jhu.edu) has joined ##uportal
[16:20:06 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[16:52:54 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[17:26:01 EDT(-0400)] * hubick (n=hubick@cs14.pc.athabascau.ca) has joined ##uportal
[17:26:42 EDT(-0400)] <hubick> ping? anyone around?
[17:26:49 EDT(-0400)] <EricDalquist> hello
[17:27:28 EDT(-0400)] <hubick> Hi, i'm Chris from Athabasca University... we having the problem with stuck threads in our ALM moveWrongNodesToLostFolder
[17:28:16 EDT(-0400)] <EricDalquist> hrm, I'm not familiar with the internals of ALM at all ... have you tried the uportal-user mailing list?
[17:29:51 EDT(-0400)] <hubick> I just had some simple questions about the code...I will read more and see if I can figure it out
[17:30:07 EDT(-0400)] <EricDalquist> ok
[17:30:37 EDT(-0400)] <EricDalquist> I can see what I can do .. there may be others that check in on the chat room that might know
[17:32:30 EDT(-0400)] <hubick> AggregatedLayoutManager from 2.5.3 GA changeSiblingNodesOrder compares node priorities... it uses '<=' for comparing priorities, and I was thus wondering if that should be simply '<' in case they are equal, in which case it could be in a loop continuously reording two of equal priority?
[17:33:44 EDT(-0400)] <EricDalquist> that sounds like it could be a problem
[17:35:33 EDT(-0400)] <hubick> I have no idea what I am doing in this code, I'm just fishing around where UP looks stuck (smile)
[17:35:43 EDT(-0400)] <hubick> I can email the list then
[17:36:01 EDT(-0400)] <EricDalquist> yeah, I would be fishing (blindly) as well
[17:36:34 EDT(-0400)] <EricDalquist> that sounds like it could be a culprit but you probably want to check with the folks on uportal-user and see if anyone there has more experience
[17:37:39 EDT(-0400)] <hubick> I added the patch that was posted before...I have to see why that isn't causing it to bail out
[17:38:12 EDT(-0400)] <hubick> The loop should be capped at "10,000"... I wonder how long it takes to hit that... more research needed
[17:38:32 EDT(-0400)] <EricDalquist> if it is doing DB io in the loop it could be a long time
[17:39:02 EDT(-0400)] <hubick> I think at that point it's already loaded everything into a data structure
[17:39:17 EDT(-0400)] <EricDalquist> ah
[17:44:01 EDT(-0400)] <lescour> hubick, have you looked at the running uportal with a debugger?
[17:44:50 EDT(-0400)] <hubick> lescour, I have not
[17:45:27 EDT(-0400)] <hubick> lescour, that would require running our production cluster in debug mode, and waiting until it's triggered, or assuming we know how to trigger it (smile)
[17:45:42 EDT(-0400)] <lescour> right, it's problematic in production
[17:47:46 EDT(-0400)] <lescour> i would definitely send a message to the list eric referenced. i would be suprised if you didn't have a reply tomorrow
[17:48:21 EDT(-0400)] <hubick> lescour, I'm gonna spend some more time reading the code and see what i can see... I was just wondering if someone here would see what I saw and go "oooh that's it!"
[17:50:17 EDT(-0400)] <lescour> np
[18:15:06 EDT(-0400)] <andrew_petro_ubu> being in that lost node method in the first place is already a bad sign, iirc
[18:34:37 EDT(-0400)] <hubick> We force all our users to share the same layout... I would like to write a patch to have uP run against a read only DB connection (smile)
[18:51:39 EDT(-0400)] <andrew_petro_ubu> if all users share the same layout, then you don't gain anything from using ALM
[18:51:47 EDT(-0400)] <andrew_petro_ubu> and can implement your portal using SLM
[18:56:22 EDT(-0400)] <hubick> andrew_petro_ubu, well, I should say that their layout isn't modifiable...but we do customize it using fragments and groups
[18:56:49 EDT(-0400)] <hubick> actually all our content is fragment tabs
[18:57:04 EDT(-0400)] <hubick> all pushed. the actual user layout is empty
[18:58:54 EDT(-0400)] <andrew_petro_ubu> DLM?
[18:59:16 EDT(-0400)] <andrew_petro_ubu> since users have no personal preferences, your upgrade path should be pretty attractive.
[18:59:34 EDT(-0400)] <andrew_petro_ubu> eep, gotta run
[19:04:04 EDT(-0400)] <hubick> andrew_petro_ubu, We went with ALM, cuz it does what we need, and we thought it would be the best supported out of the box, being the default for 2.5.3.
[19:04:46 EDT(-0400)] <hubick> the plan is to switch to whatever unified layout manager is in uP 3 anyhow
[20:45:07 EDT(-0400)] * esm (n=esm@esm.qis.net) has joined ##uportal
[23:43:57 EDT(-0400)] * JASIGLogBot (i=jasigch@jasigch.Princeton.EDU) has joined ##uportal
[23:43:57 EDT(-0400)] * Topic is 'Maven 2: Definitive Guide http://www.sonatype.com/book/index.html' set by esm on 2007-04-27 13:56:12 EDT(-0400)