uPortal IRC Logs-2012-01-10
[08:40:23 CST(-0600)] * Topic is 'http://uportal.pastebin.com/Â - https://wiki.jasig.org/display/UPC/uportal+IRC+Logs' set by EricDalquist!~apollo@76.210.64.142 on 17:44:12 CDT(-0500)
[12:11:49 CST(-0600)] <dd> hello
[12:11:54 CST(-0600)] <EricDalquist1> hi
[12:12:19 CST(-0600)] <dd> i'm trying to include clearpass as a dependency
[12:12:33 CST(-0600)] <dd> right now it is listed as a dependency in my uportal-ear/pom.xml file
[12:12:54 CST(-0600)] <athena> what version of uportal are you using?
[12:12:59 CST(-0600)] <dd> when i build, it keeps putting it in shared/lib/clearpass... instead of webapps/uPortal/WEB-INF/lib
[12:13:04 CST(-0600)] <dd> 3/2/4
[12:13:07 CST(-0600)] <dd> 3.2.4
[12:13:40 CST(-0600)] <athena> it needs to be a dependency of uportal-war/pom.xml, rather than the ear
[12:14:51 CST(-0600)] <dd> what's the difference between uportal-impl, uportal-ear, and uportal-war?
[12:14:59 CST(-0600)] <dd> in terms of where to put dependencies
[12:15:31 CST(-0600)] <EricDalquist1> dependencies should go under uportal-impl
[12:15:47 CST(-0600)] <EricDalquist1> impl is all source & libraries
[12:15:52 CST(-0600)] <EricDalquist1> war is webui
[12:16:01 CST(-0600)] <athena> er, right, thanks eric
[12:16:05 CST(-0600)] <EricDalquist1> ear is a packaging of uPortal and bundled portlets
[12:16:34 CST(-0600)] <EricDalquist1> in 4.0 impl was moved into war
[12:16:36 CST(-0600)] <dd> so if i wanted to do the facebook portlet, it would go in ear
[12:17:28 CST(-0600)] <dd> but a maven overlay would be preferred?
[12:17:33 CST(-0600)] <EricDalquist1> yeah, you add an overlay
[12:17:39 CST(-0600)] <EricDalquist1> and then reference the overlay in uportal-ear
[12:19:33 CST(-0600)] <dd> ok
[12:20:01 CST(-0600)] <dd> one more, Bookmarks Portlet is complaining that it can't find mysql connector
[12:20:21 CST(-0600)] <dd> i see it in uportal/web-inf and weather and webproxy portlet
[12:20:52 CST(-0600)] <dd> can i make it shared or do i need to edit the bookmarks pom.xml to add as a dependency
[12:21:05 CST(-0600)] <EricDalquist1> add it as a bookmarks dep
[12:21:11 CST(-0600)] <EricDalquist1> though it should be getting pulled in by the overlay
[12:21:34 CST(-0600)] <athena> that may not really be configured that way in 3.2
[12:22:02 CST(-0600)] <dd> ok
[12:22:07 CST(-0600)] <dd> thanks, big help
[12:38:22 CST(-0600)] <EricDalquist1> someday I want to refactor uPortal into a bunch of little maven modules
[12:38:42 CST(-0600)] <EricDalquist1> actually have real API seperation internally
[12:41:49 CST(-0600)] <athena> that'd be kinda neat
[13:46:12 CST(-0600)] <Arvids> hey, Eric
[13:46:22 CST(-0600)] <EricDalquist1> hi
[13:46:36 CST(-0600)] <Arvids> saw your mail regarding database dialect detection in DAOs
[13:46:55 CST(-0600)] <Arvids> seems like only solution for me... but kinda... dirty
[13:46:58 CST(-0600)] <EricDalquist1> yeah
[13:47:02 CST(-0600)] <EricDalquist1> not sure what our alternative is
[13:47:10 CST(-0600)] <Arvids> on the other hand - hibernate uses that approach a lot
[13:47:12 CST(-0600)] <EricDalquist1> since Postgres absolutely requires an explicit TX to read clobs
[13:47:18 CST(-0600)] <Arvids> indeed
[13:47:35 CST(-0600)] <EricDalquist1> brb
[13:47:36 CST(-0600)] <Arvids> in fact i´m wondering why MSSQL doesn´t require it to be done in transaction
[13:47:57 CST(-0600)] <EricDalquist1> most db's don't
[13:48:02 CST(-0600)] <EricDalquist1> well they do
[13:48:06 CST(-0600)] <EricDalquist1> but they use an implicit tx
[13:48:19 CST(-0600)] <EricDalquist1> which is much nicer from a user point of view
[13:48:30 CST(-0600)] <EricDalquist1> the db knows that reading the row + clob is a two step opperation
[13:48:45 CST(-0600)] <EricDalquist1> and deals with the tx demarcation accordingly to give you a consistent read
[13:49:56 CST(-0600)] <Arvids> ok...
[13:50:14 CST(-0600)] <EricDalquist1> I was concerned about those read-only tx entries
[13:50:21 CST(-0600)] <EricDalquist1> we had something like that before and hit performance issues
[13:52:10 CST(-0600)] <Arvids> sounds like all those places must be reviewed
[13:52:24 CST(-0600)] <EricDalquist1> well the only places they exist right now is around reads of clobs
[13:53:00 CST(-0600)] <EricDalquist1> the other thing I'd like help looking into is why TEXT is failing for mssql
[13:53:07 CST(-0600)] <EricDalquist1> er varchar(max)
[13:53:14 CST(-0600)] <EricDalquist1> since from what I was reading TEXT is deprecated
[13:53:41 CST(-0600)] <dd> hello again
[13:53:50 CST(-0600)] <dd> i'm having problems because of this bug: https://issues.jasig.org/browse/UP-2881?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel#issue-tabs
[13:54:13 CST(-0600)] <dd> is there any way around it to toggle aggregation on/off?
[13:56:41 CST(-0600)] <Arvids> Isn´t that issue fiexed, already?
[13:56:53 CST(-0600)] <dd> it is in 3.2.5, but i am on 3.2.4
[13:57:54 CST(-0600)] <Arvids> i´m not familiar with that feature, but i believe, the right choice would be to patch your portal
[13:58:45 CST(-0600)] <dd> any way to view what changes were made to the relevant files?
[13:59:13 CST(-0600)] <Arvids> ...trying to find them
[13:59:46 CST(-0600)] <Arvids> https://github.com/Jasig/uPortal/commit/f8145b40549d39ca97449ca3e8d3d887066bb417
[14:00:00 CST(-0600)] <Arvids> seems like fix is not that hard at all
[14:00:46 CST(-0600)] <Arvids> btw... is there a way to download a diff file from github?
[14:01:35 CST(-0600)] <dd> nice, thank you
[14:01:46 CST(-0600)] <dd> don't know, i've wondered that too
[14:02:09 CST(-0600)] <Arvids> it would be a nice feature
[14:02:26 CST(-0600)] <EricDalquist1> there is
[14:02:29 CST(-0600)] <EricDalquist1> but it is hidden
[14:02:33 CST(-0600)] <EricDalquist1> add .patch to the commit url
[14:02:34 CST(-0600)] <EricDalquist1> https://github.com/Jasig/uPortal/commit/f8145b40549d39ca97449ca3e8d3d887066bb417.patch
[14:05:11 CST(-0600)] <Arvids> nice catch
[14:09:10 CST(-0600)] <EricDalquist1> just some google-fu
[14:10:08 CST(-0600)] <Arvids> EricDalquist, will you have some time tomorrow to talk about possible performance improvements?
[14:10:25 CST(-0600)] <EricDalquist1> yes
[14:10:44 CST(-0600)] <athena> speeeeeeeed
[14:10:48 CST(-0600)] <EricDalquist1> though I won't be in the office until 10am
[14:10:55 CST(-0600)] <EricDalquist1> actually more like 10:30
[14:11:01 CST(-0600)] <EricDalquist1> I have a meeting across campus at 9am
[14:11:05 CST(-0600)] <EricDalquist1> not sure what time that is for you
[14:11:10 CST(-0600)] <EricDalquist1> as it is 2:10pm here
[14:11:18 CST(-0600)] <EricDalquist1> on Tuesday
[14:11:18 CST(-0600)] <Arvids> today I´ve done some initial research, but tomrorrow i´ll gather more data in order to make sure that the performance bottlenecks are consistent
[14:11:30 CST(-0600)] <EricDalquist1> did you hear back from YourKit at all?
[14:11:43 CST(-0600)] <Arvids> no
[14:11:51 CST(-0600)] <EricDalquist1> hrm, ok I'll poke them again
[14:13:16 CST(-0600)] <Arvids> i´m a little bit getting used to OSS tools... JVM monitor is simple yet does the right thing.
[14:13:39 CST(-0600)] <EricDalquist1> yeah … I'm just spoiled by the level of detail that yourkit provides
[14:13:44 CST(-0600)] <Arvids> not much, but at least gives some impression
[14:14:09 CST(-0600)] <Arvids> it´s nice if you have such opportunity
[14:15:03 CST(-0600)] <EricDalquist1> yeah, Jasig is supposed to have a few dozen OSS licenses for yourkit
[14:15:14 CST(-0600)] <EricDalquist1> to be used by project developers for improving Jasig projects
[14:15:23 CST(-0600)] <Arvids> hrmmm.. do you have any idea how many 4.0 portals are running in prodcution environment?
[14:15:32 CST(-0600)] <EricDalquist1> probably none right now
[14:15:36 CST(-0600)] <EricDalquist1> you're close
[14:15:40 CST(-0600)] <EricDalquist1> we're working on customizations right now
[14:15:45 CST(-0600)] <EricDalquist1> chicago is close
[14:15:52 CST(-0600)] <EricDalquist1> BYU just started customizations
[14:15:55 CST(-0600)] <Arvids> we´re rolling out on 01.02
[14:16:10 CST(-0600)] <EricDalquist1> exciting
[14:16:22 CST(-0600)] <EricDalquist1> you may well be the first
[14:16:23 CST(-0600)] <holdorph> i think drew wills is at oakland this week, i think he might be working on up 4 for them as well.
[14:16:24 CST(-0600)] <Arvids> didn´t think that we´re the ice-breakers
[14:16:36 CST(-0600)] <EricDalquist1> yeah, not by more than a month or two
[14:16:44 CST(-0600)] <Arvids> heh
[14:17:20 CST(-0600)] <Arvids> at least we´ll be the first ones in Europe
[14:17:31 CST(-0600)] <holdorph> whats that site that hosts free polls/surveys?
[14:18:16 CST(-0600)] <EricDalquist1> doodle?
[14:18:17 CST(-0600)] <holdorph> maybe we could send one out to the uportal-user mailing list. see what kind of response we get. Just a simple "which version of uPortal are you running in production?" with answers like "up 3.0.x"
[14:18:54 CST(-0600)] <holdorph> i think doodle is just for trying to set a meeting time
[14:19:15 CST(-0600)] <Arvids> google spreadsheets are great for simple polls
[14:19:49 CST(-0600)] <holdorph> i wonder if it's surveymonkey i'm thinking of
[14:20:59 CST(-0600)] <EricDalquist1> back on in a little bit folks, have to run home
[14:21:51 CST(-0600)] <athena> surveymonkey, yes
[14:21:57 CST(-0600)] <athena> that sounds like a useful thing
[14:28:48 CST(-0600)] <holdorph> hmm.... the free option at surveymonkey only supports 100 responses. While, i doubt we'd get that many, still seems like a frustrating limitation.
[14:30:38 CST(-0600)] <athena> that does indeed seem silly
[14:31:26 CST(-0600)] <holdorph> the first tier up from free, is a $204 annual subscription
[14:31:27 CST(-0600)] <athena> i think you can maybe make them using google docs?
[14:32:04 CST(-0600)] <holdorph> probably. i have no idea how easy or hard it is with google docs. but worth checking out, i suppose if it is quick
[14:32:19 CST(-0600)] <athena> looks plausible: http://www.online-tech-tips.com/google-softwaretips/how-to-create-an-online-survey-for-free-using-google-docs/
[14:32:27 CST(-0600)] <athena> i think that's how we're powering the current portal registration thing
[14:34:21 CST(-0600)] <holdorph> hmm... that guide is old, my version of a google doc spreadsheet doesn't look the same
[14:34:46 CST(-0600)] <holdorph> I don't have that second image "print, discuss, share, publish". style tabs.
[14:35:07 CST(-0600)] <holdorph> i have a share button, but clicking on that, doesn't give me the "to fill out a form" option
[14:40:19 CST(-0600)] <holdorph> ok, try this: https://docs.google.com/spreadsheet/viewform?formkey=dFBSVjQ5RFFrOTV4QmVPc0owOGU2S2c6MQ
[14:49:35 CST(-0600)] <holdorph> Arvids1: can you test that survey for me: https://docs.google.com/spreadsheet/viewform?formkey=dFBSVjQ5RFFrOTV4QmVPc0owOGU2S2c6MQ
[14:53:25 CST(-0600)] <Arvids2> ok, just a minute
[14:53:49 CST(-0600)] <Arvids2> ahh... very nice
[14:53:59 CST(-0600)] <holdorph> let me know when you've submitted
[14:54:12 CST(-0600)] <Arvids2> looks like VPN connecting/disconnecting spawns a lot of clones
[14:54:21 CST(-0600)] <Arvids2> submitted
[14:54:21 CST(-0600)] <holdorph> nod
[14:55:01 CST(-0600)] <holdorph> ok, nice. I didn't see any option for 'sharing' permissions, so wanted to make sure it was accessible to public, before posting the link to the mailing list.
[14:55:02 CST(-0600)] <athena> looks like a nice survey
[14:55:13 CST(-0600)] <athena> do we want to maybe ask what the institution is?
[14:55:35 CST(-0600)] <athena> my suggestion would be to add that
[14:55:47 CST(-0600)] <athena> and then maybe a checkbox to decide whether to allow us to list them as a uportal adopter
[14:55:48 CST(-0600)] <holdorph> i was going to just do one question, so people would be more likely to answer it.
[14:55:49 CST(-0600)] <Arvids2> +1 from me
[14:55:52 CST(-0600)] <athena> and then we can make a map of responses
[14:55:59 CST(-0600)] <athena> i know we don't want to add too much
[14:56:13 CST(-0600)] <holdorph> i'm worried there's some relationship to length of survey to number of responses
[14:56:18 CST(-0600)] <athena> but from a practical standpoint, it'll probably be more useful to those of us doing development and planning to have a sense of which schools are running which versions
[14:56:49 CST(-0600)] <athena> because then we can infer things about institution size, java expertise, etc. that can help us actually do something useful with those responses
[14:56:58 CST(-0600)] <Arvids2> Hmm... portal registration portlet serves for that purpose
[14:57:15 CST(-0600)] <athena> right - we don't need to ask all those questions all over again
[14:57:34 CST(-0600)] <athena> but if we know which institutions are planning to upgrade to uportal 4, we can have a sense of how big they are, etc. from information we already have
[14:57:51 CST(-0600)] <Arvids2> That reminds me that I should fill in that form
[14:58:04 CST(-0600)] <athena> and . . . that's the reason i suggested we maybe add a checkbox too
[14:58:10 CST(-0600)] <athena> then we could create a map, update the public data
[14:58:15 CST(-0600)] <athena> since i think we all know it's really out of date
[15:00:56 CST(-0600)] <Arvids2> btw, there are still two lists of deployments: https://wiki.jasig.org/display/UPC/Deployed+uPortals and http://www.jasig.org/uportal/deployments
[15:01:14 CST(-0600)] <athena> and we also have a google spreadsheet, which should probably replace both those things
[15:01:19 CST(-0600)] <athena> i'll poke the steering committee about that
[15:01:26 CST(-0600)] <holdorph> ok, sent. I know it would be nice to have the institution name, i wanted it as well. But I just worried too much that it might be just enough to cause someone to not fill it out. I rather like the anonymous nature of it, in it's current form.
[15:02:01 CST(-0600)] <athena> i guess - could have been optional
[15:02:32 CST(-0600)] <athena> as a developer i'm worried i won't be able to do much with the resulting data
[15:03:24 CST(-0600)] <holdorph> 'much' sure. i agree. but the idea only occurred to me, in terms of being able to answer Arvids question earlier.
[15:04:23 CST(-0600)] <holdorph> "is anyone else running uPortal 4". Sure I won't be able to say which organization is running it. but at least I'd be able to say "yes, about 4 schools are running it in production" or something like that
[15:05:48 CST(-0600)] <holdorph> wow, already got one response, besides Arvids. nice.
[15:06:04 CST(-0600)] <Arvids2> that was fast
[15:06:16 CST(-0600)] <holdorph> oooh, one more. up to 3. exciting.
[15:06:46 CST(-0600)] <holdorph> 4
[15:07:47 CST(-0600)] <Arvids2> ok, time to sleep. Thank you for your time. Good night I´m looking forward to see those results, because it could speed up some possible bug detection.
[15:08:10 CST(-0600)] <holdorph> g'night. later.
[15:11:24 CST(-0600)] <holdorph> 3 responses so far for up 3.1.x (current), 2 for up 3.2.x. All 5 responses plan to upgrade to up 4.0.x. 3 in jan-march, 1 in april-june, 1 in oct-dec
[15:32:12 CST(-0600)] <holdorph> 12 responses. now with 2 of the 12, saying their plans are upgrading from 3.2.x to 3.2.x. still have the 1 response with no plans to upgrade, the other 9 all plan to upgrade to up 4, this year. currently no one is running up 4 in production.
[16:14:14 CST(-0600)] <pandisaurus> hey I am trying to import some channels into uPortal 4 (the channels should be upgrading from 3.2) and I occasionally get an error (occasionally because I can run the import and get the error, then run it again without changing anything and sometimes no error occurs)
[16:14:34 CST(-0600)] <pandisaurus> the error is: Caused by: org.xml.sax.SAXParseException: cvc-pattern-valid: Value ' dailyUniverseMultimedia ' is not facet-valid with respect to pattern '[\w-_]+' for type 'fname-type'.
[16:14:48 CST(-0600)] <pandisaurus> the value is arbitrary, it changes randomly (that's the name of one of the channels)
[16:14:56 CST(-0600)] <pandisaurus> that is a portlet, in this case
[16:15:24 CST(-0600)] <EricDalquist1> hrm
[16:15:30 CST(-0600)] <EricDalquist1> not sure how that error would just go away
[16:15:51 CST(-0600)] <EricDalquist1> so you're saying that on try 1 the import fails on foo.channel
[16:15:59 CST(-0600)] <EricDalquist1> but then on try 2 foo.channel imports correctly?
[16:16:13 CST(-0600)] <pandisaurus> that is correct
[16:16:21 CST(-0600)] <pandisaurus> and bar.channel fails another time
[16:16:34 CST(-0600)] <EricDalquist1> that is very strange
[16:16:39 CST(-0600)] <pandisaurus> where foo.channel and bar.channel are completely arbitrary because it could be any of the channels
[16:16:43 CST(-0600)] <EricDalquist1> yeah
[16:16:48 CST(-0600)] <pandisaurus> and sometimes it just goes through all of them claiming successful
[16:17:25 CST(-0600)] <EricDalquist1> I'm just about to sign off for the day … could you create a jira issue at issues.jasig.org and paste in the full output of a failed run?
[16:17:31 CST(-0600)] <EricDalquist1> that would help in tracking down the problem
[16:17:40 CST(-0600)] <pandisaurus> I'll note, it says "INFO Importing 104 files of type <channel-definition script="classpath://org/jasig/portal/io/import-channel_v3-2.crn"> " then it does not say upgraded 104 times
[16:17:55 CST(-0600)] <EricDalquist1> ok, that sounds correct
[16:18:14 CST(-0600)] <EricDalquist1> even better if you can also attach the .channel file that failed in the example you post
[16:18:21 CST(-0600)] <EricDalquist1> I realize it is different channel files each time
[16:18:28 CST(-0600)] <EricDalquist1> but having the correlation for one failed run could help
[16:18:56 CST(-0600)] <pandisaurus> I can do that no problem, its just odd
[16:19:06 CST(-0600)] <EricDalquist1> I agree
[16:19:21 CST(-0600)] <pandisaurus> what determines whether it says 'upgraded' next to the channel or not?
[16:19:58 CST(-0600)] <pandisaurus> since I have you here
[16:20:35 CST(-0600)] <EricDalquist1> it should say it for all of them
[16:20:43 CST(-0600)] <EricDalquist1> since you are importing .channel files from 3.2
[16:20:57 CST(-0600)] <EricDalquist1> the XML file has to be transformed into the 4.0 format
[16:20:59 CST(-0600)] <EricDalquist1> and then imported
[16:21:10 CST(-0600)] <EricDalquist1> that transform step is what prints out the upgraded message
[16:21:54 CST(-0600)] <pandisaurus> I'm showing 24 "upgraded" messages, but it says it found 104 files to import (104 should be accurate) is there a way to see what files its skipping and why?
[16:22:08 CST(-0600)] <EricDalquist1> is that after a failure?
[16:22:16 CST(-0600)] <pandisaurus> no I'm counting when its successful
[16:22:19 CST(-0600)] <EricDalquist1> hrm
[16:22:23 CST(-0600)] <EricDalquist1> that sounds like a bug
[16:22:35 CST(-0600)] <EricDalquist1> you should be seeing either the same numbers all the way through
[16:22:40 CST(-0600)] <EricDalquist1> or it should fail with a reason
[16:22:49 CST(-0600)] <EricDalquist1> stuff shouldn't get silently ignored
[16:23:39 CST(-0600)] <pandisaurus> hmmm, that's troublesome because it is silently ignoring. "git checkout" just updates to the most current version of the trunk right? If I'm not specifying a branch?
[16:25:16 CST(-0600)] <EricDalquist1> sorry
[16:25:40 CST(-0600)] <EricDalquist1> "git checkout branch-name" checks out the specified branch
[16:25:41 CST(-0600)] <EricDalquist1> locally
[16:25:50 CST(-0600)] <EricDalquist1> to updates to the branch you are currently on
[16:25:57 CST(-0600)] <EricDalquist1> do "git pull"
[16:26:18 CST(-0600)] <EricDalquist1> which is a shortcut for doing:
[16:26:18 CST(-0600)] <EricDalquist1> git fetch
[16:26:19 CST(-0600)] <EricDalquist1> git merge origin/current-branch-name
[16:26:34 CST(-0600)] <EricDalquist1> you can determine your current branch via "git status"
[16:26:41 CST(-0600)] <EricDalquist1> have to run though