The IRC channel is #sakai on the freenode network, you can join using a web client at http://webchat.freenode.net/?channels=sakai and you can view the complete logs at http://sakai.iriscouch.com/irc/_design/viewer/index.html
10:19 <SakaiGithub> [3akai-ux/1.5.0] Merge origin/master and fix conflicts - Bert Pareyn 10:30 <PhysX> kentfitz ping 10:31 <kentfitz> physx: pong 10:31 <PhysX> kentfitz: I'm seeing a 500 error on a POST to an item's activity 10:31 <PhysX> on qa20 10:32 <PhysX> actually on every activity post I'm doing 10:32 <PhysX> kentfitz: I'm thinking it has something to do with the content not being mine, so no manager on it, and trying to write to the activity node 10:34 <PhysX> kentfitz: did you get my messages? 10:34 <kentfitz> physx: Lost my connection for a second. Can you give me an example of a specific activity? 10:34 <PhysX> kentfitz: sure, if you go to https://qa20-us.sakaiproject.org:8088/content#p=mpmimzHie/iPad-Test.txt 10:34 <PhysX> I shared this with you, but you don't manage it, Eli is the only manager 10:35 <PhysX> now comment/share it 10:35 <PhysX> the posts will return a 500 error 10:35 <kentfitz> I can confirm that, although I am able to post on other content I am not a manager of 10:36 <PhysX> kentfitz: hmm I can confirm that as well 10:37 <kentfitz> physx: Hmmmm, so what's different about this piece of content 10:37 <PhysX> kentfitz: nothing much on first sight, same permissions, not added to any groups, shared by same amount of people 10:40 <kentfitz> physx: it looks like iPad-Text was successfully shared with me btw, and our comments are being written. 10:41 <PhysX> kentfitz: yeah, only the activity isn't posted 10:42 <kentfitz> physx: That is curious for sure. I'd write it up 10:42 <PhysX> kentfitz: I'll create a ticket and look into it some more 10:42 <kentfitz> sounds good 10:46 <PhysX> kentfitz SAKIII-6048 10:46 <kentfitz> physx: Thanks 10:50 <PhysX> kentfitz: another thing I noticed on https://qa20-us.sakaiproject.org:8088/content#p=mpmimzHie/iPad-Test.txt is that the creation date is a day after the last modified date 10:50 <PhysX> and the created date is actually in the future 10:52 <kentfitz> created date shows as yesterday to me and the last modified as today to me 11:35 <stuartf> http://www.happyprog.com/tdgotchi/ 12:11 <GitHub33> [sparsemapcontent] arwhyte pushed 1 new commit to 1.3: http://git.io/C9WVOA 12:11 <GitHub33> [sparsemapcontent/1.3] [maven-release-plugin] prepare release org.sakaiproject.nakamura.core-1.3.6 - Anthony Whyte 12:11 <GitHub184> [sparsemapcontent] arwhyte pushed 1 new commit to 1.3: http://git.io/vHBf5g 12:11 <GitHub184> [sparsemapcontent/1.3] [maven-release-plugin] prepare for next development iteration - Anthony Whyte 13:38 <SakaiGithub> [3akai-ux/1.5.0] SAKIII-5540 Added new delete icon for My Message - Dean Angerer 13:48 <SakaiGithub> [3akai-ux] bp323 pushed 1 new commit to 1.5.0: http://git.io/hs7irg 13:48 <SakaiGithub> [3akai-ux/1.5.0] SAKIII-4976 remove ability to resize columns when exiting edit mode - James Sloane 14:05 <SakaiGithub> [3akai-ux] bp323 pushed 1 new commit to 1.5.0: http://git.io/4AMqNQ 14:05 <SakaiGithub> [3akai-ux/1.5.0] SAKIII-5482 - avoid jumping to the top in Google Chrome - Christian Vuerings 14:29 <SakaiGithub> [3akai-ux] bp323 pushed 1 new commit to 1.5.0: http://git.io/XUXkBA 14:29 <SakaiGithub> [3akai-ux/1.5.0] SAKIII-5386 some changes for accessibility - James Sloane 14:52 <thecarlhall> hey mrvisser 14:53 <MrVisser> hey thecarlhall 14:53 <thecarlhall> mrvisser, I was just taking a look through the perf4j PR. 14:53 <thecarlhall> am I correct in thinking that perf4j gives us timing results? 14:54 <MrVisser> thecarlhall: that's correct. take some time to see the screencast, and maybe a quick browse through the perf4j documentation, it should all be linked from the JIRA. 14:55 <thecarlhall> I ask because it looks like some things are being timed that should have near 0 return time like response.setStatus 14:55 <MrVisser> thecarlhall: ahh, yea, that is mostly for telemetry. 14:55 <thecarlhall> mrvisser, there are cases of measuring the speed of telemetry, too 14:55 <MrVisser> thecarlhall: you'll get a 0 for timing on that tag, but it's the counts you want, which perf4j provides as well. 14:56 <thecarlhall> mrvisser, so is this intended as an eventual replacement for the telemetry stuff we already have? Seems like there is some overlap. 14:56 <MrVisser> thecarlhall: if we get to the point where we can enable this in production, then yes, it can replace telemetry. 14:57 <MrVisser> thecarlhall: at the very least, we can enable it in performance tests to get timing results from suspect methods. 14:57 <thecarlhall> mrvisser, is this not enabled? where's the trigger? 14:57 <MrVisser> thecarlhall: it's a build profile -Pperf4j 14:59 <thecarlhall> so with that profile active, <guessing>it rewrites bytecode to weave the annotations into their processing otherwise they're innocuous? 14:59 <MrVisser> thecarlhall: correct. it uses aspectj 15:00 <MrVisser> thecarlhall: the perf4j jar can also be used as a java agent for on-the-fly weaving. 15:01 <MrVisser> but I don't think that would be necessary. I anticipate that there is negligible performance hit by enabling it, and that in a later release after some experience running it on bug bashes and performance tests, we can decide to enable it by default. 15:02 <MrVisser> thecarlhall: if enabled by default, I think we can do away with the telemetry bundle. 15:02 <thecarlhall> mrvisser, 2 things on that: 1) is it as fast as our current stuff and 2) can we dictate output (resmon format, etc) 15:04 <MrVisser> thecarlhall: 1) I don't know that it isn't as fast as our current stuff; 2) logs output the raw data that can be crunched, the GraphingAppender can be used to see the graphing details, and custom appenders could be used to output custom formats, if needed. 15:04 <thecarlhall> mrvisser, it definitely looks like perf4j has many more features. As long as the performance doesn't impede the app any more than the current stuff then it'll be pretty cool 15:06 <MrVisser> thecarlhall: the only significant overhead I can see is processing the JEXL for dynamic tag names. Keeping in mind that the @Profiled annotations has a "log level" parameter (e.g., level=INFO), then I think we can have levels of granularity — not much different than regular logging. 15:07 <MrVisser> so, when level=DEBUG and we're at INFO level, then the JEXL is not processed, I would imagine. Also, you can use the Perf4j StopWatch manually as well, if you wanted to bypass JEXL. Not much different than just using System.currentTimeInMillis() 15:16 <SakaiGithub> [3akai-ux] bp323 pushed 2 new commits to 1.5.0: http://git.io/sV0jWg 15:16 <SakaiGithub> [3akai-ux/1.5.0] SAKIII-4921 - BasicLTI widget sends relatively useless context_id - Christian Vuerings 15:16 <SakaiGithub> [3akai-ux/1.5.0] Merge pull request #2266 from christianv/SAKIII-4921 - Bert Pareyn 15:21 <MrVisser> thecarlhall: if you have more questions or concerns re: Perf4j, just let me know if you want a skype. 15:22 <thecarlhall> mrvisser, sure thing. My only concern is performance and that's because I don't know specifically what perf4j is doing. 15:22 <thecarlhall> mrvisser, I'll try to read up on it and watch the videos. 15:26 <MrVisser> thecarlhall: if it helps, here is the code where perf4j intercepts the method: http://grepcode.com/file/repo1.maven.org/maven2/org.perf4j/perf4j/0.9.13/org/perf4j/aop/AgnosticTimingAspect.java#AgnosticTimingAspect.runProfiledMethod%28org.perf4j.aop.AbstractJoinPoint%2Corg.perf4j.aop.Profiled%2Corg.perf4j.LoggingStopWatch%29 15:48 <thecarlhall> mrvisser, nice ref. Gives me a much better feeling. 16:36 <zacht> raydavis Have you seen or heard ctweney today? 16:36 <raydavis> zacht he just walked in. 16:36 <zacht> nice! 16:36 <raydavis> We aim to please. 16:51 <zacht> We gained ctweney and lost mrvisser. 16:51 <zacht> Coincidence? 16:52 <thecarlhall> ACTION hears dramatic, suspenseful music 16:52 <ctweney> huh zacht? I hope you only mean who's logged in to IRC... 16:52 <zacht> ctweney He's still alive as far as I know. 16:52 <zacht> ctweney Yeah, IRC. 16:53 <zacht> ctweney thecarlhall raydavis lancespeelmon I was hoping we could get on a call at some point today and hammer the next several weeks into shape 16:54 <ctweney> kind of a pre-meeting for planning tomorrow? 16:54 <ctweney> works for me 16:54 <raydavis> OK by me, too. 16:54 <thecarlhall> I'm in. My schedule is open. 16:55 <zacht> The preparation we need for the kickoff is basically to be able to say what we're each going to be working on. 16:55 <zacht> I want people to be in control of their own scope _and_ timeframe. Every time that mistake gets made, I die a little on the inside. 16:56 <lancespeelmon> sounds fine to me zacht 16:58 <zacht> ctweney raydavis thecarlhall lancespeelmon How does 11am PDT sound? 16:58 <lancespeelmon> 2p EDT 16:58 <thecarlhall> +1 17:00 <ctweney> +1 17:01 <zacht> Can you guys see this? https://trello.com/board/v1-5/50101ac7683db67213410ad0 17:02 <zacht> Probably not. One minute. 17:09 <ctweney> zacht, in case you were tempted to sleep well at night, I wrote up kern-3053 as a result of our headscratching yesterday 17:10 <zacht> ctweney What the heck is a SyntheticResource? 17:11 <ctweney> one made of petroleum products, I suppose 17:25 <zacht> mrvisser! 17:26 <MrVisser> zacht! 17:26 <ctweney> zacht: I'm going thru every bundle that has initial content to see which one creates /var as nt:folder. no luck so far. 17:27 <zacht> ctweney ah, but there are plenty of bundles that create content at a _child_ of /var 17:27 <zacht> ctweney Which means Sling has to generate all the ancestors. 17:28 <zacht> mrvisser Are you free in 30 minutes to talk sprint planning with a bunch of dudes? 17:28 <thecarlhall> ctweney, nt:folder is the default type 17:28 <MrVisser> zacht: sure 17:28 <zacht> mrvisser Cool. Look for a Trello invitation in your mail. 17:28 <thecarlhall> ctweney, bundles/search/impl/src/main/resources/SLING-INF/content/var.json is what creates /var 17:29 <ctweney> thecarlhall: not always, it seems 17:29 <ctweney> thecarlhall: sometimes /var comes back as nt:folder instead of sling:Folder 17:30 <zacht> thecarlhall We're guessing it's some kind of race. 17:30 <zacht> thecarlhall If you don't get sling:Folder, initial content fails all over the place. 17:31 <zacht> mrvisser Feel free to create cards for the backlog. 17:31 <zacht> mrvisser We'll talk about who wants what on the call. 17:32 <thecarlhall> zacht, ctweney I think nt:folder is what would get created by creating something below it and that causing a -p style parent creation 17:32 <MrVisser> zacht: OK, sounds good. 17:32 <MrVisser> zacht: offline now for about 5min to get to a skype-friendly place. 17:36 <ctweney> that's what I'd have guessed too thecarlhall, but I've just been thru every bundle that has initial content, deleting /var and then reloading the bundle, and all of them succeeded (sling:folder in every case) 17:41 <MrVisser> In my lame attempt to create a new card, I ended up changing the description of the entire board. Lets try that again. 17:41 <ctweney> thecarlhall: if I wanted to artificially delay a bundle's initial content load, is there a place I could put a breakpoint at? 17:41 <zacht> mrvisser No sweat. 17:41 <MrVisser> If only these buttons had, I don't know, titles that say what they do? 17:42 <thecarlhall> ctweney, there's a thing in sling that does the work. Give me a bit and I'll find the class. 17:42 <zacht> Be back shortly. 17:45 <thecarlhall> ctweney, there's a bundle in sling called o.a.s.jcr.contentloader 17:46 <thecarlhall> ctweney, quite possibly DefaultContentImporter is the key 17:48 <thecarlhall> ctweney, org.apache.sling.jcr.contentloader.internal.Loader looks like a good starting point. 17:56 <MrVisser> now I managed to archive a bunch of cards. Note to self: do not press 'c' in Trello. 17:56 <zacht> mrvisser Sheesh, who presses 'c'? 17:57 <MrVisser> I thought it would "C"reate a new card, like C creates an issue in JIRA :( 17:57 <zacht> mrvisser This is the kind of product you get from people who use Windows. 17:59 <lancespeelmon> where are we meeting zacht ? 18:00 <zacht> lancespeelmon I thought I'd pull everyone into Skype. Sound ok? 20:05 <MrVisser> thecarlhall: I'm rebuilding my perf4j branch to address your feedback. thanks for having a look! 20:06 <MrVisser> thecarlhall: re: perf4j/pom.xml and mixed tabs/spaces: I don't think there is a mix, it's all tabs. Should it be expanded to spaces? 20:06 <MrVisser> also, the "perf4j-enhanced" repo on my github needs to be built first. it will need to be uploaded to the sakai maven repo. 20:07 <thecarlhall> mrvisser, our style guide is 2 spaces; no tabs 20:07 <MrVisser> thecarlhall: ok. I think the eclipse template we have must not include XML files. I'll amend it now while the file has no history. good catch 21:38 <denbuzze> is anybody else having Git(Hub) issues? "git fetch stuartf" just results in: fatal: The remote end hung up unexpectedly 21:39 <thecarlhall> mrvisser, pong 21:39 <thecarlhall> denbuzze, all looks good here: https://status.github.com/ 21:39 <MrVisser> denbuzze: I get that error sometimes as a "catch-all" for other weird things 21:39 <MrVisser> i.e., bad network connection, invalid remote address definition, etc... 21:39 <denbuzze> thecarlhall: that's also the first place I looked but dean440 seems to have similar issues 21:40 <thecarlhall> denbuzze, y'all must be ahead of their monitoring & reporting :P 21:41 <MrVisser> thecarlhall: was thinking it might be better to keep the Perf4J and its logging a "black box" from the rest of the container. It has some strict requirements, such as the config has to be an XML file, and it seems tricky to me to provide a "default" log4j.xml file on the classpath where the log4j bundle could find it. 21:42 <MrVisser> this would mean that log4j would stay embedded in the perf4j bundle, with no log4j exports 21:42 <thecarlhall> mrvisser, I see. I was worried about the embedded things not being ready to look into other classloaders 21:43 <MrVisser> thecarlhall: I think that things are quite well sectioned off, the only exports from the perf4j bundle are those that are absolutely necessary to weave the perf4j things. If we did want to, for example, introduce a container-managed log4j bundle, it should be easily pluggable with no bad side-effects on the perf4j bundle. 21:44 <denbuzze1> thecarlhall mrvisser: it's working fine over WiFi ... but not over cable 21:44 <thecarlhall> denbuzze1, does that goes through a different provider? 21:45 <denbuzze1> thecarlhall: it shouldn't …(UC Berkeley) 21:53 <thecarlhall> mrvisser, I'm not quite following everything there. If we can add a container-managed log4j later, why not now? 21:54 <MrVisser> thecarlhall: because log4j is not the logging implementation for our container now, just for the perf4j bundle. 21:54 <thecarlhall> mrvisser, it is offered by the commons.log bundle though 21:55 <thecarlhall> mrvisser, may not be the full impl but only enough to use it as a client 21:56 <MrVisser> thecarlhall: didn't realize that bundle exported it. perf4j needs the full impl, I think. 21:57 <MrVisser> like, it needs direct access to the appenders and such. 21:58 <thecarlhall> I think commons.log has enough impl to pipe things into slf4j. 21:58 <thecarlhall> I haven't dug very deeply into it though 21:59 <thecarlhall> might be worth exploring the updated version of commons.log 21:59 <thecarlhall> minor details though 21:59 <thecarlhall> for now, just leave log4j embedded in your bundle and don't export it 22:00 <ixmati> anybody know if there are character limits or other restrictions when it comes to usernames and/or passwords? 22:00 <MrVisser> thecarlhall: sounds good. also, re: commons-jexl: the manifest data is bogus. doesn't have symbolic-name 22:00 <thecarlhall> mrvisser, will it not deploy without that? 22:01 <MrVisser> thecarlhall: nah, felix complains 22:01 <thecarlhall> mrvisser, I see this: Bundle-SymbolicName: org.apache.commons.jexl 22:01 <MrVisser> thecarlhall: which version? 22:01 <thecarlhall> 2.1.1 22:01 <MrVisser> perf4j depends on pre-2.x, I think. let me double-check. I think it uses 1.1 22:02 <thecarlhall> in that case, we can either wrap it ourselves or embed it. I'm leaning towards embedding it until something else needs it 22:02 <MrVisser> thecarlhall: me too. 22:03 <MrVisser> and not only because that's less work for me :) 22:03 <thecarlhall> :) 22:03 <MrVisser> I have another PR coming with spacing changes and POM file changes. 22:03 <thecarlhall> right on 22:04 <thecarlhall> what's the news on perf4j-enhanced? just a jar we need to build? 22:06 <MrVisser> thecarlhall: the perf4j-enhanced is just the perf4j jar that we would normally embed, except it was pre-run through the aspectj weaving. the aspectj plugin will do the weaving on-the-fly, but the plugin tries to inline all the class files of weaved dependencies into the bundles. 22:06 <MrVisser> and that makes me (and osgi) sad 22:06 <thecarlhall> mrvisser, cool. I'll do the work locally to see it all work but we'll need wait for the jar to get into a repo before the merge is done 22:07 <MrVisser> thecarlhall: absolutely, I expected that. I think I dressed up the POM file for all that is necessary based on my experience with the oae-maven-plugin release.. but if there's anything else needed just let me know. 22:08 <thecarlhall> mrvisser, if you find that our style guide xml import file thing needs updating, feel free to throw that in, too 22:08 <MrVisser> thecarlhall: yea I'm going to look at that. I've had trouble with XML files in the past but just haven't had time to get around to fixing up the templates. 22:09 <MrVisser> well, maybe not that I haven't had time, just always felt like there was something better to do :) 22:09 <thecarlhall> mrvisser, I'm also quite open to introducing files for IntelliJ and others /cc ctweney zacht raydavis 22:28 <ctweney> thecarlhall, mrvisser +1 for IDEA support 22:28 <thecarlhall> ctweney, feel free to submit any files that help with that. I'd love to see a coherent setup for the popular IDEs 22:39 <MrVisser> thecarlhall: shall I contact zacht about getting perf4j-enhanced "released" to the repos? Or do you have that power? 22:39 <MrVisser> zacht / anthony 22:39 <thecarlhall> mrvisser, I lack that power. Once the review is clear (I can build locally), we can ask for the jar to go into the repo 22:40 <MrVisser> cool, thanks! 22:41 <MrVisser> thecarlhall: I've updated the PR, should be "final" :) 22:41 <thecarlhall> mrvisser, checking it now 22:56 <raydavis> Death by a thousand slashes. 22:58 <thecarlhall> mrvisser, a question while this all builds. should the perf4j-enhanced project be moved to sakaiproject or put into the oae tree under libraries/? 22:59 <MrVisser> thecarlhall: if it is put under libraries/ and inherits from the base POM, I recall running into issues with circular dependencies 23:00 <thecarlhall> mrvisser, circular from perf4j-enhanced to what? 23:02 <MrVisser> thecarlhall: to the base pom itself, I believe. The aspectj plugin must include it as a dependency in order for the "aspectLibrary" declaration to work. Since the base pom includes the definition of the plugin, the base pom effectively depends on perf4j-enhanced when -Pperf4j is enabled. 23:02 <MrVisser> thecarlhall: the alternative is to have all bundles that use perf4j copy/paste the whole plugin declaration into their local POMs to get things to work.. I didn't like that. 23:03 <thecarlhall> mrvisser, eww, agreed. I see the circle now. 23:03 <thecarlhall> mrvisser, it could still be in there with a different parent or created as a separate project 23:03 <MrVisser> thecarlhall: true, I considered breaking out a "master", but seemed to be drastic, and didn't think I had enough of the picture in my head ot make that kind of move. 23:04 <MrVisser> thecarlhall: we could have a master that declares *only* meta-related information. such as the groupId, version, SCM, etc... 23:04 <thecarlhall> mrvisser, you've lost me. 23:06 <MrVisser> thecarlhall: the thing I hated about creating a new project for perf4j-enhanced is that it is a lot of copy/paste. e.g., in the POM there is the scm declaration, maven-compiler plugin, wagon-ssh, etc… If we had a "master" POM that just contained that stuff with no dependencies, then having perf4j-enhanced in the /libraries directory depending on the "master" instead of the "base" would work.. see what I mean? 23:07 <thecarlhall> mrvisser, ahh, I follow now. Certainly something to consider 23:07 <MrVisser> thecarlhall: I haven't touched CLE in a while, but as I recall they have that separation as well. 23:07 <thecarlhall> mrvisser, also very common in apache projects 23:26 <SakaiGithub> [3akai-ux] christianv pushed 2 new commits to 1.5.0: http://git.io/OtuK8g 23:26 <SakaiGithub> [3akai-ux/1.5.0] SAKIII-3583 display an error message when trying to save a document that was deleted - James Sloane 23:26 <SakaiGithub> [3akai-ux/1.5.0] Merge pull request #2028 from jsloane/SAKIII-3583 - Christian Vuerings 23:45 <denbuzze> thecarlhall ctweney: what is the best way to see when a user has last logged in? Is the only way looking at the log file? 23:46 <thecarlhall> denbuzze, I think so. We don't record anything in the db. 23:46 <ctweney> interesting question denbuzze 23:46 <ctweney> I don't think the login code records login time 23:46 <ctweney> so yeah, log file grepping ftw 23:47 <ctweney> it'd be easy enough to add last login time 23:47 <thecarlhall> it would just have to be updated in each of the auth handlers _______________________________________________ oae-dev mailing list [email protected] http://collab.sakaiproject.org/mailman/listinfo/oae-dev
