Versions Compared

Version Old Version 89 New Version 90
Changes made by

Apereo Infrastructure

Apereo Infrastructure

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Migrated to Confluence 4.0

[04:33:19 CST(-0600)] <rickard> Hmm. It appears uPortal ignores anything I put into UP_CHANNEL_MDATA.
[04:33:38 CST(-0600)] <rickard> I'm trying to localize channel names.
[04:34:18 CST(-0600)] <rickard> I've read online that CHAN_NAME needs to match CHAN_NAME from UP_CHANNEL and it does, but I still can't get it to work.
[04:50:43 CST(-0600)] <rickard> athena: You there?
[09:35:49 CST(-0600)] <athena> hi rickard
[09:35:57 CST(-0600)] <athena> i was not in fact here (smile)
[09:36:03 CST(-0600)] * athena lives on the west coast of the US
[09:40:11 CST(-0600)] <rickard> athena: That explains it (big grin)
[09:40:17 CST(-0600)] <athena> lol yes
[09:40:39 CST(-0600)] <rickard> I don't want to bug you, but did you by any chance see my post on the mailing list regarding the up_channel_mdata?
[09:40:52 CST(-0600)] <rickard> You seem to know your way around the code.
[09:41:17 CST(-0600)] <athena> looking
[09:41:56 CST(-0600)] <athena> i'm not sure the code to use the MDATA stuff exists/works anymore :/
[09:42:05 CST(-0600)] <athena> to be perfectly honest
[09:42:13 CST(-0600)] <rickard> Ok
[09:42:28 CST(-0600)] <athena> we've been talking about updating / rewriting all of that for the upcoming release of uportal, which should have a lot of i18n fixes
[09:42:29 CST(-0600)] <rickard> If that's true, that would explain why it doesn't get picked up (big grin)
[09:42:52 CST(-0600)] <athena> yeah, i have a feeling it may have gotten lost somewhere
[09:42:54 CST(-0600)] <rickard> Yeah, I've read some discussions on the subject, but I thought that was for 3.3.
[09:43:14 CST(-0600)] <athena> you might be able to fix it by making theme/structure modifications
[09:43:20 CST(-0600)] <athena> yeah, all those fixes will be for 3.3
[09:43:53 CST(-0600)] <rickard> Yeah, that's what we do today. We add it to messages.xml and then do a lookup on the channel name before we output it.
[09:44:03 CST(-0600)] <athena> gotcha
[09:44:11 CST(-0600)] <athena> if you look at channeldefinitionimpl there are messages like getName(String locale)
[09:44:13 CST(-0600)] <rickard> But it's not the most elegant solution in that it requires us to modify the theme whenever we make a change.
[09:44:26 CST(-0600)] <athena> but i'm not sure if the layout ever actually gets those locales
[09:44:44 CST(-0600)] <athena> eric might know a bit more
[09:45:44 CST(-0600)] <rickard> Ok
[09:46:45 CST(-0600)] <rickard> There is a locale attribute on channel in the theme xml, but for some reason, it's set to en_US on some channels and sv_SE on others. Regardless, the channel titles and stuff are still in English.
[09:47:11 CST(-0600)] <rickard> I'll see if Eric is here tomorrow. Maybe he knows something.
[09:47:20 CST(-0600)] <athena> yeah, asking him wouldn't be a bad idea
[09:47:33 CST(-0600)] <rickard> How is 3.2.5 coming along?
[09:47:41 CST(-0600)] <athena> also a question for eric (smile)
[09:47:45 CST(-0600)] <rickard> (big grin)
[09:47:57 CST(-0600)] <athena> i know there are some good fixes out there, but i'm not quite sure what else might be planned for that release
[09:48:02 CST(-0600)] <rickard> There's a pretty big list of unresolved bugs for 3.2.5 in Jira. Can't imagine all of them will be closed before release.
[09:48:09 CST(-0600)] <athena> agreed
[09:48:22 CST(-0600)] <athena> and probably some of them aren't very high priority
[09:48:27 CST(-0600)] <rickard> True
[09:49:10 CST(-0600)] <rickard> Another issue with i18n is that the locale switcher doesn't seem to work for guests.
[09:49:17 CST(-0600)] <athena> yes, that is absolutely true
[09:49:27 CST(-0600)] <rickard> On the list for 3.3? (smile)
[09:49:29 CST(-0600)] <athena> yes
[09:49:42 CST(-0600)] <athena> should be easier to fix with the rendering pipeline changes there
[09:49:46 CST(-0600)] <rickard> Ok
[09:49:51 CST(-0600)] <rickard> I won't bug you anymore. Thanks.
[09:50:00 CST(-0600)] <athena> i'd really like uportal's i18n support to be more solid as of 3.3 - i imagine it's kind of frustrating right now
[09:50:04 CST(-0600)] <athena> no problem (smile)
[09:50:10 CST(-0600)] <rickard> Yeah, it is.
[09:52:03 CST(-0600)] <rickard> The documentation on the subject is also a bit dated. I found references to stuff from like 2002 that surely no longer applies.
[09:52:55 CST(-0600)] <athena> well, we do have a documentation coordinator now, so if you find documentation that needs to be fixed in the manual, you might let her know
[09:53:13 CST(-0600)] <athena> all the current efforts are taking place at https://wiki.jasig.org/display/UPM32/Home
[09:53:35 CST(-0600)] <rickard> Yup
[09:53:44 CST(-0600)] <rickard> I've been working with this: https://wiki.jasig.org/display/UPC/Minimal+guide+for+Internationalization
[09:53:56 CST(-0600)] <rickard> But I guess it no longer applies.
[09:54:40 CST(-0600)] <athena> i guess step 1 does, but from what you're saying it sounds like step 2 doesn't work
[09:54:44 CST(-0600)] <athena> wonder how hard it would be to fix
[09:55:14 CST(-0600)] <rickard> Thing is, even if that is fixed, tabs still need to be localized in the theme.
[09:55:20 CST(-0600)] <athena> i know
[09:55:30 CST(-0600)] <athena> and i don't know if we'll be able to get that done for 3.3
[09:55:53 CST(-0600)] <athena> here's the wish list / planning page we have set up at the moment: https://wiki.jasig.org/display/UPC/Improving+uPortal+i18n+Support
[09:57:00 CST(-0600)] <rickard> Yeah, I saw that today.
[09:57:02 CST(-0600)] <rickard> "i18n support for tab names"
[09:57:53 CST(-0600)] <rickard> The up_channel_mdata solution seems ok to me. Wouldn't something similar be doable for tabs?
[09:58:06 CST(-0600)] <rickard> Or are tabs not store in the database like channels?
[09:58:11 CST(-0600)] <rickard> stored even
[09:58:35 CST(-0600)] <athena> they are not, unfortunately
[09:58:59 CST(-0600)] <athena> hey EricDalquist
[09:59:04 CST(-0600)] <EricDalquist> hello
[09:59:06 CST(-0600)] <rickard> howdy
[10:06:03 CST(-0600)] <rickard> EricDalquist: Did you by any chance see my post on the mailing list regarding up_channel_mdata?
[10:06:22 CST(-0600)] <rickard> athena tells me that he isn't certain if the mdata stuff exists/works anymore.
[10:06:29 CST(-0600)] <EricDalquist> not yet, on a phone call for the next hour...
[10:06:38 CST(-0600)] <rickard> Ok. No rush.
[10:06:49 CST(-0600)] <rickard> I'm leaving for the day in a few minutes anyway.
[11:23:51 CST(-0600)] <EricDalquist> athena: I have some ideas of what is happening with the mobile view
[11:24:00 CST(-0600)] <athena> oooh
[11:24:18 CST(-0600)] <EricDalquist> it appears to be initiating rendering of the portlets when just viewing the list of portlets
[11:24:34 CST(-0600)] <athena> interesting
[11:24:53 CST(-0600)] <EricDalquist> but since there are never incorperated into the post-theme content the portal doesn't wait for them to finish
[11:24:57 CST(-0600)] <EricDalquist> so we see all sorts of errors
[11:25:18 CST(-0600)] <EricDalquist> does the mobile structure XSL write them out as <channel> elements?
[11:25:43 CST(-0600)] <EricDalquist> yeah it does
[11:25:51 CST(-0600)] <athena> i honestly have no idea - i think matt wrote most fo the structure
[11:25:53 CST(-0600)] <athena> ah
[11:26:01 CST(-0600)] <athena> interesting
[11:26:01 CST(-0600)] <EricDalquist> any <channel> element written out by the structure will result in that channel being rendered
[11:26:09 CST(-0600)] <athena> i did not know thta
[11:26:11 CST(-0600)] <EricDalquist> that happens async
[11:26:12 CST(-0600)] <athena> makes sense though
[11:26:15 CST(-0600)] <EricDalquist> they get started
[11:26:20 CST(-0600)] <athena> so hopefully this is something we can fix pretty easily?
[11:26:25 CST(-0600)] <EricDalquist> well
[11:26:33 CST(-0600)] <EricDalquist> we'll have to modify the mobile structure & theme
[11:26:37 CST(-0600)] <athena> yeah
[11:26:46 CST(-0600)] <athena> only include the channel when in focus mode, basically?
[11:26:48 CST(-0600)] <EricDalquist> so in the channel list view we're writing out something like <channel-link>
[11:26:53 CST(-0600)] <athena> then have a way to have the titles in, but not the channels
[11:26:54 CST(-0600)] <EricDalquist> yup
[11:27:08 CST(-0600)] <EricDalquist> brb
[11:27:10 CST(-0600)] <athena> ok
[11:31:08 CST(-0600)] <EricDalquist> yeah so I'll take a crack at that this morning
[11:47:19 CST(-0600)] <awills> oh hello EricDalquist, good to see you
[11:48:08 CST(-0600)] <awills> would be happy to discuss header/footer fragment content in real time
[12:26:26 CST(-0600)] <EricDalquist> heh, yeah this mobile XSL issue should make the mobile view a lot faster too
[12:26:35 CST(-0600)] <EricDalquist> it was rendering every portlet in a user's layout on every request (tongue)
[12:26:41 CST(-0600)] <EricDalquist> granted it wasn't waiting for them
[12:29:33 CST(-0600)] <athena> eep!
[12:29:38 CST(-0600)] <athena> really glad you tracked that down (smile)
[12:29:40 CST(-0600)] <athena> i would not have found that
[12:29:53 CST(-0600)] <athena> awills: you following the mobile issue?
[12:30:09 CST(-0600)] <EricDalquist> yeah so the rendering pipline (in all versions) will start rendering every <channel> output by the structure
[12:30:42 CST(-0600)] <EricDalquist> then incorporate the output of <channel> output by the theme XSL
[12:30:51 CST(-0600)] <athena> gotcha
[12:31:03 CST(-0600)] <EricDalquist> the problem in the mobile ui is it is writing out <channel> elements just for navigation
[12:31:20 CST(-0600)] <EricDalquist> that is why the flyouts use <tabChannel> and not <channel>
[12:31:22 CST(-0600)] <EricDalquist> to avoid this issue
[12:31:27 CST(-0600)] <athena> makes sense
[12:31:34 CST(-0600)] <athena> seems like we can fix that pretty easily
[12:31:39 CST(-0600)] <EricDalquist> yeah
[12:31:42 CST(-0600)] <EricDalquist> working on that right now
[12:31:49 CST(-0600)] <EricDalquist> should only be a few places that need fixing
[12:33:43 CST(-0600)] <EricDalquist> dangit, I always forget to turn off the iphone useragent in firefox (tongue)
[12:33:51 CST(-0600)] <EricDalquist> and then get confused when I go to other sites
[12:34:14 CST(-0600)] <athena> lol
[12:34:16 CST(-0600)] <athena> me too
[12:38:05 CST(-0600)] <awills> athena catching up
[12:38:30 CST(-0600)] <awills> sounds like a good catch... important fix
[12:38:47 CST(-0600)] <athena> absolutely
[12:38:58 CST(-0600)] <athena> i'd determined yesterday that my previous patch didnt' fix the issue
[12:39:04 CST(-0600)] <athena> it's really hard to reliably trigger the bug
[12:40:13 CST(-0600)] <EricDalquist> well I still haven't seen that bug
[12:40:19 CST(-0600)] <EricDalquist> but I'm not going to dig more until this is fixed
[12:40:25 CST(-0600)] <EricDalquist> because it could very well be a side effect
[12:40:28 CST(-0600)] <athena> no that sounds like a reasonable candidate
[12:40:46 CST(-0600)] <athena> and should address that anyway
[12:45:21 CST(-0600)] <EricDalquist> not sure why but xslt/xpath makes me happy
[12:46:38 CST(-0600)] <athena> lol
[12:46:47 CST(-0600)] <athena> i think i'm less enthusiastic
[12:46:55 CST(-0600)] <athena> though as long as you're not doing something awful it's fine
[12:47:34 CST(-0600)] <athena> i think some of the work we did a few years ago with trying to render links for moving tabs based on permissions and priorities of the tabs before and after broke some small piece of my soul
[12:47:40 CST(-0600)] <awills> you might be in the minority on that one... though I can support the notion that XPath is satisfying
[12:47:45 CST(-0600)] <athena> it's yet to really forgive XSLT for that
[12:47:59 CST(-0600)] <athena> however using XPath type stuff in jQuery makes me happy
[12:47:59 CST(-0600)] <EricDalquist> I think it is nostalgia, my first programming job in college was 25% java 75% XSLT
[12:48:06 CST(-0600)] <EricDalquist> and we did some crazy stuff in XSL (tongue)
[12:48:08 CST(-0600)] <athena> ah (smile)
[13:01:41 CST(-0600)] <EricDalquist> trying the fix now ... … I just renamed channel to channel-nav for the navigation XML
[13:04:17 CST(-0600)] <EricDalquist> yay that works
[13:10:16 CST(-0600)] <EricDalquist> hrm
[13:10:21 CST(-0600)] <EricDalquist> another thing with the mobile UI I noticed
[13:10:35 CST(-0600)] <EricDalquist> when you hit back then refresh you get the portlet again
[13:16:20 CST(-0600)] <athena> oh hm, that's an interesting point
[13:16:32 CST(-0600)] <EricDalquist> since the URL is never changed
[13:16:35 CST(-0600)] <athena> guess we could update the current URL via js?
[13:16:38 CST(-0600)] <EricDalquist> right
[13:16:45 CST(-0600)] <EricDalquist> but that can cause a page reload I think
[13:16:49 CST(-0600)] <EricDalquist> you could do what gmail does
[13:16:49 CST(-0600)] <athena> ah
[13:16:53 CST(-0600)] <EricDalquist> and do it via anchors
[13:16:56 CST(-0600)] <athena> thought there was a way to update the URL without reloading
[13:17:03 CST(-0600)] <EricDalquist> maybe
[13:17:09 CST(-0600)] <EricDalquist> I'm not an expert with that stuff
[13:17:12 CST(-0600)] <athena> nor am i
[13:17:19 CST(-0600)] <EricDalquist> I'll file a jira for that when I'm doing the other fixes
[13:17:23 CST(-0600)] <athena> sounds great
[13:17:32 CST(-0600)] <athena> do you want me to look at updating the structure/theme?
[13:17:37 CST(-0600)] <athena> or are you already looking at that?
[13:18:13 CST(-0600)] <rickard> EricDalquist: Thanks for your responses on the mailing list. I think either I'm misunderstanding you or you misunderstood my question regarding hiding a channel when you click on the tab. What I'm after is essentially a channel that will only be displayed in focused mode.
[13:18:29 CST(-0600)] <EricDalquist> how does the user focus the channel?
[13:18:42 CST(-0600)] <rickard> They click on the link in the navigation.
[13:19:05 CST(-0600)] <EricDalquist> ok, so how does the link get there then?
[13:19:10 CST(-0600)] <EricDalquist> do you want that link hardcoded?
[13:19:20 CST(-0600)] <EricDalquist> or there dynamically because the channel is in the user's layout?
[13:20:51 CST(-0600)] <EricDalquist> athena: was there a jira for the mobile view issue?
[13:21:06 CST(-0600)] <athena> there was - we should reopen it
[13:21:19 CST(-0600)] <athena> let me find it
[13:22:33 CST(-0600)] <athena> https://issues.jasig.org/browse/UP-2857
[13:22:54 CST(-0600)] <athena> https://issues.jasig.org/browse/UP-2857
[13:22:58 CST(-0600)] <athena> er, woops
[13:22:59 CST(-0600)] <rickard> EricDalquist: Hmm. When you click on a tab, all the channels in that tab are rendered. However, you can also click on a channel in the navigation (fly-out or the "static" version) to focus that channel. What I want is for specific channels to not be rendered when you click on the tab. Instead, users will have to click on the link to the channel in the navigation.
[13:23:39 CST(-0600)] <EricDalquist> hrm, you could set a custom channel parameter on those channels
[13:23:51 CST(-0600)] <EricDalquist> and then in the structure XSL look for that parameter
[13:23:59 CST(-0600)] <rickard> Yeah, that's what I thought.
[13:24:08 CST(-0600)] <EricDalquist> and if it exists only write out the tabChannel entry but not the channel entry
[13:24:18 CST(-0600)] <rickard> But I'm not sure how to set custom parameters on channels.
[13:24:33 CST(-0600)] <rickard> I can't seem to find a way to do it in the portlet admin thingy.
[13:24:36 CST(-0600)] <EricDalquist> it can be done during publishing
[13:24:42 CST(-0600)] <rickard> I can only edit the existing parameters.
[13:25:35 CST(-0600)] <EricDalquist> thanks athena
[13:25:43 CST(-0600)] <athena> yep - reopened it
[13:25:58 CST(-0600)] <athena> EricDalquist: do you want me to try and make the change to not have channels in the structure?
[13:26:19 CST(-0600)] <EricDalquist> oh I have that done already
[13:26:25 CST(-0600)] <athena> terrific (smile)
[13:26:30 CST(-0600)] <EricDalquist> going to commit it in just a few minutes
[13:26:38 CST(-0600)] <athena> let me know and i can try and help test the fix
[13:26:49 CST(-0600)] <EricDalquist> athena: did the ability to set custom channel parameters go away in the new portlet manager?
[13:27:04 CST(-0600)] <athena> no i don't think so
[13:27:21 CST(-0600)] <EricDalquist> I'm lookin at the publishing new portlet UI
[13:27:29 CST(-0600)] <EricDalquist> and my only option is to add preferences
[13:27:35 CST(-0600)] <EricDalquist> I can't see a way to add parameters
[13:27:38 CST(-0600)] <athena> i think the CPD has to have the arbitrary parameter or arbitrary portlet section
[13:27:43 CST(-0600)] <EricDalquist> ah
[13:27:57 CST(-0600)] <athena> maybe we need to add that in for generic portlets
[13:28:02 CST(-0600)] <athena> or for everything?
[13:29:00 CST(-0600)] <EricDalquist> probably for everything
[13:29:25 CST(-0600)] <EricDalquist> especially if we're moving to re-purpose channel parameters as a way to customize the structure/theme transforms
[13:29:43 CST(-0600)] <EricDalquist> so rickard looks like you're right. It isn't there in the UI right now
[13:29:45 CST(-0600)] <EricDalquist> but it really should be
[13:31:49 CST(-0600)] <rickard> What do you call the view when a particular channel isn't focused? unfocused?
[13:31:58 CST(-0600)] <EricDalquist> normal
[13:32:07 CST(-0600)] <rickard> Ok.
[13:32:52 CST(-0600)] <rickard> In other words, I would be able to open up portlet manager and add a parameter to the Word of the Day channel (for example "hideFromNormalView") that I could then look for in my theme?
[13:33:09 CST(-0600)] <EricDalquist> yup
[13:33:16 CST(-0600)] <rickard> Ok, great.
[13:33:34 CST(-0600)] <rickard> But I can't expect this for 3.2.5, can I?
[13:33:44 CST(-0600)] <EricDalquist> yes you should
[13:33:53 CST(-0600)] <rickard> Evern better (smile)
[13:33:59 CST(-0600)] <rickard> even even
[13:34:14 CST(-0600)] <EricDalquist> we can both peer-pressure athena into helping me fix the CPDs so you can set custom parameters in the manager UI (smile)
[13:34:27 CST(-0600)] <rickard> (big grin)
[13:34:32 CST(-0600)] <athena> well, it shouldn't be hard to "fix" the CPD
[13:34:45 CST(-0600)] <athena> although i'm not actually sure that we want the arbitrary params bucket to show up by default
[13:34:55 CST(-0600)] <athena> it's very easy to add though, and it's also easy to add new parameters
[13:35:07 CST(-0600)] <EricDalquist> right
[13:35:12 CST(-0600)] <EricDalquist> I suppose that would be the other option
[13:35:14 CST(-0600)] <athena> maybe instead what you want is to add a "hideFromNormalView" parameter to the display parameters section of the shared cpd?
[13:35:21 CST(-0600)] <EricDalquist> yeah
[13:35:22 CST(-0600)] <athena> i think that would be a lot more user-friendly for your portal
[13:35:25 CST(-0600)] <athena> and it's really easy
[13:35:34 CST(-0600)] <EricDalquist> and then rickard you would just have a drop-down to choose when publishing
[13:35:41 CST(-0600)] <EricDalquist> similar to the show-in-mobile option
[13:35:46 CST(-0600)] <athena> you want uportal-war/src/main/resources/org/jasig/portal/channels/SharedParameters.cpd
[13:35:50 CST(-0600)] <athena> yes, exactly
[13:36:22 CST(-0600)] <rickard> Looking...
[13:37:14 CST(-0600)] <rickard> Looks promising (smile)
[13:39:54 CST(-0600)] <athena> that's how we're hiding channels from the mobile view
[13:40:22 CST(-0600)] <athena> though EricDalquist i guess we need to strip those out at the structure level, based on what you talked about earlier?
[13:40:50 CST(-0600)] <EricDalquist> that was already working
[13:41:06 CST(-0600)] <EricDalquist> it only copies channels over that dont have hideInMobile set
[13:43:09 CST(-0600)] <athena> oh good
[13:43:20 CST(-0600)] <athena> wasn't sure if that might be at the theme level
[13:43:38 CST(-0600)] <EricDalquist> this 'hidden' channel issue is also what was causing those errors from the locale selector
[13:43:47 CST(-0600)] <EricDalquist> it was listed in the header of a bunch of layouts
[13:43:54 CST(-0600)] <EricDalquist> which caused it to start rendering after the structure
[13:44:04 CST(-0600)] <EricDalquist> but it is never included in the rendered UI by the theme
[13:44:18 CST(-0600)] <EricDalquist> so it wouldn't finish rendering before the rest of the portlets
[13:44:27 CST(-0600)] <EricDalquist> and be rendering after the response was complete
[13:50:41 CST(-0600)] <EricDalquist> ok athena: https://developer.jasig.org/source/changelog/jasigsvn/uPortal?cs=22436
[13:50:48 CST(-0600)] <EricDalquist> just made that change in 3.2 and trunk
[13:50:55 CST(-0600)] <EricDalquist> I'll email the list and see if they can test it out again
[13:52:33 CST(-0600)] <athena> hurray!
[13:52:36 CST(-0600)] <athena> will test now
[13:53:05 CST(-0600)] <EricDalquist> great, thanks
[13:53:11 CST(-0600)] <EricDalquist> it does seem to render a lot faster too
[13:53:16 CST(-0600)] <EricDalquist> which makes sense
[13:53:23 CST(-0600)] <EricDalquist> that was a lot of work going on in the background
[13:53:27 CST(-0600)] <athena> yeah
[13:53:30 CST(-0600)] <athena> definitely makes sense
[13:53:34 CST(-0600)] <athena> glad you found that
[13:53:40 CST(-0600)] <EricDalquist> I think in trunk I'm going to add some logging about portlets that are started but never included
[13:53:48 CST(-0600)] <athena> yeah that's a great idea
[13:53:48 CST(-0600)] <EricDalquist> seems like a reasonable WARN log message
[13:57:53 CST(-0600)] <athena> yes
[14:03:33 CST(-0600)] <athena> EricDalquist: haven't been able to trip that error yet
[14:03:37 CST(-0600)] <athena> so looking good so far!
[14:03:43 CST(-0600)] <EricDalquist> yay!
[14:03:56 CST(-0600)] <EricDalquist> hopefully it really was a nice simple fix like that!
[14:04:02 CST(-0600)] <athena> yeah (smile)
[14:44:06 CST(-0600)] <rickard> athena: I have to apologize. I believe I referred to you as "he" earlier. I wasn't aware that you were "Jen". Sorry (smile)
[14:55:26 CST(-0600)] <athena> oh, no worries
[14:55:31 CST(-0600)] <EricDalquist> https://wiki.jasig.org/display/PLT/Weather+Portlet+Configuration
[14:55:37 CST(-0600)] <athena> i know my nick isn't terribly obvious (smile)
[14:55:46 CST(-0600)] <athena> thanks for updating that eric!
[14:55:54 CST(-0600)] <EricDalquist> athena: can you take a quick look at that page and see if it is missing anything?
[14:55:55 CST(-0600)] <EricDalquist> yup
[14:56:54 CST(-0600)] <athena> looks good to me
[14:57:03 CST(-0600)] <EricDalquist> great
[14:57:05 CST(-0600)] <athena> only note i have is that the WWO icons are now mapped back to the original icons
[14:57:12 CST(-0600)] <EricDalquist> ah ok
[14:57:16 CST(-0600)] <athena> so they're the same as what we used for accuweather
[14:57:23 CST(-0600)] <athena> which actually appear to have come from a firefox plugin
[14:57:29 CST(-0600)] <athena> we should investigate the license on those
[14:57:41 CST(-0600)] <EricDalquist> we sent WWO a request yesterday for a request limit increase
[14:57:42 CST(-0600)] <EricDalquist> we have a key
[14:57:44 CST(-0600)] <EricDalquist> and it is working
[14:57:51 CST(-0600)] <athena> terrific!
[14:57:54 CST(-0600)] <EricDalquist> but we'll go way over 500/hour
[14:57:56 CST(-0600)] <athena> so were you hitting the limit?
[14:57:59 CST(-0600)] <EricDalquist> so it isn't in prod yet
[14:58:05 CST(-0600)] <athena> yeah i don't know what the typical use profile for most schools is
[14:58:11 CST(-0600)] <athena> perhaps yahoo is better for some heavier usage
[14:58:21 CST(-0600)] <EricDalquist> our stats show we average around 6000 renders/hour during the day of the weather portlet
[14:58:34 CST(-0600)] <EricDalquist> yeah except we also contacted them about an API key
[14:58:39 CST(-0600)] <EricDalquist> and haven't heard back yet (tongue)
[14:58:56 CST(-0600)] <athena> ick
[14:59:06 CST(-0600)] <athena> how many requests / hour do you think you have w/ caching in place?
[14:59:15 CST(-0600)] <EricDalquist> I'm not really sure
[14:59:17 CST(-0600)] <EricDalquist> we have 4 servers
[14:59:29 CST(-0600)] <athena> ah - caching doesn't go between them?
[14:59:32 CST(-0600)] <EricDalquist> nope
[14:59:34 CST(-0600)] <EricDalquist> not by default
[14:59:39 CST(-0600)] <EricDalquist> and it is a lot of work to set that up
[14:59:48 CST(-0600)] <athena> ah, wasn' tsure if maybe you'd done that
[15:00:02 CST(-0600)] <EricDalquist> so we would be limited to 125 unique weather locations viewed/hour
[15:04:37 CST(-0600)] <EricDalquist> ok
[15:04:41 CST(-0600)] <EricDalquist> oops wrong window (tongue)
[15:14:12 CST(-0600)] <rickard> athena, EricDalquist: Modifying SharedParameters.cpd worked great. Thanks.
[15:14:19 CST(-0600)] <EricDalquist> yay!
[15:14:20 CST(-0600)] <athena> terrific
[15:16:27 CST(-0600)] <awills> EricDalquist the DLM non-permitted channel trimming works as we discussed... I can see all the behavior we expected: (1) channels you're not auth'ed to see aren't included in your layout DOM, and (2) several copies of things you are allowed to see do show up, if the data is there and the feature is enabled
[15:17:14 CST(-0600)] <awills> so i'm wondering if it's better to (1) turn on the feature by default in 3.3, and count on people cleaning up their data, or (2) making it an opt-in thing for 3.3 as well as 3.2
[15:18:21 CST(-0600)] <awills> I would like to have it on, all other things being equal... it's a feature I feel I can do a lot with
[15:32:50 CST(-0600)] <EricDalquist> sorry awills was on a phone call
[15:33:23 CST(-0600)] <EricDalquist> I could see making it an opt-in thing for 3.2
[15:33:26 CST(-0600)] <EricDalquist> and default in 3.3
[15:33:46 CST(-0600)] <awills> terrific, thanks
[15:33:48 CST(-0600)] <EricDalquist> I'm still wondering if it would be good to code in some de-duplication for header/footer content with a toggle for that as well
[15:34:13 CST(-0600)] <EricDalquist> since I think the use-case for more than 1 portlet with the same fname in the header or footer is probably very rare
[15:34:18 CST(-0600)] <awills> ah sure, could be a nice tweak
[15:34:36 CST(-0600)] <EricDalquist> have de-dupe enabled and have it log a WARN level message about doing the dedupe
[15:34:46 CST(-0600)] <EricDalquist> then people can either fix their data or disable it
[15:34:47 CST(-0600)] <awills> even a big WARNing would be very helpful
[15:34:50 CST(-0600)] <EricDalquist> depending on their needs
[16:52:02 CST(-0600)] <awills> anyone recognize this line as recent?: this.resourceCache.put(element); //do a cache put to notify the cache the object has been modified
[16:52:14 CST(-0600)] <EricDalquist> yup
[16:52:17 CST(-0600)] <EricDalquist> I added it
[16:52:34 CST(-0600)] <EricDalquist> modifying cache entries by reference is bad for wanting to use distributed caching
[16:52:56 CST(-0600)] <EricDalquist> since the cache replication code then has no way to know the entry was modified
[16:52:57 CST(-0600)] <awills> [artifact:mvn] Failed tests:
[16:52:57 CST(-0600)] <awills> [artifact:mvn] testCachedNotModified(org.jasig.portal.utils.cache.CachingResourceLoaderImplTest)
[16:53:04 CST(-0600)] <awills> Unexpected method call put
[16:53:15 CST(-0600)] <EricDalquist> it isn't that recent though
[16:53:24 CST(-0600)] <EricDalquist> probably added 6-9 months ago
[16:53:29 CST(-0600)] <awills> which? the test class?
[16:53:45 CST(-0600)] <EricDalquist> the line you asked about
[16:53:49 CST(-0600)] <awills> ah i follow
[16:53:52 CST(-0600)] <EricDalquist> though I am making some assumptions
[16:54:05 CST(-0600)] <EricDalquist> what version of uPortal and what class?
[16:54:43 CST(-0600)] <awills> trunk... CachingResourceLoaderImplTest
[16:55:05 CST(-0600)] <EricDalquist> does it work for you with a clean checkout?
[16:55:38 CST(-0600)] <awills> had just done svn update... was doing clean initportal
[16:55:56 CST(-0600)] <awills> i'll try again
[16:55:59 CST(-0600)] <EricDalquist> huh
[16:56:14 CST(-0600)] <EricDalquist> they pass locally and bamboo seems happy with the tests in trunk as well
[16:56:34 CST(-0600)] <EricDalquist> you're on Windows right? I wonder if the test fails on windows for some reason?
[16:56:35 CST(-0600)] <holdorph> windows problem?
[16:56:47 CST(-0600)] <EricDalquist> it uses temp files and plays with last-modified timestamps
[16:56:57 CST(-0600)] <awills> ok... could be a fluke
[16:56:59 CST(-0600)] <holdorph> seems like an odd one, given what awills pasted though
[16:57:15 CST(-0600)] <EricDalquist> well that code is for loading files and caching the result
[16:57:21 CST(-0600)] <EricDalquist> then auto-reloading them when they change
[16:57:36 CST(-0600)] <EricDalquist> and the unit test does that verification with temp files and by messing with the last modified values
[16:57:48 CST(-0600)] <EricDalquist> maybe windows doesn't honor the last modified values being set
[16:58:39 CST(-0600)] <awills> still wraping my brain around these moving pieces
[16:58:57 CST(-0600)] <awills> looks like put() gets called no matter what
[17:01:59 CST(-0600)] <EricDalquist> baxk in 15
[17:25:42 CST(-0600)] <EricDalquist> did you get that test figured out awills?
[17:29:25 CST(-0600)] <awills> i didn't... it continued to happen when i ran it again
[17:29:38 CST(-0600)] <EricDalquist> hrm
[17:30:03 CST(-0600)] <EricDalquist> I guess we should eventually see if we can get a bamboo builder to run our unit tests on windows for us too
[17:30:26 CST(-0600)] <awills> i don't see all the peices of the puzzle, but to me it seems there would have to be some sort of expect(cache.put...);;
[17:30:39 CST(-0600)] <awills> in the test
[17:31:07 CST(-0600)] <EricDalquist> could you email me the txt file from the failed unit test?
[17:31:13 CST(-0600)] <awills> sure
[17:34:33 CST(-0600)] <awills> done
[17:34:38 CST(-0600)] <EricDalquist> thanks