uPortal IRC Logs-2009-04-01

Owned by Apereo Infrastructure
Last updated: Apr 01, 2009Version comment

[07:54:58 EDT(-0400)] * michelled (n=michelle@CPE001310472ade-CM0011aefd3ca8.cpe.net.cable.rogers.com) has joined ##uportal
[07:55:51 EDT(-0400)] * Arvids (n=arvids@213.175.95.54) has joined ##uportal
[07:58:50 EDT(-0400)] * dstn (n=dstn@unaffiliated/dstn) has joined ##uportal
[08:05:09 EDT(-0400)] <dstn> grr
[08:06:45 EDT(-0400)] <dstn> portlet preferences only allow storing in the action phase but I really need to store on the view
[08:43:15 EDT(-0400)] <athena> yes
[08:43:19 EDT(-0400)] <athena> it's frustrating
[08:43:53 EDT(-0400)] <athena> anthony and i have some ongoing conversation about how to handle setting the news set id in the preferences because of that problem
[09:06:42 EDT(-0400)] * athena (n=athena@99.129.100.66) has joined ##uportal
[09:06:44 EDT(-0400)] * lennard1 (n=sparhk@wsip-98-174-242-39.ph.ph.cox.net) has joined ##uportal
[09:07:41 EDT(-0400)] <dstn> athena, can you elaborate on some of the ideas you and anthony had related to storing the id in the render phase?
[09:07:57 EDT(-0400)] <athena> yes
[09:08:12 EDT(-0400)] <athena> so have you looked at the trunk of the calendar portlet recently?
[09:08:27 EDT(-0400)] <athena> it's using AJAX for all calendar calls
[09:08:51 EDT(-0400)] <athena> so when you pull up the portlet for the first time, it displays the main page, and that that page does an ajax request to get all the data
[09:09:16 EDT(-0400)] <dstn> I haven't
[09:09:19 EDT(-0400)] <athena> ok
[09:09:39 EDT(-0400)] <dstn> aside from using ajax...any other ideas
[09:09:39 EDT(-0400)] <athena> anyway, if we agree to pull the data via ajax, rather than in that main controller
[09:09:58 EDT(-0400)] <athena> then the data will always be pulled during an action request
[09:10:08 EDT(-0400)] <dstn> i c
[09:10:50 EDT(-0400)] <athena> which means that it becomes much simpler to store the id when it's created, since you don't need to worry about render phases
[09:10:54 EDT(-0400)] <athena> does that make some sense?
[09:10:57 EDT(-0400)] <dstn> yep
[09:11:47 EDT(-0400)] <athena> ok
[09:11:54 EDT(-0400)] <athena> i know that it does force us to require javascript
[09:12:03 EDT(-0400)] <athena> but i'm not clear on whether that's an actual problem or not
[09:12:31 EDT(-0400)] <athena> i can't really think of a reasonable browser that doesn't have javascript these days, and accessibility concerns don't seem to be focusing on javascript usage
[09:14:07 EDT(-0400)] <dstn> unfortunately, I don't think I have time to do that significant of a change
[09:14:25 EDT(-0400)] <dstn> but thanks for the idea, good to know
[09:15:18 EDT(-0400)] <athena> yeah, and the database structure really has to be re-done
[09:15:35 EDT(-0400)] <athena> does that sound reasonable to you? i don't think anthony and i had settled on a final solution
[10:51:40 EDT(-0400)] * colinclark (n=colin@bas2-toronto09-1176131209.dsl.bell.ca) has joined ##uportal
[10:53:33 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[11:08:10 EDT(-0400)] * michelled (n=michelle@CPE001310472ade-CM0011aefd3ca8.cpe.net.cable.rogers.com) has joined ##uportal
[11:10:56 EDT(-0400)] * awills (n=awills@wsip-98-174-242-39.ph.ph.cox.net) has joined ##uportal
[11:41:52 EDT(-0400)] * holdorph (n=holdorph@wsip-98-174-242-39.ph.ph.cox.net) has joined ##uportal
[11:58:21 EDT(-0400)] <athena> EricDalquist: i had some thoughts about starting to gradually maven-ify our build process
[11:58:27 EDT(-0400)] <EricDalquist> great
[11:59:44 EDT(-0400)] <athena> basically yale would like to be able to build the uportal ear, put it in the maven repository, then run something on the production server to deploy the pre-built artifact
[12:00:26 EDT(-0400)] <athena> from initial experimentation it's actually pretty easy to take the TomcatEarDeployer stuff you wrote and create a small maven plugin to call that code to deploy the artifact
[12:00:52 EDT(-0400)] <EricDalquist> great
[12:00:55 EDT(-0400)] <athena> yeah
[12:01:22 EDT(-0400)] <athena> it seems like it might be worth also exploring how to edit the uportal build.xml to call that plugin rather than the ant macro
[12:01:29 EDT(-0400)] <athena> although i'm less sure how to organize all that
[12:01:50 EDT(-0400)] <EricDalquist> well that would be doing a command line mvn call
[12:01:53 EDT(-0400)] <athena> but does that sound reasonable to gradually start transitioning our ant tasks and macros to maven?
[12:01:53 EDT(-0400)] <EricDalquist> which the build.xml supports
[12:01:56 EDT(-0400)] <athena> yeah
[12:01:59 EDT(-0400)] <EricDalquist> I think so
[12:02:02 EDT(-0400)] <athena> great
[12:02:06 EDT(-0400)] <EricDalquist> I think I'd like to do it all in a single release though
[12:02:12 EDT(-0400)] <athena> yes, definitely
[12:02:12 EDT(-0400)] <EricDalquist> so if we start doing it now
[12:02:20 EDT(-0400)] <EricDalquist> when we do a 3.2 release it is all maven
[12:02:24 EDT(-0400)] <athena> i was thinking maybe start working on it in trunk for 3.2?
[12:02:26 EDT(-0400)] <athena> yeah, exactly
[12:02:28 EDT(-0400)] <EricDalquist> yup
[12:02:41 EDT(-0400)] <athena> since we can use the ant tasks plugin, i don't think it should be too horrible
[12:02:58 EDT(-0400)] <athena> some tasks might even be able to be directly turned into ant plugin calls
[12:03:12 EDT(-0400)] <athena> and then we'll have access to things like profiles, etc.
[12:03:19 EDT(-0400)] <EricDalquist> yeah
[12:03:35 EDT(-0400)] <athena> turns our profile support is actually pretty painful in the maven-ant-tasks
[12:03:43 EDT(-0400)] <EricDalquist> I found out we have to be careful with marking build plugins in the root pom <inherit>false</inherit>
[12:03:56 EDT(-0400)] <EricDalquist> yeah, I'm not supprisied
[12:04:00 EDT(-0400)] <athena> yeah everything is inherited by default, right?
[12:04:08 EDT(-0400)] <EricDalquist> those ant tasks are not a full replacement for maven
[12:04:10 EDT(-0400)] <EricDalquist> yeah
[12:04:43 EDT(-0400)] <athena> so i'm thinking for yale's build, it may be useful to use classifiers to mark built artifacts as being configured for a particular deployment environment
[12:04:46 EDT(-0400)] <athena> does that sound reasonable?
[12:05:17 EDT(-0400)] <EricDalquist> http://www.ja-sig.org/wiki/display/UPC/Moving+to+an+all+Maven+build
[12:05:23 EDT(-0400)] <EricDalquist> yes
[12:05:25 EDT(-0400)] <EricDalquist> I think so
[12:05:31 EDT(-0400)] <EricDalquist> I'm not that familiar with classifiers
[12:05:45 EDT(-0400)] <EricDalquist> the only thing I remember about them is it breaks if they have different dependencies
[12:05:58 EDT(-0400)] <EricDalquist> since you don't get 1 pom per classified artifact
[12:06:06 EDT(-0400)] <EricDalquist> you get 1 pom which all classified artifacts use
[12:07:55 EDT(-0400)] <EricDalquist> so lets try and use that wiki page to document what the work is going to be doing
[12:08:02 EDT(-0400)] <EricDalquist> just so we have a point to collaborate on
[12:08:10 EDT(-0400)] <athena> ah, interesting
[12:08:20 EDT(-0400)] <athena> and great, yes, having a shared page would be really helpful
[12:09:06 EDT(-0400)] <athena> by the way, i think you'd mentioned that you weren't yet using the maven release plugin for uPortal?
[12:10:25 EDT(-0400)] <EricDalquist> right, I'm not
[12:10:31 EDT(-0400)] <EricDalquist> partially because of the ant build
[12:10:44 EDT(-0400)] <EricDalquist> there are some version numbers and things I have to munge by hand
[12:13:02 EDT(-0400)] <athena> ah ok
[12:13:14 EDT(-0400)] <athena> i was wondering what the impediments to getting that in order were
[12:13:31 EDT(-0400)] <EricDalquist> the stuff in bootstrap primarily
[12:13:39 EDT(-0400)] <EricDalquist> there are some poms there that are used for utility tasks
[12:14:53 EDT(-0400)] <athena> ah ok
[12:39:20 EDT(-0400)] * Arvids (n=arvids@213.175.95.54) has left ##uportal
[12:41:08 EDT(-0400)] <athena> yay, got anthony and i's conversation back on-list
[12:41:35 EDT(-0400)] <athena> eric, i dont' know if you have any thoughts about making the bookmarks, calendar, and news portlets all share some similar set creating and resolving strategies
[12:53:26 EDT(-0400)] * SusanB (i=user@susan-x200.its.yale.edu) has joined ##uportal
[13:03:13 EDT(-0400)] <athena> i think getting this maven plugin together would let us not have a build.properties anymore
[13:03:37 EDT(-0400)] <EricDalquist> probably not
[13:03:46 EDT(-0400)] <EricDalquist> maven doesn't really read from .properties files
[13:03:49 EDT(-0400)] <EricDalquist> which is REALLY annoyuing
[13:04:11 EDT(-0400)] <EricDalquist> these are some of my issues with using maven for more than just a build tool
[13:04:19 EDT(-0400)] <EricDalquist> but I don't think we have much of a choice
[13:04:30 EDT(-0400)] <athena> if we used maven we shouldn't need it at all
[13:04:41 EDT(-0400)] <EricDalquist> and I'm not sure what you mean by "eric, i dont' know if you have any thoughts about making the bookmarks, calendar, and news portlets all share some similar set creating and resolving strategies"
[13:04:44 EDT(-0400)] <athena> it looks like the only things we're using it for now is basically properties for the tomcat deployment
[13:04:48 EDT(-0400)] <EricDalquist> yeah
[13:04:56 EDT(-0400)] <EricDalquist> and the clean toggles at the bottom
[13:05:12 EDT(-0400)] <athena> oh, sorry - anthony just sent an email about some of the NewsSetResolving stuff, which i'd originally modeled off the bookmarks portlet
[13:05:21 EDT(-0400)] <athena> right, so those could all become maven profile properties
[13:05:27 EDT(-0400)] <athena> if we were using maven to do this stuff
[13:05:42 EDT(-0400)] <EricDalquist> which list for that email?
[13:05:55 EDT(-0400)] <athena> portlet-dev
[13:06:07 EDT(-0400)] <EricDalquist> my email must be slow (tongue)
[13:06:17 EDT(-0400)] <EricDalquist> which is sad since I work in the same building as the listserv hardware
[13:06:34 EDT(-0400)] <athena> huh, you know, i don't think i've gotten it yet either
[13:06:39 EDT(-0400)] <athena> just the copy that was addressed to me
[13:06:48 EDT(-0400)] <EricDalquist> lists may be slow here
[13:06:49 EDT(-0400)] <athena> but it does look like he cc'd portlet-dev
[13:06:51 EDT(-0400)] <athena> wird
[13:06:51 EDT(-0400)] <EricDalquist> they get overloaded at times
[13:06:53 EDT(-0400)] <athena> ah
[13:07:01 EDT(-0400)] <athena> so maybe it just hasn't gone through the list yet
[13:07:24 EDT(-0400)] <EricDalquist> yeah
[13:07:58 EDT(-0400)] <athena> so for maven, i'm envisioning something like "mvn install uportal:deploy-ear"
[13:08:15 EDT(-0400)] <athena> which could then pick up those properties that are currently in build.properties from a profile or the command line
[13:08:53 EDT(-0400)] <EricDalquist> sounds good
[13:09:05 EDT(-0400)] <EricDalquist> I wonder if there is any way to declare dependencies like you can in ant
[13:09:17 EDT(-0400)] <athena> hm, i'm not sure - what do you have in mind?
[13:09:18 EDT(-0400)] <EricDalquist> like 'deploy-ear' triggers an install if the latest code isn't already installed
[13:09:30 EDT(-0400)] <athena> gotcha
[13:09:47 EDT(-0400)] <athena> i'm not sure if there's a way to do it short of calling "install"
[13:09:52 EDT(-0400)] <EricDalquist> yeah
[13:09:54 EDT(-0400)] <EricDalquist> and that may be ok
[13:10:03 EDT(-0400)] <athena> yeah i mean it's easy enough to do on the same command line
[13:10:07 EDT(-0400)] <EricDalquist> the code right now pretty much does that every time
[13:10:10 EDT(-0400)] <athena> yeah
[13:10:10 EDT(-0400)] <EricDalquist> yaeh
[13:10:19 EDT(-0400)] <EricDalquist> I just want the CL to be as easy as possible
[13:10:22 EDT(-0400)] <athena> definitely
[13:10:32 EDT(-0400)] <athena> i really don't want to require some hideous string of arguments
[13:10:43 EDT(-0400)] <EricDalquist> like it would be bad if doing 'mvn deploy-ear' grabs a snapshot from developer.jasig.org and deploys that (tongue)
[13:11:02 EDT(-0400)] <athena> yeah, that's the one thing i'm worrying about a little bit right now
[13:11:27 EDT(-0400)] <athena> i think it actually would be really nice to be able to deploy something from a repository
[13:11:38 EDT(-0400)] <athena> but we don't want to create something that will do that by accident
[13:11:44 EDT(-0400)] <EricDalquist> right
[13:11:54 EDT(-0400)] <EricDalquist> so part of this will be documenting how to run a local maven repository
[13:14:59 EDT(-0400)] <athena> yeah
[13:18:14 EDT(-0400)] <athena> really i think some of the problem is that we have configuration information in an artifact that's not identified with that configuration
[13:20:22 EDT(-0400)] <EricDalquist> ?
[13:21:30 EDT(-0400)] <athena> so to maven, the uportal artifacts on a locally built system look exactly the same as what's in the jasig repository
[13:21:42 EDT(-0400)] <EricDalquist> ah yeah
[13:21:45 EDT(-0400)] <athena> even though they have different database properties and such
[13:22:08 EDT(-0400)] <EricDalquist> I really think the end goal should be an overlay build
[13:22:21 EDT(-0400)] <EricDalquist> with a groupid specific for the deployer
[13:22:36 EDT(-0400)] <athena> yes
[13:22:44 EDT(-0400)] <athena> the problem is keeping the overlays in syn
[13:22:46 EDT(-0400)] <athena> sync
[13:22:51 EDT(-0400)] <EricDalquist> perhaps we even look at distributing a template
[13:22:52 EDT(-0400)] <EricDalquist> yes
[13:22:58 EDT(-0400)] <athena> yeah we'd need to distribute the overlay
[13:23:11 EDT(-0400)] <athena> and be responsible for changing files in the overlay whenever the base files are changed
[13:23:32 EDT(-0400)] <athena> i know manchester has started putting artifacts in the repository that contain filter tokens
[13:23:33 EDT(-0400)] <EricDalquist> yup
[13:23:45 EDT(-0400)] <athena> and just filtering the artifacts themselves rather than using overlays, for that reason
[13:23:50 EDT(-0400)] <athena> i think yale's going to follow the same strategy
[13:39:51 EDT(-0400)] <EricDalquist> neat: [java] Caused by: java.lang.NullPointerException
[13:39:51 EDT(-0400)] <EricDalquist> [java] at org.dom4j.io.XMLWriter.writeNode(XMLWriter.java:1297)
[13:40:22 EDT(-0400)] <EricDalquist> that's from trying to do a crn-export from our 3.1ish data with the 3.1-ga code (smile)
[13:40:29 EDT(-0400)] <EricDalquist> I'm thinking we have some data issues (tongue)
[13:41:00 EDT(-0400)] <athena> oh, eesh
[13:42:40 EDT(-0400)] <awills> oh noes
[13:43:06 EDT(-0400)] <awills> do you want to post more of the stack?
[13:43:39 EDT(-0400)] <EricDalquist> sure
[13:48:01 EDT(-0400)] <EricDalquist> http://uportal.pastebin.com/m70b51f88
[13:48:15 EDT(-0400)] <EricDalquist> the line in dom4j: http://www.dom4j.org/dom4j-1.6.1/xref/org/dom4j/io/XMLWriter.html#1283
[13:49:13 EDT(-0400)] <EricDalquist> so the problem is somewhere in the export a null node is getting appended
[13:49:22 EDT(-0400)] <EricDalquist> I'm wondering if append-node should detect that
[13:50:24 EDT(-0400)] <awills> probably should... you're further than me into the issue – are you sure it's being appended? not attempting to write a null node?
[13:51:42 EDT(-0400)] <awills> looks like it's writing a null document
[13:59:17 EDT(-0400)] <EricDalquist> ah, that could be the case too
[14:07:11 EDT(-0400)] * Sememmon (n=Sememmon@uni1.unicon.net) has joined ##uportal
[14:07:40 EDT(-0400)] <EricDalquist> is there an easy way to save a complete web page and CSS?
[14:08:00 EDT(-0400)] <EricDalquist> Firefox doesn't seem to traverse down to CSS/images included by CSS
[14:10:14 EDT(-0400)] <athena> i don't know eric
[14:10:19 EDT(-0400)] <athena> i've yet to do that successfully either
[14:21:54 EDT(-0400)] <EricDalquist> well after a lot of manual file copies I have a full save of one of our portal pages: https://mywebspace.wisc.edu/dalquist/web/MUM_UI/uP31Merge/mum_academics.html
[14:26:03 EDT(-0400)] <EricDalquist> oh yeah I know what is going on with the export
[14:26:07 EDT(-0400)] <EricDalquist> I thought I had fixed this (tongue)
[14:27:17 EDT(-0400)] <EricDalquist> so the issue is some of the export tasks may return null (and this is valid)
[14:27:26 EDT(-0400)]

<EricDalquist> like: <write-document node="$

Unknown macro: {crn(export-layout.crn)}

" file="$

Unknown macro: {EXPORT_DIR}

/layout/$

Unknown macro: {SAFE_USER_NAME}

.layout"/>


[14:27:41 EDT(-0400)] <EricDalquist> so the question is what should write-document do when passed a null node
[14:34:22 EDT(-0400)] <awills> yeah... we can either (1) wrap <write-document> with other operations that detect and exclude nulls or (2) enhance <write-document> to respond to nulls in some way (like log.warn())
[14:35:15 EDT(-0400)] <EricDalquist> so for now I'm just wrapping with an IF
[14:35:26 EDT(-0400)] <EricDalquist> but I'll leave it up to you on what to do with write-document
[14:35:37 EDT(-0400)] <awills> #2 would be more succinct and approachable, if we feel we have a good handle on the general case
[14:35:46 EDT(-0400)] <EricDalquist> yeah
[14:35:57 EDT(-0400)] <EricDalquist> I think #2 would be a good fix as well
[14:36:04 EDT(-0400)] <awills> yeah so do it
[14:36:08 EDT(-0400)] <awills> I*
[14:36:10 EDT(-0400)] <awills> oops
[14:36:20 EDT(-0400)] <EricDalquist> perhaps add an ignore-null="true" attribute to
[14:36:21 EDT(-0400)] <EricDalquist> too*
[14:36:42 EDT(-0400)] <EricDalquist> so you can flag that it is 'expected' to be passing in null nodes as to not get overloaded with WARN messages
[14:36:47 EDT(-0400)] <awills> hmmm, that makes some sense as well
[14:37:42 EDT(-0400)] <awills> what about a CANCEL_ON_NULL or somesuch, to indicate that nulls are expected and should cancel the operation?
[14:38:25 EDT(-0400)] <EricDalquist> sure, ignore may be a better term ... cancel makes me think 'cancel the whole script'
[14:39:51 EDT(-0400)] <awills> yeah, i'm talking about something slightly different... the operation itself, not the log msg
[14:39:59 EDT(-0400)] <EricDalquist> ah
[14:58:39 EDT(-0400)] <EricDalquist> awills: in your new layout export work
[14:59:05 EDT(-0400)] <EricDalquist> are you having to create an instance of IUserLayoutManager for each layout you're exporting?
[14:59:37 EDT(-0400)] <awills> nope, just one instance of IUserLayoutStore defined in importExportContext.xml
[14:59:50 EDT(-0400)] <EricDalquist> ok
[15:00:20 EDT(-0400)] <awills> i need a rudimentary instance of IPerson and UserProfile / layout export
[15:00:38 EDT(-0400)] <awills> be back in a few
[15:00:42 EDT(-0400)] <EricDalquist> k
[15:24:55 EDT(-0400)] * colinclark (n=colin@142.150.154.101) has joined ##uportal
[16:04:54 EDT(-0400)] <EricDalquist> hrm ...
[16:05:38 EDT(-0400)] <EricDalquist> so I added support a while back for transient portlet entities
[16:05:49 EDT(-0400)] <EricDalquist> I'm running into a problem when trying to do the data export though
[16:06:14 EDT(-0400)] <EricDalquist> because I can only reverse engineer their subscribe ID by creating a IUserLayoutManager
[16:06:23 EDT(-0400)] <EricDalquist> which is a pita to do outside of a running portal
[16:09:59 EDT(-0400)] <awills> these are portlet preferences?
[16:10:12 EDT(-0400)] <EricDalquist> yeah ... part of it is my brain is a bit friend
[16:10:17 EDT(-0400)] <EricDalquist> so I need to think about this a bit more
[16:10:53 EDT(-0400)] <EricDalquist> so the problem is: how do we export portlet preferences for transient portlets
[16:11:17 EDT(-0400)] <EricDalquist> they way they are stored in the database is if you have a transient portlet entity it gets an ID of 'ctf.<portlet def id>'
[16:12:12 EDT(-0400)] <EricDalquist> so that say a transient instance of the weather portlet always comes up with the same preferences data for you
[16:12:17 EDT(-0400)] <EricDalquist> since it would always have the same definition id
[16:12:27 EDT(-0400)] <EricDalquist> so I guess to export it I'd have to user a different identifier
[16:15:26 EDT(-0400)] <EricDalquist> it wouldn't be a layout:xpath
[16:15:45 EDT(-0400)] <EricDalquist> it would be ... a marker that it is transient & a fname
[16:16:07 EDT(-0400)] <EricDalquist> so that on import the correct ctf.<portdefid> could be determined
[16:16:09 EDT(-0400)] <EricDalquist> :/
[16:25:32 EDT(-0400)] <athena> EricDalquist: I don't know if you might have a little advice for mavenifying the deploy-war task
[16:25:48 EDT(-0400)] <athena> for deploy-ear i basically just re-used the java code you'd written and called it from the new Mojo
[16:26:13 EDT(-0400)] <athena> but i'm not quite sure where deploy-war needs to be headed, mostly because i don't actually know that much about the structure of a war file
[16:26:13 EDT(-0400)] <EricDalquist> probably going to have to write some code like was written for deploy-ear
[16:26:16 EDT(-0400)] <athena> yeah
[16:26:21 EDT(-0400)] <EricDalquist> it just needs to be extracted
[16:26:34 EDT(-0400)] <athena> so can i extract it from a war just like you've extracted it from the ear?
[16:26:42 EDT(-0400)] <athena> i wasn't sure if the structure was similar enough for that to work
[16:26:45 EDT(-0400)] <EricDalquist> ah
[16:26:51 EDT(-0400)] <EricDalquist> so no
[16:26:57 EDT(-0400)] <athena> (smile)
[16:27:01 EDT(-0400)] <EricDalquist> the ear file is extracted and the contents are placed in specific locations
[16:27:04 EDT(-0400)] <EricDalquist> the WAR file is just a ZIP
[16:27:14 EDT(-0400)] <EricDalquist> just extract it into container/webapps/
[16:27:14 EDT(-0400)] <athena> ok, that's what i thought
[16:27:16 EDT(-0400)] <athena> yeah
[16:27:25 EDT(-0400)] <EricDalquist> there are some tricky bits with it though
[16:27:33 EDT(-0400)] <EricDalquist> primarily around the fastWarDeploy stuff
[16:27:37 EDT(-0400)] <athena> i was going to ask if you thought it was sufficient to just use a zip utility or whether there are other things we need to consider
[16:27:47 EDT(-0400)] <EricDalquist> the reason that is there is make live-deployes/redeployes less painful
[16:27:53 EDT(-0400)] <athena> yeah i saw that your current code seems to unzip to a temp directory as long as it's not windows
[16:27:59 EDT(-0400)] <athena> which i think we could replicate with maven
[16:28:14 EDT(-0400)] <EricDalquist> if you just unzip it directly into tomcat, tomcat will start loading the webapp before the unzip is done
[16:28:19 EDT(-0400)] <EricDalquist> which is bad
[16:28:31 EDT(-0400)] <athena> ahhh, i see
[16:28:34 EDT(-0400)] <athena> yes, that is bad (smile)
[16:28:45 EDT(-0400)] <EricDalquist> so the trick is extract to a temp dir, then delete the existing webapp then move the new one in
[16:28:55 EDT(-0400)] <EricDalquist> except on windows, because windows file locking results in exceptions
[16:29:08 EDT(-0400)] <EricDalquist> you can overwrite a file that is open and locked
[16:29:11 EDT(-0400)] <EricDalquist> but you can move or delete it
[16:29:19 EDT(-0400)] <EricDalquist> which is really weird
[16:29:35 EDT(-0400)] <athena> that is weird
[16:29:45 EDT(-0400)] <EricDalquist> but that is windows
[16:30:40 EDT(-0400)] <athena> lol
[16:30:42 EDT(-0400)] <athena> yeah
[16:59:21 EDT(-0400)] * michelled (n=michelle@CPE001310472ade-CM0011aefd3ca8.cpe.net.cable.rogers.com) has joined ##uportal
[17:03:34 EDT(-0400)] <EricDalquist> hrm
[17:03:45 EDT(-0400)] <EricDalquist> apparently all of our users have the header channel in their layouts
[17:03:51 EDT(-0400)] <EricDalquist> but it wasn't rendering in 3.0
[17:03:54 EDT(-0400)] <EricDalquist> now it is in 3.1
[17:03:56 EDT(-0400)] <EricDalquist> weird
[17:09:32 EDT(-0400)] <athena> oh weird
[17:09:39 EDT(-0400)] <athena> do we not have anything in the header anymore?
[17:10:06 EDT(-0400)] <EricDalquist> not sure what you mean by that?
[17:10:58 EDT(-0400)] <awills> normally header & footer content only renders if there's XSL like this placing it on the page: <xsl:copy-of select="/channel[@fname = 'header']"/>
[17:11:19 EDT(-0400)] <EricDalquist> ah
[17:11:31 EDT(-0400)] * SusanB (i=user@susan-x200.its.yale.edu) has left ##uportal
[17:11:56 EDT(-0400)] <awills> so we could have 100 channels in the data that don't show up, if nothing's explicitly putting them in (wink)
[17:13:25 EDT(-0400)] <athena> are we just not using it now though? i mean, did we used to explicitly render it?
[17:14:00 EDT(-0400)] <EricDalquist> well no in 3.1 it appears to be rendering
[17:14:07 EDT(-0400)] <EricDalquist> the problem is UW didn't have it rendering previously
[17:14:24 EDT(-0400)] <EricDalquist> so all of a sudden we have extra content in the head
[17:14:28 EDT(-0400)] <EricDalquist> header*
[17:14:30 EDT(-0400)] <EricDalquist> which we don't want
[17:15:23 EDT(-0400)] <athena> ah
[17:16:53 EDT(-0400)] <awills> this all comes of not forking your theme of course (wink)
[17:17:02 EDT(-0400)] <awills> (jk)
[17:17:06 EDT(-0400)] <EricDalquist> (smile)
[17:19:03 EDT(-0400)] <athena> where does the temp directory get set?
[17:19:56 EDT(-0400)] <athena> sorry - specifically wondering about java.io.tmpdir in uportal build.xml
[17:20:07 EDT(-0400)] <EricDalquist> it gets provided by Java
[17:20:31 EDT(-0400)] <EricDalquist> java.io.tmpdir is a special system property that the JVM sets
[17:20:36 EDT(-0400)] * michelled (n=michelle@CPE001310472ade-CM0011aefd3ca8.cpe.net.cable.rogers.com) has joined ##uportal
[17:20:37 EDT(-0400)] <athena> oh!
[17:20:39 EDT(-0400)] <athena> right.
[17:20:40 EDT(-0400)] <EricDalquist> I think you can override it
[17:20:54 EDT(-0400)] <EricDalquist> by passing your own value in for that system property
[17:21:07 EDT(-0400)] <athena> no really i'm just trying to mimic it in the maven plugin
[17:21:15 EDT(-0400)] <EricDalquist> ah yeah
[17:21:25 EDT(-0400)] <EricDalquist> System.getProperty("java.io.tmpdir")
[17:21:26 EDT(-0400)] <athena> presumably there's a way to get the same thing in the mojo
[17:21:30 EDT(-0400)] <athena> like that (smile)
[17:22:52 EDT(-0400)] * lennard1 (n=sparhk@wsip-98-174-242-39.ph.ph.cox.net) has joined ##uportal
[17:27:49 EDT(-0400)] <EricDalquist> duh ... thanks awills
[17:27:58 EDT(-0400)] <EricDalquist> I had messed up part of the merge
[17:28:05 EDT(-0400)] <EricDalquist> and left in:
[17:28:06 EDT(-0400)] <EricDalquist> <xsl:call-template name="login.channel"/>
[17:39:59 EDT(-0400)] <awills> glad to help (smile)
[17:41:36 EDT(-0400)] <holdorph> so... anyone tried a fresh clean, 'ant initportal' using maven 2.1.0 underneath and the uPortal 3.1.0 source distribution?
[17:42:55 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[17:42:56 EDT(-0400)] <holdorph> nm, i think there is tomcat 6 instructions I forgot
[17:46:56 EDT(-0400)] <EricDalquist> uhg ... I wonder how hard adding CSS namespacing to the webproxy portlet would be
[17:47:13 EDT(-0400)] <holdorph> fwiw, it seems to have built and run, at least one time with maven 2.1.0
[17:48:26 EDT(-0400)] <EricDalquist> holdorph: so are you running into problems with 2.1?
[17:48:47 EDT(-0400)] <holdorph> no, i forgot to edit the two tomcat files you have to touch in tomcat 6
[17:48:51 EDT(-0400)] <EricDalquist> ah
[17:48:58 EDT(-0400)] <holdorph> i forgot because I've only done it once or twice.
[17:49:14 EDT(-0400)] <holdorph> you didn't have to edit tomcat files in 5.5, so it's not ingrained in me yet
[17:49:22 EDT(-0400)] <holdorph> i'm sure after 1-2 more classes it will be
[18:53:51 EDT(-0400)] * lennard1 (n=sparhk@wsip-98-174-242-39.ph.ph.cox.net) has left ##uportal
[18:59:40 EDT(-0400)] * lennard1 (n=sparhk@wsip-98-174-242-39.ph.ph.cox.net) has joined ##uportal
[19:13:03 EDT(-0400)] <holdorph> man there is definitely a display bug in channel manager in up 3.1.0 theme
[20:14:30 EDT(-0400)] * michelled (n=michelle@CPE001310472ade-CM0011aefd3ca8.cpe.net.cable.rogers.com) has joined ##uportal
[20:25:40 EDT(-0400)] * lennard1 (n=sparhk@wsip-98-174-242-39.ph.ph.cox.net) has left ##uportal
[21:51:54 EDT(-0400)] * colinclark (n=colin@bas2-toronto09-1176131209.dsl.bell.ca) has joined ##uportal
[23:45:05 EDT(-0400)] * Sememmon (n=Sememmon@ip70-190-32-223.ph.ph.cox.net) has joined ##uportal
[23:57:26 EDT(-0400)] * Sememmon (n=Sememmon@ip70-190-32-223.ph.ph.cox.net) has joined ##uportal