Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
* fasseg joins01:39
* kaarefc joins03:35
* kaarefc leaves03:55
* leenata joins05:56
* nbanks joins06:27
* github-ff joins07:48
[fcrepo4] nigelgbanks opened pull request #75: Merged the generated models for the request into a single model, before (master...nodes-return-more-than-one-document) http://git.io/5mtGjQ
* github-ff leaves
* jcoyne joins08:15
* jcoyne leaves08:39
* awoods_ joins08:40
<pivotal-bot_>Andrew Woods added comment: "Subsumed by: https://www.pivotaltracker.com/s/projects/684825/stories/51314651" https://www.pivotaltracker.com/story/show/5116453908:59
Andrew Woods delivered "Optimize indexing configuration" https://www.pivotaltracker.com/story/show/51164539
* nbanks_ joins09:03
<pivotal-bot_>Andrew Woods edited "Examine (profile) update data operations." https://www.pivotaltracker.com/story/show/5124224309:04
Andrew Woods added comment: "Was a new, "update datastream" test added to the jmeter suite for this item?" https://www.pivotaltracker.com/story/show/5124224309:05
* nbanks leaves09:06
* gregjansen joins
* kaarefc joins09:25
* kaarefc leaves09:31
* kaarefc joins09:34
* kaarefc leaves09:53
* kaarefc joins
<pivotal-bot_>A. "Codiselic" Soroka added comment: "No, it was there already, inside the module for Fedora4" https://www.pivotaltracker.com/story/show/5124224309:55
* jcoyne joins10:00
* kaarefc leaves10:02
* kaarefc joins10:03
* nbanks_ leaves10:09
* ajs6f joins
* kaarefc leaves10:17
* ajs6f1 joins10:20
<barmintor>aawoods++ // nice work on clustering
<ajs6f1>Can we run the tests against a cluster now?10:21
(The load test, I mean.)
* ajs6f leaves
<barmintor>ajs6f1: The changes I made to get the grizzly stuff working could be rolled into http-commons if someone could sort out a jaxb error w/ web.xml loading in grizzly
<ajs6f1>basrmintor: are you talking about the COntainerWrapper class10:22
<barmintor>(that is, so that the test servlet configs could be expressed as xml rather than done programtically)
ajs6f1: yes
<ajs6f1>Hm. A JAXB error? Weird.
<barmintor>Pretty sure it uses jaxb to parse the wb.xml
<ajs6f1>Ah. And is it like a situation where we're not presenting a valid web.xml beacuse we're conf'ing programmatically?
<barmintor>more a situation that I did it progr. b/c I couldn't get the wb.xml to work right away
<ajs6f1>Oh, we should be doing in in web.xml?10:24
<barmintor>ajs6f1: that would let us re0use the container wrapper10:25
<ajs6f1>Are we not doing that now? I thought the container-wrapper was a construction specifically for building a web container using Spring?10:26
<barmintor>some things can't be bult in Spring (the filters, for example)
<ajs6f1>Well, that actually makes me happy.10:28
I'm not a complete idiot.
It really was bloody (&)(&ing impossible.10:29
* barmintor doesn't touch that one
But you tried using web.xml and couldn't get it to parse?
<barmintor>yes this is true
this is the jaxb problem
a null pointer, actually, which makes me think there's an optional grizzly dep with the web.xml jaxb bindings for w/e version of the spec we're using10:30
<ajs6f1>Okay. Maybe we can talk about this with voices today? I feel like I need to get a little help coming up to speed with how you made it work— I'm not at all familiar with the new Grizzly dep you intro'd.10:31
<aawoods>re: clustering, I was seeing some issues (flakiness) with futures3-5, so I switched to AWS. I will check my Puppet work into a new futures git-repo some time soon.10:39
<ajs6f1>barmintor: I'm a little confused— it looks like the filters for OAuth aren't being engaged at all for the tests… should I ass them back into the Spring as beans? It looks like your code expects to pick them up in the container wrapper by name...10:41
* barmintor is pretty sure it was 'ass'10:42
<ajs6f1>I don't know how to verb ass.
<barmintor>anyway, no- they're not spring beans
they're objects the container creates that happen to know to look for a Spring app context10:43
<ajs6f1>where should they get built? They're not being initialized or deployed?
<barmintor>and then use a bean name to pull a delegate
<ajs6f1>Sorry, that's what I meant— the beans being delegated to.
They don't appear to exist.
But there is a section of Spring XML:
Bad c&p.10:44
Urg. Can't c&p with angle-brackets?
Anyway, in rest.xml there's a comment "authn filters" with nothing beneath it?
Is that where they go?
<barmintor>--> security.xml10:45
<ajs6f1>Ah ha!
<barmintor>they definitely get built, b/c I got an error when I mstyped the name of one
<ajs6f1>Why am I not seeing init logging, then… checking...
<cbeer>ajs6f1: can we bring that disseminator discussion to ff-tech?
<ajs6f1>Right on, cbeer. I'd like to think it's an opportunity to reach out and get some more people involved.10:46
* leenata leaves
<ajs6f1>Was I right in thinking that Andrew just misunderstood what you meant?
<aawoods>ajs6f1: me?10:47
<cbeer>ajs6f1: i think what i said was definitely in the spirit of:
"We haven't done any work on disseminators yet because none of the use cases submitted by institutions or individuals mentioned them very prominently, and we're working understaffed to a horribly tight timeline. That does _not_ imply that we won't ever or that they've disappeared from the horizon of development for the community. "10:48
* github-ff joins
[fcrepo-auth-oauth] ajs6f pushed 1 new commit to master: http://git.io/o3srzg
fcrepo-auth-oauth/master 74794a1 ajs6f: Making tests fail.
* github-ff leaves
<ajs6f1>awoods: no, cbeer.
cbeer: Cool. That's what I figured.
It's funny how face-to-face communications is so prone to noise. We should do fewer conferences. :)10:49
<aawoods>Is there an email thread happening on a list that is interesting?
<cbeer>aawoods: yes, ajs6f1 will move it to ff-tech
* eddies joins
* eddies leaves
* eddies joins
<aawoods>cbeer: where is it currently?
<ajs6f1>aawoods: Yes, as of this morning.
<cbeer>aawoods: i stirred up some disseminator partisans at the hydra partners meeting10:50
aawoods: it's a uva-internal discussion
<barmintor>Down with disseminator partisans
<ajs6f1>aawoods/cbeer: I don't know if any of those UVa people are actually on ff-tech.
<cbeer>ajs6f1: i probably also said something like "if you want it, it's on you to build it. i'm not touching WSDL."
<barmintor>All praise the 1000 year REST!
<ajs6f1>cbeer: WSDL is just irrlevant.10:51
<barmintor>I think it would be easy to re-implement much of it over the MODE base, fwiw
tho not the WSDL part
<ajs6f1>barmintor: yes. I'm already doing it.
<cbeer>ajs6f1: just saying, i probably confused the fcrepo3 implementation with the possibility of doing something like disseminators (ever).10:52
<ajs6f1>cbeer: Exactly. And that's what I tried to explain to Andrrw, althought I think it was he who made the confusion.
<barmintor>ajs6f1: are you doing the whole "this node has this contract mixin, and it provides it for nodes of this other mixin type"?10:53
<ajs6f1>will explain more later, got to walk back to the office...10:55
* ajs6f1 leaves
* nbanks joins10:56
<eddies>get up. stand up11:00
<jcoyne>I and I are standing up11:02
Hmm, what is the proper verb to use in that sentence?
* ajs6f joins11:04
barmintor: I haven't got as far as pre-runtime binding yet. I've only gotten as far as thinking through the nature of declaring the structure of a disseminator.11:10
<cbeer>aawoods: how important is windows support before OR?11:12
<aawoods>cbeer: It is definitely a concern I have... Windows support.
I am guessing we will want to engage with Windows users at OR11:13
<barmintor>ajs6f: interested to hear it as it develops
<aawoods>That would be a big blow, if we have to say "but if you are on Windows, sorry\"
<ajs6f>For the nonce in my mind, a diss is an ordinary object, with datstreams with names like "GET", "POST" (the verbs), and in those datastreams the content for each is an URL-tempalte.
And there's no binding before runtime (until we figure out how we want to do that) so any diss can be used against any object.11:14
and the method, headers, and params just carry through.11:15
(No backend authN management if we can help it!)11:16
* nbanks leaves
* nbanks joins
<aawoods>fasseg: that link is not producing anything for me11:19
<fasseg>grml this one: https://docs.google.com/file/d/0B5nd_qlYdcqyZ0ZiVFB0WnZlTVU/edit?usp=sharing
<barmintor>eddies: ok, I'll be at my desk for a while11:20
"a while" = pretty much the rest of the day until 4pm
<pivotal-bot_>Andrew Woods added comment: "Can you add your findings to the wiki, ala: https://wiki.duraspace.org/display/FF/Profiling+Fedora+4+Delete" https://www.pivotaltracker.com/story/show/5124224311:22
<aawoods>gregjansen: Would you mind creating (three?) new tickets to breakdown your current work?11:26
<gregjansen>aawoods: sure thing
<fasseg>I used this config for the test: https://github.com/ModeShape/modeshape-examples/tree/master/modeshape-filesystem-store-example/src/main/resources11:32
<cbeer>ok, i'm an idiot - my test was broken by a typo.. so i'll push the first round soon, and the async thing next.11:36
<aawoods>fasseg: It is probably worthwhile to make sure your team can run fcrepo4 on Windows...11:43
as we have no other Windows members on the team.
<ajs6f>and HP/UX, and CPM-M. All the major players. :)
* barmintor shivers at the mention of HP-UX
<aawoods>I would argue that we should have support for: Windows and *nix
<ajs6f>Let's not go crazy!
Apple Dos 3.3 should do everyone just fine.
<aawoods>I know, it is a stretch
<ajs6f>barmintor: I hate to say it, but from the logging, it looks like the filters aren't getting applied… did you see them actually engage?11:46
<barmintor>ajs6f: Hey man, I got them created and ran the tests. You want me to actually know what's going on?11:47
<cbeer>barmintor: you had some tickets in mind, right? not looking for people to dump tickets on you?
<ajs6f>I don't expect anyone to do anything I wouldn't do, so, no, I wouldn't expect you to know what's going on.
<barmintor>cbeer: I don't really have anything in mind, but eddies might
<cbeer>barmintor: ok -- i have https://www.pivotaltracker.com/story/show/46108229 that is now unblocked11:48
<pivotal-bot_>feature: Backup a running repository (started) / owner: Chris Beer
<cbeer>and i wonder if it "just works"
aawoods: i'm not going to do it, but i see there's some tweaks for indexing in a cluster11:54
to send index updates around over jgroups or jms11:55
<ajs6f>cbeer: is the alternative (which we, presumably, are now using) that each cluster member generates its own index updates, duplicatively?11:57
<cbeer>ajs6f: it's unclear from the config.11:58
<ajs6f>Fair enough.
<aawoods>cbeer: It would probably be useful to note such config opportunities with inline xml comments, at least.11:59
<cbeer>aawoods: this is in the modeshape json config.. i don't think json has a comment syntax (right?)12:00
<aawoods>no, comments and json do not overlap
<cbeer>i guess i could add a README
<aawoods>unless you use something like a "description" element
README is probably cleaner
<cbeer>aawoods: unfortunately, modeshape does schema validation on its json files12:01
<ajs6f>MODE does have that (a description element).
<cbeer>ajs6f: yeah, but you'd also need to fill out the required elements
<ajs6f>cbeer: Don't we have to do that now?
<cbeer>ajs6f: sorry. if i create the contextually-appropriate block to put the comment in the description field, i also need to fill out the required fields for that block. sure, i could put it at the top-level or something... or i just create the README rider and call it good12:02
<ajs6f>cbeer: Not trying to make life difficult: I assumed that we'd want to put in-line comments right next to whatever config we were commenting on, so I assumed we had such config.12:03
<cbeer>ok, or i don't understand this configuration well enough to make an educated comment12:06
so a ticket it is!12:07
* github-ff joins12:08
[fcrepo4] cbeer created configuration-to-fcrepo-jcr (+1 new commit): http://git.io/_e3f9w
fcrepo4/configuration-to-fcrepo-jcr 7947724 Chris Beer: move various modeshape/infinispan configurations into fcrepo-jcr
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer opened pull request #76: move various modeshape/infinispan configurations into fcrepo-jcr (master...configuration-to-fcrepo-jcr) http://git.io/RXF1QQ
* github-ff leaves
<eddies>barmintor: call?12:09
<barmintor>eddies: there in a sec12:10
* github-ff joins12:12
[fcrepo4] cbeer force-pushed configuration-to-fcrepo-jcr from 7947724 to 15d0105: http://git.io/L3scEg
fcrepo4/configuration-to-fcrepo-jcr 15d0105 Chris Beer: move various modeshape/infinispan configurations into fcrepo-jcr
* github-ff leaves
<cbeer>aawoods: i'm still trying to figure out how kitchen sink wires into all that12:15
but it should be good enough for webapp testing
(i suspect there's a java opt we need to add to the kitchen sink to get it to use our kitchen sink config)12:16
<barmintor>eddies: https://github.com/futures/fcrepo-auth-oauth/blob/master/src/test/java/org/fcrepo/auth/oauth/integration/api/ContainerWrapper.java#L8812:17
* nbanks leaves12:19
<bljenkins>Project fcrepo-fixity-corrupter build #67: SUCCESS in 49 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/67/12:23
<ajs6f>eddies/barmintor: Are y'all talking about this Spring stuff (which is still beating me up— the filters seem to exist, but not to engage!)
<barmintor>ajs6f: no, but will look at it when off call12:26
<ajs6f>barmintor: It's cool— just wanted to make sure I couldn't help. As usual, I couldn't.
<bljenkins>Project fcrepo-fixity build #248: STILL UNSTABLE in 4 min 27 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/248/
<ajs6f>Does anyone else think it's funny that our fixity application seems to have historically been the least stable part of our codebases?12:27
Not ha-ha funny, sigh-funny.
<cbeer>escowles did the htmlprovider/dataset context stuff, right?12:34
<ajs6f>Context stuff?
<eddies>barmintor: https://docs.google.com/a/yourmediashelf.com/file/d/0B5nd_qlYdcqyZ0ZiVFB0WnZlTVU/edit
<cbeer>ajs6f: the thing that lets the htmlprovider have a way to figure out what the graph is "about" so it can choose the right template12:35
<eddies>barmintor: using this config https://github.com/ModeShape/modeshape-examples/blob/master/modeshape-filesystem-store-example/src/main/resources/my-repository-config.json
<cbeer>pretty sure it was escowles, but we've touched a bunch of the lines since then
<ajs6f>cbeer: Yes, he did, although I think there wasn't a separatye issue for it...12:36
<bljenkins>Project fcrepo4 build #770: UNSTABLE in 15 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/770/12:37
<eddies>barmintor: http://www.yourkit.com/docs/12/help/agent.jsp12:41
barmintor: specifically, in my catalina.sh, I included YJP_HOME="/Applications/YourKit_Java_Profiler_12.0.5.app"
<eddies>and I append YJP_OPTS to my CATALINA_OPTS
<barmintor>so much for portable
<eddies>did anyone else ever get YourKit to attach to mvn jetty:run?
or even just Tomcat without all the crap I pasted above?
<cbeer>ajs6f: ok, question for you now. I'm looking at https://github.com/futures/fcrepo4/blob/master/fcrepo-http-commons/src/main/java/org/fcrepo/responses/HtmlProvider.java#L10412:43
<barmintor>eddies: https://code.google.com/p/maven-yourkit-plugin/wiki/GettingStarted ?
<cbeer>(sorry, context: i'm trying to add a template for search results)
(and search results don't have a jcr node, obviously)
<ajs6f>Erg. The templates are loaded entirely based on the node type of the results...
That has to change.
<cbeer>ajs6f: ok, i've been trying to phrase my question.. how painful is that?12:44
can we iterate over a known directory?
<ajs6f>Sure, in java.io
But what known directory?
Maybe a known part of the classpath?
<cbeer>ajs6f: apparently "templatesLocation"
<ajs6f>That's in the classpath, not a directory on disk.
the question is, how is the new (query result) tempalte going to get chosen.
<cbeer>oh, rightr.
maybe i should just hard-code a search results template for now
and we can figure out a generic way later?
<ajs6f>We can put everything in that directory into play, but how (based on the response, will we know when to use the query-results-temaplte.12:46
<cbeer>(as of now, at least, search results is a weird one-off)
<ajs6f>maybe using the "primaryType" was too limited?
Yeah, that's true.
It's the only non-node-backed result, for now.
<cbeer>ok. i'll do something messy (with tests) and we can deal with it later when we have another example of this case12:47
<ajs6f>k. I think we want to avoid tangling presentation with the data, but we may need to come up with a layer of indirection in between to get from -assertions in the data- to -which template to use-.12:48
<cbeer>ajs6f: i know! let's use disseminators!
<ajs6f>We _are_ using disseminators. It's _all_ disseminators. YOU ARE A DISSEMINATOR.12:49
* nbanks joins12:53
<cbeer>ajs6f: bah. we do have another non-node-backed graph. namespaces.12:58
<ajs6f>k. Then we have to do the job right.
So what's the real rule to select a template?
Based on the request URI?
<cbeer>hm. we do have the URI12:59
sometimes a single URI maps to multiple templates
s/URI/URI pattern/
<ajs6f>Yeah, yeah. Like a given URI could be an object or a datastream.
So that's not enough, but it might be one input.
URI form + primary type if it exists?13:00
<cbeer>ajs6f: do we have access to method annotations in the providers?
We do for ijection providers.
I'm not sure about response writers.
Lemme check.
<cbeer>i think for many of these non-node backed templates, it'd be fine to leave an annotation saying "this is an X" or13:01
"just use this template!"
<ajs6f>Yup, we do:
https://jsr311.java.net/nonav/javadoc/javax/ws/rs/ext/MessageBodyWriter.html#isWriteable(java.lang.Class, java.lang.reflect.Type, java.lang.annotation.Annotation[], javax.ws.rs.core.MediaType)
<cbeer>is that a stupid idea?
<ajs6f>No… as long as we're sure that this is the minority of cases, which it certainly seems like it to me, and that default (no annotation) behavior is easy to understand (just use the primary type to select? something else?)13:02
Or are you suggesting explicitly annotating every resource method?
No, rereading your remark, you are not.13:03
<cbeer>i think a default that introspects the graph (using primary type or otherwise) is fine
(and, we could do that here.. but it's kinda annoying to jump through that many hoops just to get the right information in the right place)
<ajs6f>Yeah, as an "exception-handling" mode (&^(^, there's no primary type, what do I do!?) it works fine.
Well, there are ontologies for discussing query results.13:04
<cbeer>and we're almost using them
<ajs6f>e.g. Stanbol has one.
<cbeer>but then we end up with a:
is the a primary type? do this
is there a search results property? do that
is there a namespace thing? do something else
<ajs6f>Yeah, that
's why I say that as long as we're talking about the exceptions...13:05
a few annotations feels cleaner.
Much as I like making EVERYTHING RDF.
<eddies>barmintor: i saw that maven-yourkit plugin but didn't try it. last commit was in 2007
<cbeer>(although i wonder if we only have those 3 cases..)
<ajs6f>Oh, jeez.
<cbeer>oh, transactions.
that's 4.
time to do it right.
<ajs6f>Transactions? There's an endpoint with an RDF representation for transactions?13:06
Damn, we _are_ into RDF.
<cbeer>there isn't yet, i think
but probably should be..
<ajs6f>Yeah, I guess that we've bought the ticket. Let's get on the train.13:08
<cbeer>ok. i'll hack something together and someone else can make it nice13:10
<eddies>barmintor (and others): I just got YourKit to work with mvn jetty:run on fcrepo-webapp13:11
<ajs6f>The way to make it nice is to… add more RDF!
<eddies>MAVEN_OPTS="-Xmx512m -XX:MaxPermSize=128m -agentpath:/Applications/YourKit_Java_Profiler_12.0.5.app/bin/mac/libyjpagent.jnilib=delay=10000"
<eddies>(the critical thing being the -agentpath)
and i think for linux users, it's a .so rather than .jnilib
<ajs6f>I'd be curious as to whether there's a big diff between Jetty and Tomcat. (Or other containers, for that matter.)
<eddies>ajs6f: i'm hoping it's essentially a constant difference though, so i can not worry about it =)13:13
(for profiling purposes, anyway)
<cbeer>ajs6f: hm. i guess we could also use annotations to direct the htmlprovider to the right predicate to use13:15
i'm not doing that.
<ajs6f>cbeer: Definitely phase 2.13:16
cbeer: How about start with 'constant" annotations, and another ticket to "paramterize" them?13:17
(You can put me on that, as a holding pattern.)
<ajs6f>cbeer: How about metaprogramming? Is that cool? :)
<cbeer>ajs6f: do you know off-hand how to go from a Proxy back to the original value?13:22
<ajs6f>You mean a dynamic proxy?
I don't think you can.
Not without inspecting JVM-impl types.13:23
Where are we using dyn proxies
<cbeer>ajs6f: the annotations on https://jsr311.java.net/nonav/javadoc/javax/ws/rs/ext/MessageBodyWriter.html#writeTo(T, java.lang.Class, java.lang.reflect.Type, java.lang.annotation.Annotation[], javax.ws.rs.core.MediaType, javax.ws.rs.core.MultivaluedMap, java.io.OutputStream)
are a bunch of $Proxy objects13:24
hm. maybe it doesn't matter
<ajs6f>The annotations on that method itself, or the annotations received as a parameter?
If you are looking for a particular annotation (as a test to decide what to do)
<cbeer>sorry, answered my own question. turns out it doesn't matter
<ajs6f>you can use Class.isAssignableFrom().13:26
Oh, cool.
afk bbi513:32
* ajs6f leaves
* ajs6f joins13:33
<cbeer>ajs6f: any suggestions for writing unit tests about these annotations?13:36
* nbanks_ joins
<cbeer>oh, i guess i can just mock like crazy.13:37
<barmintor>Fedora 4: Mock like Crazy13:38
<ajs6f>Fedora 4: We're not mocking with you, we're mocking at you.13:39
* nbanks leaves13:42
* fasseg leaves
* github-ff joins13:49
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/XK1owQ
fcrepo4/master 1a76a7b Chris Beer: Merge pull request #76 from futures/configuration-to-fcrepo-jcr...
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer created html-template-annotation (+1 new commit): http://git.io/KGAfSg
fcrepo4/html-template-annotation 8fb5c80 Chris Beer: hack in a search results html template.
* github-ff leaves
<cbeer>ajs6f: can you review https://github.com/futures/fcrepo4/pull/77?
<ajs6f>On it.
* github-ff joins13:50
[fcrepo4] cbeer opened pull request #77: hack in a search results html template. (master...html-template-annotation) http://git.io/vyR66w
* github-ff leaves
<ajs6f>Did you make me the requestor on a ticket, or do you just want me to check and merge the code?
<barmintor>if you're going to make al your methods private, you should have a good api13:51
* barmintor is annoyed13:52
<cbeer>ajs6f: no ticket.
<ajs6f>cbeer: cool.
<cbeer>this is just a means to an end
but i did file the "make it better" issue
<ajs6f>barmintor: a good API and an AWESOME type system.
<cbeer>whenever pivotal-bot_ feels like telling us
pivotal-bot_: hi?
pivotal-bot_: hi
<ajs6f>pivotal_bot_: New ticket: MAKE EVERYTHING BETTER FOR EVERYONE13:53
pivotal-bot_: hello
* pivotal-bot_ leaves
* pivotal-bot joins13:54
<cbeer>pivotal-bot: hello
<ajs6f>OMG. They got to pivotal_bot. And she was the best coder of us all!
<pivotal-bot>Hello, cbeer
<cbeer>guess it isn't on my end. pivotal must not be sending webhooks13:55
<ajs6f>cbeer: is it worth breaking out the "extra" templates into a conf file?
<cbeer>ajs6f: i don't know. i felt bad about hard-coding them there13:56
but not bad enough to do anything about it
I just worry about us forgetting that that's what we did.
urg, rain! afk13:57
* ajs6f leaves
* ajs6f joins14:01
cbeer: Can we not kill VelocityViewer? Isn't it on;y used for the query endpoint?
<gregjansen>whew, I finally got my bagit events y'all14:02
<ajs6f>cbeer: I think that would let us kill the http-api dep on Velocity, as a consequence.
grejansen: did you successfully manage to make a reservation at UPEI dorms? I can't get anyone to answer my calls of emails there!
<bljenkins>Yippie, build fixed!
Project fcrepo4 build #771: FIXED in 13 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/771/
Chris Beer: move various modeshape/infinispan configurations into fcrepo-jcr
<gregjansen>I made an online reservation at Glenndenning Residence. url in a sec14:03
<ajs6f>Online? cool. I'm theresville.
<gregjansen>so no, UPEI it is not
<bljenkins>Project fcrepo-fixity-corrupter build #68: SUCCESS in 58 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/68/
<gregjansen>it is 2km from site14:04
<ajs6f>Oh, wait— there's another school in Ch'town?
I'll hitch a ride on a skateboard holding on to your bike.
<gregjansen>cool, yeah it is cheap as heck14:05
<ajs6f>Now I've got to call them and ask if they will accept a purchase order from UVa. Yes, believe it or not that's the only way I can pay for accomodations.
<gregjansen>no reimbursements system there?
<ajs6f>If you're willing to wait six months.
That's been the average for me.14:06
Frequently I have to get Martha Sites to personally call someone at Procurtement and chew them out.
Don't even start me.
<bljenkins>Project fcrepo-fixity build #249: STILL UNSTABLE in 3 min 57 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/249/
* travis-ci joins14:10
[travis-ci] futures/fcrepo4#653 (master - 1a76a7b : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/9a53bdc1757a...1a76a7b24a7b
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7957490
* travis-ci leaves
<barmintor>NB: Grizzly bombs if webAp@metadata-complete doesn't exist14:11
<ajs6f>barmintor: What?14:13
is that a servlet 3.0 thing?
The "this web.xml is all the config" attribute?
<ajs6f>But our i-tests don't even _have_ a web.xml?!
<barmintor>it's not required, that's a bug in Grizzly 1.9.57
my local branch does
that's where my jaxb npe was coming from
<ajs6f>So should all of them?
<barmintor>not yet
but soon
and for the rest of their lives14:14
<ajs6f>Or just when we're trying to use a web.xml, it needs to be complete...?
<ajs6f>and you're saying we should use web.xml everywhere we can?
<barmintor>a web.xml in the local project that is being tested will allow a generic ContainerWrapper to be defined in http-commons14:15
when and if I get this wortking
<bljenkins>Project fcrepo4 build #772: UNSTABLE in 12 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/772/
<barmintor>there's two bugs I'd submit a fix to grizzly for if Oracle didn't use that BS svn repo14:16
<ajs6f>Not an "if", only a "when". It's barmintor, after all.
What about modules that don't expose web resources...?
<barmintor>ajs6f: this *is* http-commons we're talking about :D
<ajs6f>e.g. fcrepo-jms should continue to set up with just Spring...
<barmintor>I'm assuming that if you're not testing a web resource, you're also not using grizzly14:17
<ajs6f>k, right.
<pivotal-bot>A. "Codiselic" Soroka added comment: "Sure, but the setup is exactly like eddies, except I think I used more loops. Perhaps just the scr..." https://www.pivotaltracker.com/story/show/51242243
<ajs6f>Wow. I added that comment several hours ago.14:18
<pivotal-bot>Gregory Jansen added "Fire created, updated and removed node events for BagIt federation" https://www.pivotaltracker.com/story/show/5141584714:20
Gregory Jansen estimated "Fire created, updated and removed node events for BagIt federation" as 3 points https://www.pivotaltracker.com/story/show/51415847
Gregory Jansen started "Fire created, updated and removed node events for BagIt federation" https://www.pivotaltracker.com/story/show/51415847
Chris Beer finished "Profile report of jmeter test plan" https://www.pivotaltracker.com/story/show/5099792914:21
Gregory Jansen added "Create an integration test that copies projected bagit nodes to internal nodes via Workspace.copy()" https://www.pivotaltracker.com/story/show/51416177
Gregory Jansen started "Create an integration test that copies projected bagit nodes to internal nodes via Workspace.copy()" https://www.pivotaltracker.com/story/show/51416177
Gregory Jansen edited "Create a sequencer that copies projected bagit nodes to internal nodes." https://www.pivotaltracker.com/story/show/50748291
<cbeer>ajs6f: i'll take a look at the velocity viewer. i had to re-introduce it to get an html representation when there are no search results
<ajs6f>cbeer: interesting. The null case…
Maybe there's special logic for when there are no results...?14:22
But we shouldn't need that.
<pivotal-bot>Chris Beer added "Investigate alternative indexStorage options (especially in a clustered configuration)" https://www.pivotaltracker.com/story/show/51417691
<cbeer>and just make our own empty graph?
<pivotal-bot>Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/76" https://www.pivotaltracker.com/story/show/51314651
<cbeer>hey, there's the webhooks
<pivotal-bot>Chris Beer added comment: "(I wouldn't call them recommended.. but at least show some of the different options)" https://www.pivotaltracker.com/story/show/51314651
<ajs6f>yeah, they're all coming in a rush.
<pivotal-bot>Edwin Shin added comment: "@Nigel: wherever you land tomorrow w/ Islandora, please make sure your latest work is deployed to futures6" https://www.pivotaltracker.com/story/show/48351401
<cbeer>ajs6f: and, there IS special logic for no results and TEXT_HTML
<pivotal-bot>Chris Beer finished "Create recommended single-node and multi-node configs" https://www.pivotaltracker.com/story/show/51314651
<ajs6f>cbeer: Okay, well, I think we can rip that out.14:23
<cbeer>ajs6f: right.. and just give the provider an empty model?
<ajs6f>Yep. That's what the results are, right?
<cbeer>sure, i guess.14:24
ok, i'll take a stab at that this afternoon
other than that, good to go?
<ajs6f>Echoing, like the vast chambers of out unfulfilled desires for disseminators.
cbeer; yeah, I don't really care about the hard coded "eception" templates, because we log them when they are used.
cbeer: so, yeah, good to go.14:25
<pivotal-bot>Chris Beer started "Fix search results HTML template" https://www.pivotaltracker.com/story/show/51085649
<ajs6f>cbeer: unless you want to rewrite in SCALA!
<cbeer>ok, heading to the office.14:26
<pivotal-bot>Chris Beer started "add HTML representation to namespace endpoint" https://www.pivotaltracker.com/story/show/5109795314:28
Chris Beer added "Parameterize Dataset HTML response " https://www.pivotaltracker.com/story/show/51426639
Chris Beer estimated "Parameterize Dataset HTML response " as 3 points https://www.pivotaltracker.com/story/show/5142663914:29
<barmintor>ajs6f: ok, I now see what is going om14:30
<ajs6f>barmontor: that makes one of us.14:31
<barmintor>ajs6f: your delegates also need to live in web-app space (web.xml)
<ajs6f>barmintor: doesn't that take us right back to the start? The delgates need beans (like SessionFactory) that live in child contexts...14:32
<barmintor>they need to get thier Spring beans "by hand" like the Spring filters do
<ajs6f>Oh, jeez. That sounds brittle.
<barmintor>Spring doesn't have access to FilterConfig's and what not
that's the app container's job14:33
<ajs6f>So they reach into child contexts and pull out the beans they need?
By name?
<barmintor>well, not exactly
they reach into the only context they have
the app contexty
and yes, by nma
err… name
<ajs6f>So the SessionFactory bean, and everything it relies on (like the repo stuff) have to come back up to the app context...?
<ajs6f>Okay. I'll try that.
your delegates can elect not to rely on FilterConfigs
<ajs6f>They don't.
They just impl the interface, but they're ordianry beans, with mutators.14:35
so to speak
If they don't need FilterConfig, how dos that help?
<barmintor>it lets Spring init them14:36
<ajs6f>Spring does that now. Via PostConstruct.
That's why I wasn't seeing the init take place earlier.
<barmintor>I'm responding here to your question earlier- "Why don't I see the init logging?"
<ajs6f>Oh, right. Sorry, I fixed that and didn't tell you.
<barmintor>well what's your problem now?
<ajs6f>They init. They exist. They just don't filter.
They never get called in the course of a request.
<barmintor>they don't get mapped to?
<ajs6f>I guess. I tried to inspect the filters set up via the Contexts, but the ServletContext in the Sprin gtest framework is a mock which doesn't support most ops.
Including retrieving the filters in play.14:38
<ajs6f>I'm going back to my office. afk bbi10. Thanks for the help so far!
* ajs6f leaves
* ajs6f joins14:46
Wow. We've been under Flash Flood Watch for several days now. Now Tornado Watch, too. I remember when I was a kid and summer thunderstorms were just a lot of noise and fireworks.14:50
Now they're scary monsters. I feel like my dog, like hiding under the bed.
* fasseg joins14:54
* jcoyne leaves14:55
<pivotal-bot>A. "Yefremination" Soroka unstarted "Parameterize Dataset HTML response " https://www.pivotaltracker.com/story/show/5142663914:56
* jcoyne joins
* github-ff joins15:07
[fcrepo4] cbeer force-pushed html-template-annotation from 8fb5c80 to 268aea9: http://git.io/0UiWRg
fcrepo4/html-template-annotation 268aea9 Chris Beer: hack in a search results html template.
* github-ff leaves
<ajs6f>cbeer: was that the death of VelocityViwer?15:08
<cbeer>oh, yeah. i should probably git rm that
<ajs6f>I think we can kill the velocity dep in the pom.xml, too.
Cause we're getting that from http-commons15:09
<cbeer>(although that's a little weird too..)
i'm not sure why the html templates don't live with the http-api
<ajs6f>Maybe they should… I don't why I put them in commons. I guess I htought of them "suppiorting" in some way?15:10
<cbeer>i guess they're not tightly bound to our REST API
and the RDF generating stuff is in kernel15:11
so, sure. i'll kill velocity in http-api
ok. pushed to the branch. i didn't finish running the tests, but off to lunch15:15
hopefully travis will let us know what it thinks
* github-ff joins15:16
[fcrepo4] cbeer force-pushed html-template-annotation from 268aea9 to 231da5b: http://git.io/0UiWRg
fcrepo4/html-template-annotation 231da5b Chris Beer: Add the annotation @HtmlTemplate to allow JAX-B methods to hint at...
* github-ff leaves
<bljenkins>Yippie, build fixed!15:18
Project fcrepo4 build #773: FIXED in 10 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/773/
Project fcrepo-fixity-corrupter build #69: SUCCESS in 54 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/69/15:19
Project fcrepo-fixity build #250: STILL UNSTABLE in 3 min 7 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/250/15:21
* github-ff joins15:22
[ff-jmeter-madness] fasseg pushed 1 new commit to master: http://git.io/NMD2mw
ff-jmeter-madness/master 7557cdd fasseg: added simple 'create node' test for vanilla modeshape
* github-ff leaves
<bljenkins>Project fcrepo4 build #774: UNSTABLE in 11 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/774/15:29
<fasseg>created this page from my modeshape results so far: https://wiki.duraspace.org/pages/viewpage.action?pageId=34654157 it's still missing a comparison between sync/async though, which might be useful..15:30
<pivotal-bot>Andrew Woods added comment: "A wiki capture of the results screenshot with some text that gives the results context seems reasonable." https://www.pivotaltracker.com/story/show/5124224315:46
* github-ff joins16:22
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/VLu5-A
fcrepo4/master 907ddd2 Chris Beer: Add the annotation @HtmlTemplate to allow JAX-B methods to hint at...
* github-ff leaves
<pivotal-bot>Chris Beer finished "Fix search results HTML template" https://www.pivotaltracker.com/story/show/5108564916:25
Chris Beer finished "add HTML representation to namespace endpoint" https://www.pivotaltracker.com/story/show/51097953
<ajs6f>barmintor: This is really quite weird. At this point, the tests are passing, because the /token endpoint is returning 200s. Which is coo. Except it's not actually executing. I put some logging in, when I noticed that no actual tokens were appearing, and then I found out that something is returning 200s from the URL of the /token endpoint, but it's not the code I wrote. So things are going from not working to surreal.16:40
<ajs6f>Something like that.
* travis-ci joins16:43
[travis-ci] futures/fcrepo4#657 (master - 907ddd2 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/1a76a7b24a7b...907ddd20516f
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7961858
* travis-ci leaves
<pivotal-bot>Chris Beer started "java.lang.NullPointerException at org.fcrepo.syndication.RSSPublisher" https://www.pivotaltracker.com/story/show/51155279
<cbeer>ajs6f: i'm taking a look at that bug ^16:44
and it looks like the eventbus isn't getting wired into the rss or webhooks stuff
for reasons i can't explain yet..
(of course.. i can't figure out how the kitchen sink and webapp tests are even loading spring..)16:45
<ajs6f>cbeer: Like in the i-tests?16:46
<cbeer>yeah. i assume there's some maven magic going on
<ajs6f>Maven? How would Maven inject the shizzit?
<bljenkins>Yippie, build fixed!16:47
Project fcrepo4 build #775: FIXED in 24 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/775/
Chris Beer: Add the annotation @HtmlTemplate to allow JAX-B methods to hint at
<cbeer>ajs6f: well, both those projects are wars
<ajs6f>But the i-tests work out of containers built by Spring, right?
* github-ff joins
[fcrepo-bagit-modeshape-federation-connector] gregjan pushed 1 new commit to master: http://git.io/KzPjAA
fcrepo-bagit-modeshape-federation-connector/master 0706091 Gregory Jansen: factored out the BagItWatchService, as we didn't need to subclass WatchService...
* github-ff leaves
<cbeer>maybe not on either of these?
<ajs6f>Hm, maybe not. I just assumed they were like the others...16:48
<bljenkins>Project fcrepo-fixity-corrupter build #70: SUCCESS in 47 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/70/
<pivotal-bot>Gregory Jansen added comment: "add and remove events are now fired for projected bag nodes" https://www.pivotaltracker.com/story/show/5141584716:49
<ajs6f>But even then, how would the build tools change what's getting innjected?
<bljenkins>Project fcrepo-fixity build #251: STILL UNSTABLE in 3 min 22 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/251/16:50
Project fcrepo-bagit-modeshape-federation-connector build #19: UNSTABLE in 3 min 39 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-bagit-modeshape-federation-connector/19/16:51
Project fcrepo-kitchen-sink build #354: SUCCESS in 4 min 24 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/354/16:52
<barmintor>Is there a reason we're using version 1.9.x of grizzly? It looks like the current version is 2.3.3?17:01
<ajs6f>barmintor: fasseg installed grizzy.
I'm outta here. See y'all later.
* ajs6f leaves
<eddies>cbeer: war-producing modules run against jetty (via jetty-maven-plugin). so the ITs for fcrepo-webapp and fcrepo-kitchen-sink are indeed different from the other ITs which use grizzly17:08
<pivotal-bot>Gregory Jansen added comment: "Bag folders do not have enough state to really broadcast property updates or addition and removal of chil..." https://www.pivotaltracker.com/story/show/5141584717:09
<cbeer>eddies: ok. i'm going to try a few more things and then give up on this ticket.
i think there's something about component scanning i don't understand17:10
<eddies>cbeer: how are you triggering that NPE? in a test?
(i haven't caught up on the earlier irc chatter)17:11
<cbeer>any request to /rest/fcr:rss is a NPE
or /rest/fcr:webhooks
tests or in fact.
<eddies>cbeer: i don't see org.fcrepo.syndication in context:component-scan of fcrepo-kitchen-sink's rest.xml (or fcrepo-webapp's)17:16
<cbeer>eddies: yeah, i saw that. and added it locally.
but no change.
i wonder if the rest.xml i think is getting loaded isn't..17:21
<eddies>are you using kitchen-sink or fcrepo-webapp?17:22
<cbeer>kitchen sink. webapp doesn't have rss or webhooks enabled.
<eddies>cbeer: just check fcrepo-kitchen-sink/target/fcrepo-kitchen-sink-4.0-SNAPSHOT/WEB-INF/classes/spring/rest.xml17:24
(assuming you've done a mvn install or package)17:25
<cbeer>looks right.17:28
i'll mention it on standup tomorrow
and try to get any takers.17:29
* github-ff joins17:44
[fcrepo-kitchen-sink] eddies pushed 2 new commits to master: http://git.io/J4KoPg
fcrepo-kitchen-sink/master 2501488 Edwin Shin: Moves Metrics-related classes and dependencies to fcrepo4/fcrepo-webapp....
fcrepo-kitchen-sink/master 4391872 Edwin Shin: Now that ModeShape 3.3 is out, uses ${modeshape.version}
* github-ff leaves
<bljenkins>Project fcrepo-kitchen-sink build #355: UNSTABLE in 2 min 49 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/355/17:47
* gregjansen leaves17:48
<pivotal-bot>Chris Beer started "Tidy up HTML implementation to ensure all data we want "designed" is exposed" https://www.pivotaltracker.com/story/show/5074165917:58
Chris Beer finished "Tidy up HTML implementation to ensure all data we want "designed" is exposed" https://www.pivotaltracker.com/story/show/50741659
Chris Beer estimated "Tidy up HTML implementation to ensure all data we want "designed" is exposed" as 3 points https://www.pivotaltracker.com/story/show/50741659
<eddies>there's something odd happening with github for fcrepo417:59
<cbeer> cbeer: Heads up. Just in case. https://github.com/mbklein/puppet-fcrepo4
<eddies>i pushed a commit, got an error, and my latest commit doesn't show on github.com. but if i clone the fcrepo4 repo, i see that latest commit...
<cbeer>from mbklein ^
eddies: sometimes github lags.18:00
<pivotal-bot>Chris Beer added "Create a puppet script for fcrepo4" https://www.pivotaltracker.com/story/show/51445649
Chris Beer estimated "Create a puppet script for fcrepo4" as 0 points https://www.pivotaltracker.com/story/show/51445649
Chris Beer started "Create a puppet script for fcrepo4" https://www.pivotaltracker.com/story/show/51445649
Chris Beer finished "Create a puppet script for fcrepo4" https://www.pivotaltracker.com/story/show/51445649
<pivotal-bot>Chris Beer added comment: " https://github.com/mbklein/puppet-fcrepo4" https://www.pivotaltracker.com/story/show/5144564918:01
Chris Beer delivered "Create a puppet script for fcrepo4" https://www.pivotaltracker.com/story/show/5144564918:02
Chris Beer unstarted "java.lang.NullPointerException at org.fcrepo.syndication.RSSPublisher" https://www.pivotaltracker.com/story/show/51155279
* jcoyne leaves18:03
<eddies>ok, y'all can now use jmx to get metrics reporting on fcrepo-webapp and kitchen-sink, which should hopefully pair nicely w/ any profiling exercises18:06
<pivotal-bot>Chris Beer added "Update to metrics 3.0.0" https://www.pivotaltracker.com/story/show/51446085
Chris Beer started "Update to metrics 3.0.0" https://www.pivotaltracker.com/story/show/51446085
<eddies>man, i just updated to the 3.0.0-SNAPSHOT :-P18:07
<cbeer>so, we just need the infinispan 5.3.0 release18:08
and then we're on full releases of stuff
<pivotal-bot>Chris Beer accepted "Update to metrics 3.0.0" https://www.pivotaltracker.com/story/show/5144608518:11
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/DvPbdA
fcrepo4/master 8ddc9ce Chris Beer: update to metrics 3.0.0
* github-ff leaves
* nbanks_ leaves18:16
<pivotal-bot>Chris Beer delivered "Fix search results HTML template" https://www.pivotaltracker.com/story/show/5108564918:17
Chris Beer delivered "add HTML representation to namespace endpoint" https://www.pivotaltracker.com/story/show/51097953
<bljenkins>Project fcrepo-fixity-corrupter build #71: SUCCESS in 40 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/71/18:23
Project fcrepo-fixity build #252: STILL UNSTABLE in 3 min 12 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/252/18:25
<eddies>i think we need to update grizzly18:28
version we're using still uses servlet 2.518:29
but i'm off to bed
* travis-ci joins18:32
[travis-ci] futures/fcrepo4#658 (master - 8ddc9ce : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/d4741a8f25c9...8ddc9cebe822
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7964899
* travis-ci leaves
<cbeer>aawoods: ping?18:35
on a call
<cbeer>aawoods: is it anyone on the steering group's responsibility to update the https://wiki.duraspace.org/display/FF/Fedora+Futures+Home page?18:36
i noticed it hadn't been touched since CNI
so i'm doing stuff..
* nbanks joins18:43
<cbeer>"Open Repostitories Conference"18:46
* nbanks leaves18:50
<aawoods>cbeer: I think the short answer is 'no'. It is also interesting to note the similarity with this other page on the wiki: https://wiki.duraspace.org/display/FF/Fedora+Futures+Prospectus19:30
<cbeer>aawoods: because i just copied content from that page :)
after modeshape linked to it, i realized there's nothing there that says wtf this project is.19:31
<aawoods>Clearly, we have some work cut out in terms of our documentation19:36
<cbeer>(not it!)
<aawoods>Actually, I think leading into OR2013 we need to pull together the features and capabilities that have been proven/tested in the Fedora4 effort to-date.19:37
We need to be able to point people at the wiki and have them gain a clear picture of where things stand.
<cbeer>hm. where are those sprint-end docs?19:39
<aawoods>We can coordinate the best way to get that done.
Some of them are here: https://wiki.duraspace.org/display/FF/TT+Meetings
<cbeer>hm. does moving pages break existing links?19:40
<aawoods>it should not, no.
It depends on how the links were created.
you can use full http links, or wiki syntax links
the former break
<cbeer>ok. i'm going to assume we don't, and move that page to the top-level https://wiki.duraspace.org/display/FF/Sprint+Reports19:41
i guess when I said "not it", i meant i'd shuffle stuff around for a couple minutes.
<aawoods>that is the spirit, cbeer.19:42
<cbeer>i hate confluence.
or, shhhhh19:43
however, you must admit, clicking through all of those meeting pages to try to figure out what Fedora4 can currently do is a task very few will endure.19:44
interesting, I was just in the process of creating a "fcrepo4-aws-puppet" project19:45
<cbeer>"mbklein: It's just daemonizing mvn jetty:run as its container; no time to deal with standalone jetty or tomcat yet. But it's 10 minutes from "vagrant up" to "curl http://localhost:8080/rest" and it's repeatable. So there you are."19:51
so.. when can we get northwestern on board?
<aawoods>Sounds like we already have19:52
I will bring it up at tomorrow's SC meeting
cbeer: How did you connect with mbklein?19:53
<cbeer>aawoods: #projecthydra
<aawoods>does he commit to hydra?19:54
<cbeer>yes. and developer on avalon and a bunch of other stuff
<aawoods>maybe I will hand off my puppet work to mb.
which does deal with stand-alone tomcat19:56
* github-ff joins
[fcrepo-auth-oauth] barmintor pushed 1 new commit to master: http://git.io/oztHFw
fcrepo-auth-oauth/master 3613db3 Benjamin Armintor: actually getting the resource to work via upgraded grizzly and jersey libs
* github-ff leaves
<barmintor>I'm going home. That was stupidly painful for the triviality of my goal.19:58
<aawoods>no pain, no pain
* nbanks joins20:17
* nbanks leaves20:22
* fasseg leaves20:27
* fasseg joins20:55
* aawoods leaves21:31
* nbanks joins22:18
* nbanks leaves22:23
* jcoyne joins22:43
* jcoyne leaves22:51
* nbanks joins23:19
* nbanks leaves23:23
* barmintor_ joins23:44
* barmintor_ leaves23:48

Generated by Sualtam