[23:58:43 EDT(-0400)] * esm (n=esm@esm.qis.net) has joined ##uportal <EricDalquist> <pom file="$ /pom.xml"/> <apetro_work_desk> NPE: a.getFoo() returned and was immediately dereferenced with .getBar()
[00:49:51 EDT(-0400)] * esm (n=esm@esm.qis.net) has joined ##uportal
[08:05:47 EDT(-0400)] * jayshao (n=jayshao@jayshao.rutgers.edu) has joined ##uportal
[08:21:49 EDT(-0400)] * esm (n=esm@esm.qis.net) has joined ##uportal
[09:30:11 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[10:01:02 EDT(-0400)] * esm (n=esm@asdf.dkc.jhu.edu) has joined ##uportal
[10:05:15 EDT(-0400)] * jayshao_ (n=jayshao@wg-206-hill012.rutgers.edu) has joined ##uportal
[10:10:06 EDT(-0400)] <esm> off to get some coffee then to play with portlet session funkyness
[10:10:07 EDT(-0400)] <esm> yay
[10:10:12 EDT(-0400)] <EricDalquist> 
[10:11:09 EDT(-0400)] <EricDalquist> Keep us up to date, I don't have cycles to go and help with direct debugging but I can sure make wild guesses (ok, semi-educated and I have tracked session bugs before so maybe I can help a bit there)
[10:14:00 EDT(-0400)] <esm> cool definitly will do
[10:16:22 EDT(-0400)] <EricDalquist> hey
[10:16:26 EDT(-0400)] <EricDalquist> can you do a simple check
[10:16:40 EDT(-0400)] <EricDalquist> can you downgrade to tomcat 5.5.20 and try to reproduce it?
[10:19:49 EDT(-0400)] <esm> not very easily - the server is shared and they are testing another app on it this week as well.
[10:20:03 EDT(-0400)] <EricDalquist> np
[10:20:16 EDT(-0400)] <esm> i may try to deploy it locally, it becomes difficult then, connecting to enterprise authentication, sso, etc.
[10:20:49 EDT(-0400)] <EricDalquist> is the portlet a simple test app or a real app?
[10:24:52 EDT(-0400)] <esm> a real app
[10:25:00 EDT(-0400)] <esm> i could put together a test one
[10:25:22 EDT(-0400)] <EricDalquist> I may actually do some testing today
[10:25:36 EDT(-0400)] <EricDalquist> we're in the process of moving from TC 5.5.20 to 5.5.23 so I have both environments set up
[10:25:48 EDT(-0400)] <EricDalquist> I'll see if I see anything in one env but not the other
[10:25:51 EDT(-0400)] <esm> i put a debug log in the jira issue
[10:25:53 EDT(-0400)] <esm> cool
[10:27:51 EDT(-0400)] <EricDalquist> I'm looking at the TC changelog and there was a change in session management from 20 to 23
[10:28:02 EDT(-0400)] <EricDalquist> btw if you want to have a fun read: http://issues.apache.org/bugzilla/show_bug.cgi?id=37356 Get down to the bits where Remy jumps in 
[10:28:08 EDT(-0400)] <esm> LOL
[10:31:46 EDT(-0400)] <esm> yeah
[10:33:03 EDT(-0400)] <esm> I can't reproduce on TC 5.5.23 with Pluto 1.1
[10:33:13 EDT(-0400)] <esm> so it may be some sort of interactiong
[10:37:13 EDT(-0400)] <EricDalquist> I'm pretty convinced we're only going to see it with pluto 1.0
[10:44:57 EDT(-0400)] <EricDalquist> hrm ... no luck reproducing the error with a simple session put/get portlet on our portal with tc 5.5.23
[10:45:53 EDT(-0400)] * pberry (n=pberry@waldorf.CSUChico.EDU) has joined ##uportal
[10:55:42 EDT(-0400)] <EricDalquist> oh boy ... ChannelPublisher is some interesting code
[10:55:58 EDT(-0400)] <EricDalquist> it loads File or File[] objects from the command line
[10:56:07 EDT(-0400)] <EricDalquist> then does a getName() on them to convert them to strings
[10:56:18 EDT(-0400)] <EricDalquist> then uses the ResourceLoader to get an InputStream
[10:56:24 EDT(-0400)] <EricDalquist> seems like a rather roundabout approach
[11:03:06 EDT(-0400)] <pberry> "the long way around"
[11:07:42 EDT(-0400)] <pberry> on a side note, Google Reader finally got a search feature.
[11:12:06 EDT(-0400)] * colinclark (n=atrcwrk2@142.150.154.101) has joined ##uportal
[11:28:33 EDT(-0400)] <EricDalquist> yay I think I have the chanpub tool working
[11:32:56 EDT(-0400)] <pberry> sweet
[11:33:15 EDT(-0400)] <EricDalquist> the next one is deploy
[11:33:21 EDT(-0400)] <EricDalquist> and then initportal in the mavenized version should work
[11:43:16 EDT(-0400)] <esm> ok lunchtime.
[11:43:58 EDT(-0400)] <esm> I think this is probably a pluto problem
[11:44:24 EDT(-0400)] <esm> and it may be an interaction when multiple portlets are doing session related things in the same render cycle
[11:44:31 EDT(-0400)] <EricDalquist> ah
[11:45:06 EDT(-0400)] <esm> i'm observing that attributes of the portlet session are not being updated properly, such as last accessed time.
[11:45:19 EDT(-0400)] <esm> and the portlet context is managed improperly as well
[11:47:51 EDT(-0400)] <EricDalquist> is this going to be a fairly easy thing to fix or a core problem with the 1.0 architecture
[12:03:30 EDT(-0400)] <EricDalquist> I really wish there was a way to install just the parent pom withouth doing the children
[12:07:50 EDT(-0400)] <EricDalquist> ahha!
[12:07:51 EDT(-0400)] <EricDalquist> <artifact:install>
[12:07:52 EDT(-0400)]
[12:07:52 EDT(-0400)] <EricDalquist> </artifact:install>
[12:08:18 EDT(-0400)] <EricDalquist> if I do that from the ant script it will install the parent pom without building the children :-D
[12:08:34 EDT(-0400)] * EricDalquist waits for esm to tell him why this is a very bad idea
[12:38:22 EDT(-0400)] * Topic is 'disowning children' set by apetro_work_desk on 2007-09-06 12:38:22 EDT(-0400)
[12:41:13 EDT(-0400)] <EricDalquist> you and those topics ...
[13:08:12 EDT(-0400)] * colinclar1 (n=atrcwrk2@142.150.154.101) has joined ##uportal
[13:11:00 EDT(-0400)] <esm> ah back
[13:11:24 EDT(-0400)] <esm> i hope not a core arch issue. i don't really know.
[13:11:40 EDT(-0400)] <EricDalquist>
[13:11:41 EDT(-0400)] <esm> as far as installing a pom only i wish i coudl do that too
[13:11:48 EDT(-0400)] <esm> so go for it
[13:12:07 EDT(-0400)] <EricDalquist> I think you're going to get a kick out of this new build.xml
[13:12:35 EDT(-0400)] * esm grins
[13:12:58 EDT(-0400)] <esm> i'll get a kick out of just about anything if i can get this session stuff resolved.
[13:13:03 EDT(-0400)] <EricDalquist> I bet :/
[13:13:18 EDT(-0400)] <EricDalquist> so it looks like the context isn't getting handled correctly?
[13:13:31 EDT(-0400)] <esm> at least, yes.
[13:14:00 EDT(-0400)] <esm> the PortletSession is backed by an HttpSession. I'm not sure if that association is being managed properly
[13:14:13 EDT(-0400)] <esm> I'm trying to figure out which reference isn't being managed correctly
[13:14:25 EDT(-0400)] <esm> that will at least point me to the right code to look at
[13:14:49 EDT(-0400)] <esm> right now i'm just tracing my way through the objects from the request back to the session back to the context/attributes
[13:15:14 EDT(-0400)] <EricDalquist> could sessions be getting swapped between portlets?
[13:15:46 EDT(-0400)] <esm> yeah exactly
[13:15:50 EDT(-0400)] <esm> that is what i'm wondering
[13:16:17 EDT(-0400)] <EricDalquist> not the easiest to do but would a test portlet (deployed multiple times with different webapps) that dumps the contents of the session help test that?
[13:16:55 EDT(-0400)] <EricDalquist> if each portlet sticks something unique in the session it should be fairly obvious what went where
[13:16:59 EDT(-0400)] <EricDalquist> if that is the problem
[13:17:55 EDT(-0400)] <esm> i have a portlet right now that dumps the session on every request, and after two or three renders I end up dumping a session that has anohter portlet's context object, and no session attributes.
[13:18:19 EDT(-0400)] <esm> so it appears that the session objects get swapped, or at least my portlet's session is being replaced.
[13:18:29 EDT(-0400)] <esm> the session id doesn't change.
[13:18:37 EDT(-0400)] <EricDalquist> should there be session attributes with from the other context?
[13:18:59 EDT(-0400)] <EricDalquist> the session id is the same for all webapps if you have the root session path set to /
[13:19:09 EDT(-0400)] <EricDalquist> which you likely do so portlet/servlet session sharing works
[13:19:40 EDT(-0400)] <esm> ah right
[13:19:57 EDT(-0400)] <esm> well, i'm continuing to dig
[13:21:01 EDT(-0400)] * jayshao (n=jayshao@jayshao.rutgers.edu) has joined ##uportal
[13:27:21 EDT(-0400)] <esm> EricDalquist: mvn -?
[13:27:23 EDT(-0400)] <esm> mvn -N
[13:27:53 EDT(-0400)] <EricDalquist> huh?
[13:28:05 EDT(-0400)] <EricDalquist> ahhh
[14:05:56 EDT(-0400)] <esm> i had a heck of a time compiling 1.0.1 on this laptop
[14:06:14 EDT(-0400)] <esm> maven 1 couldn't get the dependencies
[14:06:26 EDT(-0400)] <EricDalquist> no fun
[14:06:32 EDT(-0400)] <esm> so i ended up downloading a number of them by hand
[14:06:55 EDT(-0400)] <esm> the repo with the maven 1 artifacts sends a redirect, but the maven 1.0.2 doesn't follow...
[14:07:03 EDT(-0400)] <esm> find. -name
[14:07:03 EDT(-0400)] <esm> er
[14:21:18 EDT(-0400)] <esm> i love npe;s
[14:21:43 EDT(-0400)] * esm sings
[14:21:50 EDT(-0400)] <esm> there's null pointers like null pointers
[14:21:51 EDT(-0400)] <EricDalquist> lol
[14:22:04 EDT(-0400)] <EricDalquist> going that well huh?
[14:22:24 EDT(-0400)] <esm> heheh just forgot to protect it with an if/then
[14:22:29 EDT(-0400)] <esm> yeah
[14:22:32 EDT(-0400)] <esm> npe are the worst
[14:22:43 EDT(-0400)] <esm> followed up just behind that with ioe's with no init cause.
[14:23:11 EDT(-0400)] <EricDalquist> I really enjoy that sun decided not to include any usefull info in ClassCastExceptions
[14:23:26 EDT(-0400)] <esm> ah its a three-way heat then
[14:23:26 EDT(-0400)] <EricDalquist> would it have been that hard to say Class A failed to cast to Class B?
[14:23:27 EDT(-0400)] <esm> yeah
[14:23:53 EDT(-0400)] <esm> heh yeah
[14:24:11 EDT(-0400)] <esm> and to support public IOException ( String message, Throwable cause )
[14:24:14 EDT(-0400)] <esm> its hard
[14:25:09 EDT(-0400)] <EricDalquist>
[14:26:54 EDT(-0400)] <apetro_work_desk> yeah, I've occasionally wondered about that
[14:27:11 EDT(-0400)] <apetro_work_desk> like, maybe it would even be less performant, but it would be interesting to have some kind of debugging runtime environment in which
[14:27:23 EDT(-0400)] <esm> i think they've fixed ioe's constructor in java 6 i think.
[14:27:24 EDT(-0400)] <apetro_work_desk> ClassCastException just went ahead and knew what the illegal cast was
[14:27:25 EDT(-0400)] <esm> ...
[14:27:36 EDT(-0400)] <apetro_work_desk> even went ahead and quoted the offending line of source code
[14:27:37 EDT(-0400)] <EricDalquist> yeah, I could live with a JVM parameter
[14:27:38 EDT(-0400)] <apetro_work_desk> like, why not?
[14:28:22 EDT(-0400)] <apetro_work_desk> Even NullPointerException. At some sufficiently verbose debugging level (and actually, it seems fairly typical for this level of information to be retained in the generated class files), just go ahead and tell me what variable was null.
[14:28:57 EDT(-0400)] <EricDalquist> NPEs are why I hate code like 'a.getFoo().getBar().execute()'
[14:29:09 EDT(-0400)] <EricDalquist> (since we're on this rant track)
[14:29:35 EDT(-0400)] <apetro_work_desk> +1
[14:29:44 EDT(-0400)] <apetro_work_desk> but you'd hate that less if NPE told you what was null, right?
[14:30:08 EDT(-0400)]
[14:30:32 EDT(-0400)] <EricDalquist> yup
[14:30:32 EDT(-0400)] <apetro_work_desk> so, it's a pointless rant track, you're right
[14:30:37 EDT(-0400)] <apetro_work_desk> but it's also conceptually interesting
[14:30:49 EDT(-0400)] <apetro_work_desk> in that, everything needed to implement that is already there in the running JVM, isn't it?
[14:30:49 EDT(-0400)] <EricDalquist> that would be greate
[14:30:52 EDT(-0400)] <EricDalquist> great even
[14:54:48 EDT(-0400)] <jayshao> disowning children?
[14:55:03 EDT(-0400)] <jayshao> actually – I'm in a state of owning several of them myself...
[14:55:21 EDT(-0400)] * jayshao (n=jayshao@jayshao.rutgers.edu) has left ##uportal
[14:55:45 EDT(-0400)] * jayshao (n=jayshao@jayshao.rutgers.edu) has joined ##uportal
[14:58:15 EDT(-0400)] * jayshao (n=jayshao@jayshao.rutgers.edu) has joined ##uportal
[14:58:39 EDT(-0400)] <jayshao> pberry how goes the traffic?
[15:07:23 EDT(-0400)] <EricDalquist> hrm ... apparently my macrodef foo is a bit more than ant can handle
[15:19:57 EDT(-0400)] <esm> 586 [Thread-134]Returning portlet session [org.apache.pluto.core.impl.PortletSessionImpl@17dc299] id [44734573A6F25389578B7F60F13DC302.uPortalStage]
[15:20:10 EDT(-0400)] <esm> 599 [Thread-134]Returning portlet session [org.apache.pluto.core.impl.PortletSessionImpl@17dc299] id [44734573A6F25389578B7F60F13DC302.uPortalStage]
[15:21:05 EDT(-0400)] <esm> er woos
[15:21:07 EDT(-0400)] <esm> woops
[15:28:12 EDT(-0400)] <EricDalquist> hrm ... looks like my macrodef foo works in ant 1.6 but not 1.7
[15:28:16 EDT(-0400)] <EricDalquist> off to bugzilla!
[16:08:30 EDT(-0400)] <EricDalquist> so esm if you want a break from looking at portlet sessions: http://developer.ja-sig.org/source/browse/jasigsvn/up2/branches/working-maven2/build.xml?r=42416
[16:27:58 EDT(-0400)] <pberry> jayshao: it's calmed down a bit. We don't hit 1k concurrent until 11am
[16:28:33 EDT(-0400)] <pberry> eventually we'll fall into a stable pattern where we max out at 800 concurrent from 2pm-4pm with another "spike" around 6.
[18:29:55 EDT(-0400)] * colinclark (n=atrcwrk2@142.150.154.101) has joined ##uportal
[18:54:09 EDT(-0400)] * esm (n=esm@esm.qis.net) has joined ##uportal
[19:31:36 EDT(-0400)] * colinclark (n=atrcwrk2@142.150.154.101) has joined ##uportal
[20:02:12 EDT(-0400)] * colinclark (n=atrcwrk2@142.150.154.101) has left ##uportal
[21:58:55 EDT(-0400)] * colinclark (n=atrcwrk2@bas5-toronto63-1096706696.dsl.bell.ca) has joined ##uportal
[21:59:04 EDT(-0400)] * colinclark (n=atrcwrk2@bas5-toronto63-1096706696.dsl.bell.ca) has left ##uportal
[23:21:57 EDT(-0400)] * colinclark (n=atrcwrk2@bas5-toronto63-1096706696.dsl.bell.ca) has joined ##uportal
[23:22:06 EDT(-0400)] * colinclark (n=atrcwrk2@bas5-toronto63-1096706696.dsl.bell.ca) has left ##uportal
Unknown macro: {basedir}
Unknown macro: {null}
General
Content
Integrations