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

Using timezone: Eastern Standard Time
* anusha joins08:39
* eddies leaves08:41
* github-ff joins08:52
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/WpIupA
fcrepo4/master c65b26f ajs6f: Fix for https://www.pivotaltracker.com/story/show/44066767
* github-ff leaves
* anusha leaves09:16
* eddies joins10:08
* eddies leaves
* eddies joins
* JasonDGI joins
* anusha joins10:09
<cbeer>https://www.pivotaltracker.com/projects/68482510:12
* cbeer leaves
* cbeer joins
* ajs6f joins10:14
<barmintor>good morning ajs6f10:15
<ajs6f>Hoy, all1
<cbeer>ajs6f: we're standing up now. did JasonDGI call you in?10:17
<ajs6f>No, I thought it was at 212?
11?
* fasseg joins
<cbeer>yeah, but we finished breakfast early
<JasonDGI>calling you adam
<ajs6f>Ohhhhkay.10:18
<JasonDGI>ajs6f ring ring10:19
<ajs6f>We need to stick a specific time for the standup.
<cbeer>ajs6f: we need to know what timezone we're in :P
<ajs6f>Yes, you do.10:20
Sure, I'm here. Why?10:30
<JasonDGI>i just got a skype alert that you went offline
<ajs6f>I'm on mute, so you can't here my keyboard.
<JasonDGI>:)
you should tell us to mute because our keyboards are too loud ;)10:31
<ajs6f>Yep.10:33
This issue:https://www.pivotaltracker.com/story/show/4402174310:35
no longe rakes any sense.
After fcrepo4.
I'm going to mark it finished, to indicate that.
<barmintor>ajs6f: ok. delivering.10:39
<ajs6f>Thanks, B.
<eddies>https://github.com/futures/ff-jms-solr-indexing10:45
<fasseg>adam: https://github.com/futures/ff-jms-solr-indexing/blob/master/src/test/java/org/fcrepo/index/Fedora4Mock.java10:46
here the framework used: http://www.simpleframework.org/
<ajs6f>Thanks, E and F.10:47
<fasseg>and the server mock is started in this class: https://github.com/futures/ff-jms-solr-indexing/blob/master/src/test/java/org/fcrepo/index/JMSIndexingTest.java
at line 6010:48
<cbeer>eddies: https://github.com/futures/ff-jmeter-testResults/tree/master/2013-02-ModeshapeTests/results
<fasseg>hah direct line links thanks to jason: https://github.com/futures/ff-jms-solr-indexing/blob/master/src/test/java/org/fcrepo/index/JMSIndexingTest.java#L60-7010:49
<ajs6f>That looks quite reasonable.10:51
Okay, guys. I'm going to hang up. Seems like standup is over.10:53
<barmintor>ajs6f: is there a submodule that contains the JAXB/JMS stuff?
<ajs6f>JAXB/JMS?
Theres JAXB used in kernel.
And all the JMS stuff is jms.10:54
Never the twain shall meet.
<barmintor>yes, but there is some code that serializes a Message to text
<ajs6f>Where are you looking? In jms?
<barmintor>and we suppose the converse
fasseg is asking
<ajs6f>No, Apache Abdera is being used to construct Atom messages.10:55
<cbeer>anusha: https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/FedoraObjects.java
<ajs6f>Is that what F is asking about?
How do the Atom messages get created?
<barmintor>ajs6f: ok, that's what I was looking at. maybe we need a ticket to implement a deserialization of the messages.
<ajs6f>Well, if we keep just Atom, that wold be Abdrea. If we have pluggable serializations (which would be awesome) we sure need some contracts around that.10:56
Are you going to start working on the definition of a FedoraEvent?
<barmintor>ajs6f: is it? I couldn't find an Abdera "readFrom" method, only the "writeTo"
i.e., I could only see Abdera's serialization10:57
<ajs6f>I think I'm still not understanding the question properly.
The onnly serialization used _is_ Abdera's.
I don't have independenet code to serialize.
<barmintor>yes, but Abdera doesn't provide a parser
<ajs6f>I'm more confident in Abdera's than in any JAXB code I would write.
<cbeer>ajs6f: anusha is going to start on the Merritt REST API for fcrepo4, should she just copy fcrepo-webapp and hack from there?
<ajs6f>Does fcrepo-webapp work?10:58
http://abdera.apache.org/docs/api/org/apache/abdera/parser/package-summary.html
<cbeer>ajs6f: nope, there's no code in there :P
<ajs6f>Then no, she shouldn't.
<fasseg>adam: my question was if there was a submodule, that has all the JMX annotated classes, which I could include as a meven dependency to access this objects...somtehing like fcrepo4-datamode or sth..
*JAXB10:59
<cbeer>https://github.com/futures/fcrepo4/tree/master/fcrepo-kernel/src/main/java/org/fcrepo/jaxb/responses11:00
barmintor: ^
<barmintor>ajs6f: incoming
<ajs6f>Oh, then this has nothing to do with Abdera or Atom. No, those JAXB classes (I assume that's what you mean) are in kernel. They could usefully be broken out. fcrepo-responses? They don't repreesent a data model. Just HTTP reposnses.
<fasseg>I think these classes should go into something like fcrepo4-common: https://github.com/futures/fcrepo4/tree/master/fcrepo-kernel/src/main/java/org/fcrepo/jaxb/responses
<ajs6f>+!11:02
<barmintor>Being right means never having to say you're sorry :P
<cbeer>ajs6f: do you know what's left to make fcrepo-webapp work? just copy the relevant source files in there?
<ajs6f>fcrepo4-common sounds good to me.
No, it should not have any source.11:03
If it's not building, it's because one of the other modules isn't building.
It depends on them.
<cbeer>oh, so it's just there to build a WAR?
<ajs6f>Right. If A. wants to build a Merritt API...
perhaps she should be working against fcrepo-kernel?
And creating a optional module fcrepo-merritt?11:04
Does that makes sense as any answer to your question?
Oh, wait, Anusha, you're in room. Sorry to talk about you in the third person.
<anusha>:)
<cbeer>ajs6f: oh, i forgot, fcrepo-dc has a JAX-RS resource too... so we can just do something like that, right?11:05
<anusha>I'm just listening in, so that's fine
<ajs6f>Yeah, it'd be along those lines.
Just be aware that you'll have the thing with testing that I had, although F.'s suggestion should clear all that up.
And be aware that we need to find a way to gather up all the Spring contexts that optinal modules supply and initialize them in a webapp. Currently, the webapp would just start the kernel's Spring context.11:06
A Karaf framwork would handle this for us. That might be the quikest way forward, and it would put the least presure on module devs to know anything about Fcrepo. Just "put your Spring here and it will get launched".11:07
Or we could define our own convention.
* JasonDGI leaves11:09
<ajs6f>afk
* eddies leaves11:14
* eddies joins11:16
* eddies leaves
* eddies joins
<cbeer>eddies: got a sec?11:18
ajs6f: i might take the dc test chore on myself11:22
unless you've started it and want to fight me for it.
oh, i see you started that ticket.
maybe we should be patient11:23
<eddies>http://stackoverflow.com/questions/2247199/share-test-resources-between-maven-projects
* github-ff joins11:25
[fcrepo4] cbeer created merritt (+1 new commit): http://git.io/LZNW0g
fcrepo4/merritt 7fb650a Chris Beer: stub out merritt REST API module
* github-ff leaves
<barmintor>cbeer++
<cbeer>barmintor: other than it doesn't work :P11:26
but ajs6f will fix it up for the -dc one
and anusha will do the real work
so, ajs6f, should the fcrepo REST API in fcrepo-kernel move to.. fcrepo-legacy-api or something?11:27
<eddies>http://stackoverflow.com/questions/174560/sharing-test-code-in-maven
<cbeer>barmintor: whenever you want to talk MODE/ISPN, let me know.
<barmintor>cbeer: looking at docs now11:28
<eddies>anusha, cbeer: http://stackoverflow.com/questions/174560/sharing-test-code-in-maven#174670
* JasonDGI joins11:29
<eddies>(actually, i guess that was more for ajs6f)
<barmintor>http://docs.jboss.org/infinispan/4.0/apidocs/org/infinispan/Cache.html#putAsync%28K,%20V%2911:31
* eddies leaves11:39
<cbeer>barmintor: https://github.com/ModeShape/modeshape/blob/master/modeshape-jcr/src/main/java/org/modeshape/jcr/value/binary/infinispan/InfinispanBinaryStore.java11:46
* JasonDGI leaves11:48
<ajs6f>back
Eddie— I think we're actually doing pretty good on our orthogonality as far as sharing code/test code between modules. It's more about needed to share execution contexts. that's what I really want to avoid. But I think we'll be all right. I'm going to take a look at the dc-generator thing and try to apply F.'s suggested framework there.11:50
<cbeer>barmintor: https://github.com/ModeShape/modeshape/blob/master/modeshape-jcr/src/main/java/org/modeshape/jcr/value/binary/FileSystemBinaryStore.java
barmintor: https://issues.jboss.org/browse/MODE-179011:52
ajs6f: can you ping anusha and me when you fix the dc generator test?11:55
<ajs6f>Sure. It's taking a hilw because I
m working on a different machineand have to get everything all set u. :(
<cbeer>ajs6f: would it help if I made a mess of it in the meantime, or do you think once you have everything set up it'll come together pretty quick?11:57
<ajs6f>Hm.
<cbeer>(and i'm not necessarily looking for things to do)
(so telling me "hands off" is a great answer :P)
<ajs6f>Then no. It shouldn't take long. FAMOUS LAST WORDS.
<cbeer>k11:58
(for just future reference, I want to look at org.infinispan.loaders.keymappers.Key2StringMapper someday.)12:05
* JasonDGI joins
<ajs6f>I'm more and more wondering if for some of what we want to do, we're going to have to go around ModeShape itself and talk to INSPN.12:06
<barmintor>ajs6f: it's like you're in the room with us12:07
<cbeer>ajs6f: or do some MODE/ISPN work ourselves. barmintor and I are thinking about alt. implementations for some of the modeshap/ispn interfaces.
<ajs6f>For asynch?12:08
<cbeer>ajs6f: yes.
and other places.
<ajs6f>Right.
<cbeer>i mean, there's no reason modeshape needs to present only a synchronous API to ISPN, right?
it's just convenient.
(and I think we should float this by #modeshape sometime, once we think it's not unreasonable)12:09
<ajs6f>Nope, except that there is no JCR-defined asynch API to implement. But MODE already does things that go beyond JCR (e.g. sequencing).
<cbeer>ah, true.12:10
JCR 3 here we come!12:12
<ajs6f>Actually, rhauch@modeshape told me in IRC that they are confident that much of their extension work _will_ go into the next JCR spec. (2.1) But how fast? I'm not holding my breath.12:19
Got it.12:39
Let me tighten the logging and so forth, and I'll push it (fixed fcrepo-dc).12:40
I did not use the mock framework, because what was being tested was all server side.12:41
* github-ff joins12:47
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/jSehKg
fcrepo4/master 73b7e66 ajs6f: Fixed fcrepo-dc so that build works: https://www.pivotaltracker.com/story/show/44093327
* github-ff leaves
* github-ff joins12:54
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/USaxWA
fcrepo4/master 3f21576 ajs6f: Changed module tests to use memory-resident JCR repository for much faster test execution
* github-ff leaves
<cbeer>ajs6f++12:55
i was thinking about doing that
and, anusha, i'll incorporate ajs6f's dc work in our branch, unless you beat me to it.
<ajs6f>Now I'm going to do the same thing for some of the other test apparatus— the ActiveMQ broker in fcrepo-jms, for example.
* github-ff joins13:09
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/OcSzmQ
fcrepo4/master 114bde2 ajs6f: Changed ActiveMQ broker used in tests to an in-memory, in-VM broker for better test execution speed
* github-ff leaves
<ajs6f>C. and A., are you able to work against the kernel? Are you writing the Merritt stuff as a module under fcrepo4? Or is there anything blocking you or slowing you down from doing that?13:10
<cbeer>ajs6f: anusha is looking at your patch now.13:12
<ajs6f>The dc-fix patch?
<cbeer>ajs6f: i'm going to look at -dc closer, because i might have a question for you about it
yes
<ajs6f>Cool. Ping me for help.
I'm going to tighten the use of test resources in the kernel. There's no need for it to set up clustering!13:13
<cbeer>also what i was thinking about doing :)
so, a big +1
ajs6f: ok, first question.. in fcrepo4, where can i run mvn jetty:run now?13:17
<ajs6f>Hm. You may not be able to. Can you try it at the kernel?
<cbeer>nope. so here's what i tried:
running at the top-level (no such plugin)
running at -kernel (no such plugin)13:18
running it at webapp (can't find the fcrepo-* things)
<ajs6f>I think it got left behind somehow.
<cbeer>is the webapp error
<ajs6f>?
<cbeer>in fcrepo-webapp, when i run mvn jetty:run, I get that error13:19
<ajs6f>Can't find dependencies?
<cbeer>that's what it says.. so i guess it's a path problem or something, because the tests run fine at the top level
<ajs6f>Have you done a mvn install?13:20
If not, the webapp has no place from which to get the built dependencies.
Mvn is not like an IDE— it has no contextual awareness, so to speak. If the dependency isn't either in a repo or manually added to the classpath, it's not there.13:21
vs. in say, Eclipse...
which can figure out that a dependency is actually a neighboring project.
<cbeer>ah. so mvn test does something else too, huh?
<ajs6f>So have you successfully done mvn install in the top level?
<cbeer>ajs6f++
nope.
now i have
and it works13:22
<ajs6f>Cool.
mvn is really smart, but really dumb.
<cbeer>or, almost worked.
spring errors now:
org.springframework.beans.factory.BeanDefinitionStoreException: IOException parsing XML document from ServletContext resource [/WEB-INF/classes/spring/master.xml]; nested exception is java.io.FileNotFoundException: Could not open ServletContext resource [/WEB-INF/classes/spring/master.xml]
i guess i can copy them into -webapp?
or do something else?
<ajs6f>That's an error in the web.xml. Let me look at that right now.13:23
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/8yhV1A
fcrepo4/master 1f7521e ajs6f: Now using non-clustered INSPN for kernel tests, for better test speed
* github-ff leaves
<ajs6f>Okay. Here's what's happening:13:24
in the old ff-modeshape-prototype, all code was being built into one war.
In the new fcrepo4, the code is coming in in jars.
So we have to get a little smarter about how we address things. This is what I meant this morning about having used relative paths.13:25
They all broke.
This is another.
Let me figure out what the right path to get to that context now is.
'cause it's stuck inside fcrepo-kernel.
Actually...13:26
what do you guys think about have one webapp context that loads whatever we want, and it lives in the webapp artifact.
?
Kind of the way Ben did the fcrepo3 stuff.13:27
<cbeer>if that makes sense, go for it. i don't know enough.13:29
and maybe that piece will help address my other question, which is how does the fcrepo-dc JAX-RS stuff get exposed through the webapp13:30
but i think that piece will make it all make sense to me
<ajs6f>Yeah, this is really crucial.
We've started to break things up into modules. Now how do they come back together?13:31
Especially in the absence of an actual module framework, which we don't yet have.
<cbeer>barmintor/eddies: ^^^?13:32
<barmintor>ajs6f: that's how I would do it for now13:33
<ajs6f>One ring to rule them all?
<barmintor>yeah
with different contexts for testing13:34
<ajs6f>Okay. For now, it will have to include the config files, too.
<barmintor>(of course)
<ajs6f>yeah, they all already have their own test contexts.
<cbeer>ajs6f: are you going to do that?
<ajs6f>this is just about assembling a webapp.
<barmintor>well, you can configure the spring web app context to load all the config files in a directory
<ajs6f>Yeah, Ben's gentle words have given me the confidence to go on.
<barmintor>so it doesn't have to be one literal context config, just one app context13:35
<ajs6f>Of course— it'll amount to maybe fie files, each nice and short.
<barmintor>ok, just making sure we were on the same page
<ajs6f>repo.xml, rest.xml, jms.xml, etc.
We still have no adjustability in the modularization, but I dont even want to think about that without a framework.13:36
<cbeer>+113:38
(lunch)13:39
* barmintor leaves13:40
* barmintor joins14:09
<cbeer>ajs6f: i guess i'm still waiting on your context fixes, right? so i'll shift gears, either to:14:22
- moving the fcrepo 3.x API methods into a separate module (if that makes sense to do?)
- a new API method we talked about over lunch as worth exploring
<ajs6f>Just about done/.14:23
Maybe a few more minutes.
<cbeer>- some of the fcrepo-common stuff (which, i guess might just be in fcrepo-kernel if -kernel is clean?)
ok, i'll just study up for the other tasks then
<ajs6f>There's definitely stuff in fcrepo-kerne that could go into a commons bundle.14:25
Like the JAXB stuff.
Git it.
Got it.
<cbeer>Good.14:26
<ajs6f>Give me a minute to ust test through some curl stuff and make sure everything is sane.
<cbeer>while you're doing that, can you be sure to check the /objects/{pid}/oai_dc endpoint too?
<ajs6f>Yes, I just did, and it's the one that doesn't work yet. {grin}14:28
I'm pretty sure I know what's happening here
It's the same thing you ran into originally.14:29
CXF doesn't want to allow two jaxrs:servers on the same adress ('rest/")
* eddies joins14:30
* eddies leaves
* eddies joins
<ajs6f>What do you think, for a quick fix, of offering dc at "/oai_dc/{pid}'?
<cbeer>sure14:31
<ajs6f>Oaky, that ought to do it. Let me check.
<cbeer>yeah, some day we should consolidate them, i guess14:32
but i suppose the proper module framework will help?
<ajs6f>Oh, yeah.14:34
Hm… I'm not sure that worked. Something more sublte is going on here.14:35
Maybe you better ignore me for now.
<cbeer>want to push to a branch, and i can help flail?14:40
<ajs6f>Good though.14:41
thought/
give me a sec to do that.
* eddies1 joins14:42
* eddies1 leaves
* eddies1 joins
* eddies leaves
* eddies joins
* eddies leaves
* eddies joins
* eddies1 leaves
* github-ff joins14:45
[fcrepo4] ajs6f created FlailingCXF (+1 new commit): http://git.io/YryO-g
fcrepo4/FlailingCXF b5743df ajs6f: Cannot get CXF JAX-RS paths correct
* github-ff leaves
<ajs6f>Ah, this is starting to become clearer.
Somehow CXF isn't recognizing DublinCore as a JAX-RS resource class.14:46
But why not?
Wait, that got it working!
Okay, I'll push that chris, and you can use it.
But its "/oai/objects/{pid}/oai_dc" and not the "/rest/objects/{pid}/oai_dc" you wanted.14:47
<cbeer>ajs6f++ # better you than me
yeah, i'll live with that for now
because i assume we can fix it up with a proper module framework or something
<ajs6f>Okay. I'll get that out.14:48
Well, I think I need to really deep-dive into how CXF does what it does.
I need to understand the Spring config we're using.
Because I thought jaxrs:server elements could share addresses, but now I think I was wrong.
So if you take that branch, Chris, mvn jetty:run should work in the webapp. It works for me. Just with the URL difference for oai.14:50
<cbeer>ok. think we should push that through to master for now?14:52
(or, i think it should be pushed to master, do you disagree?)14:54
<ajs6f>No, we might as well. I think I've started to get it. I think what we want to do is surprisingly, change JAX-RS libraries. Here's why:
CXF doesn't offer (that I can find) any kind of scanning service to find JAX-RS resrouces.14:55
That's why we have to explictly configure every bean that provides an edpoint.
But other libraries do.
<cbeer>i thought we were using resteasy?
<ajs6f>So we could just have our resrouces scanned, like we did at first with RESTEasy.14:56
<cbeer>guess not any more huh?
i must have missed that change
<ajs6f>No, we moved to CXF really early.]
But maybe we should move back.
I tell you what.
Let's go ahead and merge this temp fix.
I don't want to leave you and Anusha in the lurch.
<cbeer>thanks14:57
<ajs6f>But I will do a better fix going forward.
* github-ff joins
[fcrepo4] cbeer opened pull request #1: Cannot get CXF JAX-RS paths correct (master...FlailingCXF) http://git.io/Y4hftw
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/GYEMuQ
fcrepo4/master 34e49df Chris Beer: Merge pull request #1 from futures/FlailingCXF...
* github-ff leaves
<cbeer>ajs6f++
<ajs6f>Can you write me a ticket so I don't forget?
Like, "Use alternative JAX-RS library to avoid explicity resource configuration"
<cbeer>sure. resteasy is as good a choice as any?14:58
<ajs6f>Sure.
The more I thin about it, the more I realize how much I've been fighting CXF.
Not a good sign.
<cbeer>would you rather I take the ticket? at least if it's easy, i should be able to knock it out now
<ajs6f>Hell, yeah.
<cbeer>:)
<anusha>Adam, your changes are working for me. Thanks a lot14:59
<ajs6f>I'm going out for a pipe. I'll check in when I get back to see if you have any questions. Then I'm going to go out and do a few errands and go home, where I'll check in again.
Just remember...
there are spring contexts in every "module", in
<cbeer>ajs6f: one quick question
<ajs6f>Yeah?15:00
<cbeer>i think i asked before, and maybe ignored the answer
should the fcrepo 3.x REST API move to a different module?
<ajs6f>Two questions in reeply:
1) Do we have any other API?
(yet)
2) What would be the contract between the kernel and that new module?
<cbeer>nearly. anusha is doing one.
<ajs6f>Sweet.15:01
<cbeer>(CDL Merritt)
<ajs6f>And she's writing directly against JCR?
<cbeer>yes, because we have nothing else to write against yet.
but once she's done, i can see us splitting the common bits back into -kernel, right?
<ajs6f>Maybe JCR is the contract that answers question numer 2, for now.
Yessir.
I would love for kernel to be nothing but JCR and our eventing.
With a minimum operation repo being kernel and your choice of HTT api .15:02
kernel _with_ your choice, that is.
You'd need at least one API.
<cbeer>maybe.15:03
anyway, cool. i'll hack on this and see if i can get anywhere quickly
<ajs6f>We might begin to add more to the kernel (as previously discussed) like linking against external services.
Yeh. bbs
<cbeer>i think a scanning approach will be more code4lib-developer-friendly
<ajs6f>It's more developer friendly period.15:04
* eddies leaves
<ajs6f>I dont know why CXF doesn't do it, but I can't find it if they do.
<JasonDGI>cbeer: why dont i have access to the tomcat logs folder on futures6?
<cbeer>JasonDGI: sudo to root.15:05
* eddies joins
* eddies leaves
* eddies joins
<barmintor>Jesus. Maybe glacier backup should be a sequencer? Fuck this. I'm working on something else.15:45
<cbeer>barmintor: want to help convert the CXF back to Resteasy?15:46
<barmintor>cbeer: when did we go to cxf? why are we thrashing on this?15:47
<cbeer>barmintor: when we went to spring
<barmintor>cbeer: oh. why are we going back to resteasy?
<cbeer>barmintor: because cxf can't have multiple jaxrs:servers listening on the same base path
and doesn't do auto-discovery
<barmintor>hmm
how did I do this in FCF3?15:48
<cbeer>ajs6f: barmintor is going to try to replicate the fcrepo 3 CXF stuff in fcrepo-webapp15:53
trying to get resteasy working right with spring and maven was causing me trouble
* barmintor leaves15:54
* barmintor_ joins
<ajs6f>Urg. Sounds like you guys are not having fun. Maybe this was why I went to CXF to begin with (I've been trying to remember). Spring-friendliness...16:02
<cbeer>yeah
<ajs6f>But shouldn't it be orthogonal?16:03
Or is the lifecycle of scanning conflicting with Spring bean instantiation?16:04
I'm going out to do my errands. I'll check back at some point. If you guys get frustrated, push a branch and I'll pick it up. Good luck!16:06
* github-ff joins16:08
[fcrepo4] anusharanganathan pushed 3 new commits to merritt: http://git.io/HsOLXQ
fcrepo4/merritt 528ab44 Anusha Ranganathan: Merge branch 'master' into merritt
fcrepo4/merritt 66b5222 Anusha Ranganathan: Merged Adam's changes into fcrepo-merritt (so build works - https://github.com/futures/fcrepo4/commit/73b7e66ec6f64d0d27b578dccecd2341ac8cf7e1)
fcrepo4/merritt 80d731b Anusha Ranganathan: Merge branch 'master' of https://github.com/futures/fcrepo4 into merritt
* github-ff leaves
* ajs6f leaves
<cbeer>barmintor_: do you have your trader joe's receipt?
<barmintor_>cbeer: I'm not sure
<cbeer>know how much it was?16:09
* eddies leaves16:18
* eddies joins16:19
* eddies leaves
* eddies joins
<cbeer>barmintor_: how's it going?16:30
<barmintor_>trying something out16:31
<cbeer>k. i'm going to refactor out the fcrepo 3.x legacy api into a separate module16:35
localhost:9999/rest/objects/my_pid/oai_dc16:42
* barmintor_ leaves
* avessey1 leaves16:47
* barmintor joins16:55
<cbeer>curl -X POST http://localhost:8080/rest/objects/my_pid/16:56
curl http://localhost:8080/rest/objects/my_pid/oai_dc
* barmintor leaves16:59
* ajs6f joins17:15
* github-ff joins
[fcrepo4] eddies pushed 1 new commit to master: http://git.io/vlVQmg
fcrepo4/master dbc521a Edwin Shin: Adds our local Nexus repo for distribution
* github-ff leaves
<ajs6f>Any luck with the JAX-RS COnundrum?17:16
On another note, I have discovered the perfect code name for this release: "Axe Cop".
<cbeer>ajs6f: yes, maybe
barmintor is doing stuff within CXF
and having auto-inject trouble, but is apparently pretty close17:17
<ajs6f>Cool. Please do push a branch at somse point. I'd love to see what you guys are up to.
<cbeer>(but he's sitting in the living room, so doesn't get wifi)
<ajs6f>I thought Ben got Wifi natively.
<cbeer>if i can figure out this test-jar stuff, i'm going to push the fedora rest api module thing.
<ajs6f>Excellent.17:21
* github-ff joins17:26
[fcrepo4] anusharanganathan pushed 1 new commit to merritt: http://git.io/FnJ2Vg
fcrepo4/merritt b4904f5 Anusha Ranganathan: Added get service method to the fcrepo-merritt api
* github-ff leaves
* github-ff joins17:28
[fcrepo4] eddies pushed 1 new commit to master: http://git.io/-iOiwg
fcrepo4/master 0246961 Edwin Shin: changing updatePolicy for JBoss Maven repo to default (i.e. daily)
* github-ff leaves
* barmintor joins17:32
ajs6f: I can't resolve FOMXLParser. cheer says it's some scala thing. help?17:33
<ajs6f>Are you getting that from your IDE?17:34
<barmintor>IDE and from jetty startup
<ajs6f>Hm. Jetty startup. That's weird.
What's your IDE?17:35
<barmintor>eclipse
who generates that class?
<ajs6f>src/main/scala/org/fcrepo/foxml/FOXMLParser.scala
Have you set up your Eclipse for Scala?17:36
<barmintor>I guess not. Is there going to be more scala in the project?17:37
<ajs6f>There may not be any.17:38
We could rewrite the trivial FOXML parser in Java.
<eddies>+1
<barmintor>I think for using scala/jruby/jython/etc we want to either have a lot of it or none of it, right?17:39
<ajs6f>Right.
So we're jus tgoing to do more Java.
or we're going to learn how to do other things.
aybe we need to move the foxml stuff out of the kernel.
Then we can not wrorry about it for now.
It doesn't belong there anyway.
<barmintor>ok
<ajs6f>Why not just cut it out for now.17:40
<barmintor>tell me to do something!
<ajs6f>I'll bring it back when the build is more stable.
<barmintor>ok
<ajs6f>Cut out FedoreOXML, FOXMLSequencer, and FOXMLParser.
And cut out the section on sequencers in the ModeShape config JSON file.
<barmintor>wait, cut out the sequencer?17:41
<ajs6f>It's not going to run without a parser.
If you want FOXML, you either parse it or not, right?
<barmintor>is there a test suite for the parser?
<ajs6f>Noope.17:42
<barmintor>I don't want to throw the baby out w/ the bathwater
<ajs6f>Well, we're not going to lose much.
It took me about twenty minutes.
That's the great hting about Scala and XML.
It does it good.
<barmintor>ok
<ajs6f>There's a ticket or two in this:
* github-ff joins
[fcrepo4] cbeer created fcrepo-legacy-api (+1 new commit): http://git.io/kKkVUA
fcrepo4/fcrepo-legacy-api c7ae1d9 Chris Beer: move legacy api to its own module
* github-ff leaves
<ajs6f>module for serializations, module for FOXML implementation.17:43
<cbeer>ajs6f: can you review the pull request that's coming through in a sec?
* github-ff joins
[fcrepo4] cbeer opened pull request #2: move legacy api to its own module (master...fcrepo-legacy-api) http://git.io/uOfZAA
* github-ff leaves
<cbeer>(there's an unfortunate testing dependency from fcrepo-dc -> fcrepo-legacy-api to deal with.. but that seems like it's for another day)
<ajs6f>Yeah, I didn't rewrite any of those tests, because I wanted to get the build running.17:44
<cbeer>hey, whoa. i see some weird stuff in that diff
hold on
<ajs6f>I'm not doing anything faast— I'm trying to cook dinner.17:45
<cbeer>hm. guess it was just in one place. weird.17:47
<ajs6f>Okay, so should I look at that branch.
?
<cbeer>yes17:48
i'm pushing a commit now
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to fcrepo-legacy-api: http://git.io/Qz9RQA
fcrepo4/fcrepo-legacy-api 347739b Chris Beer: continue using import static instead of FedoraObjects.*
* github-ff leaves
<cbeer>that'll clean up the thing i thought was weird about it
https://github.com/futures/fcrepo4/pull/2/files
mainly I want a sanity check that I removed the right bits from fcrepo-kernel17:49
and copied the right bits over
<ajs6f>How do I fetch that brnahc?17:50
I tried get fetch fcrepo-legacy-api
but it doesn't show up in the output of git branch17:51
<cbeer>ajs6f: if you go to https://github.com/futures/fcrepo4/pull/217:53
and click the (i) next to "Merge with caution"17:54
github helpfully tells you how to get the code.
* barmintor leaves
<ajs6f>Urg.17:55
ajs6f-mbpro-3:modeshape-prototype ajs6f$ git checkout -b fcrepo-legacy-api origin/fcrepo-legacy-api
fatal: git checkout: updating paths is incompatible with switching branches.
Did you intend to checkout 'origin/fcrepo-legacy-api' which can not be resolved as commit?
That makes no sense, does it?
<cbeer>did you git fetch origin first?17:56
<ajs6f>Ah, I had fetched the branch.
That did it.
Did you intentionally leave the JAXB stuff in there? And AbstractResource? It's not a problem, just wondering.17:58
Other than that, looks good to me.
<cbeer>yes, I think fasseg wants to use some of the JAXB stuff elsewhere (e.g. to parse the XML response)18:00
and there's a reason I left AbstractResource there, but I can't remember if it was a good one
maybe fcrepo-dc depended on it?18:01
<ajs6f>Oh, yeah. It probably does.
Those guys maybe should be factored out to a fcrepo-api-helpers or something like that.18:02
But anyway, looks cool to me.
<cbeer>cool18:03
let me get it in before barmintor makes me a merge conflict
<ajs6f>Say, whatever happened with the beer sponsorship? DId that come together in the end?
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/Emj6Nw
fcrepo4/master 1c97276 Chris Beer: Merge pull request #2 from futures/fcrepo-legacy-api...
* github-ff leaves
<cbeer>ajs6f: i think duraspace gave us some money18:06
* github-ff joins
[fcrepo4] eddies pushed 2 new commits to master: http://git.io/JVU_gA
fcrepo4/master 60a399e Edwin Shin: Ignoring some dependency details for the reporting plugin to speed up site generation, c.f.: http://jira.codehaus.org/browse/MPIR-137
fcrepo4/master 2d7ccbf Edwin Shin: Merge branch 'master' of github.com:futures/fcrepo4
* github-ff leaves
<cbeer>ajs6f: do you know if I can use java properties for the port number in, e.g.:
<jaxrs:server address="http://localhost:8080/rest" >
<ajs6f>Yes, but you might have to use SPring prpoperty placeholder for that.18:08
<cbeer>k
<ajs6f>Ben did it for fedora.home in 3.x, so he knows how to do it
<cbeer>completely right on all accounts.18:14
* github-ff joins18:15
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/TGrJxg
fcrepo4/master 1ccc9bb Chris Beer: allow the port used for the integration test server to be configured as a java property
* github-ff leaves
<ajs6f>So we're sticking with CXF?18:16
<cbeer>yes, barmintor (has almost) wired something in
<ajs6f>Awesome!
<cbeer>I couldn't figure out resteasy + maven + spring
<ajs6f>Fair enough.
<cbeer>and CXF does seem to be working
<ajs6f>I can't wit to see how he did it. That would be a really useful technique.
<cbeer>it's the same as fcrepo 3, i think
<ajs6f>I never knew how that worked, either. :)18:17
<cbeer>i think he was just struggling with the scala stuff
<ajs6f>Sorry.
I really didn't think it would be a big deal.
<cbeer>i remember seeing the error he was having
but i don't remember what i did to fix it
<ajs6f>It shouldn't have gone into the kernel.
Probably:
1) Install a Scala plugin for Eclipse.
2) Add the "Scala Nature" to the project.
<cbeer>(this was in plain old mvn)18:18
<ajs6f>Oh.
I never had a problem with the maven.
Weird.
Well, no matter.
As long as he's moving again.
* ajs6f leaves18:28
* github-ff joins18:29
[fcrepo4] eddies created gh-pages (+1 new commit): http://git.io/CyYLNw
fcrepo4/gh-pages fc1111f Edwin Shin: initial commit
* github-ff leaves
* github-ff joins18:31
[fcrepo4] eddies pushed 1 new commit to gh-pages: http://git.io/KA7VWg
fcrepo4/gh-pages b339867 Edwin Shin: Wagon: Deploying site to repository
* github-ff leaves
* github-ff joins
[fcrepo4] eddies pushed 1 new commit to gh-pages: http://git.io/MghVrA
fcrepo4/gh-pages 5c455d5 Edwin Shin: Wagon: Deploying site to repository
* github-ff leaves
* github-ff joins
[fcrepo4] eddies pushed 1 new commit to gh-pages: http://git.io/sUR73g
fcrepo4/gh-pages a8be507 Edwin Shin: Wagon: Deploying site to repository
* github-ff leaves
* github-ff joins18:32
[fcrepo4] eddies pushed 1 new commit to gh-pages: http://git.io/hNQ14A
fcrepo4/gh-pages 710c59d Edwin Shin: Wagon: Deploying site to repository
* github-ff leaves
* ajs6f joins18:33
* fasseg leaves18:38
<cbeer>ajs6f: ok, i'm just being stupid i think. so i made the commit to get spring to start the jax-rs server in the right place18:40
but
https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/test/java/org/fcrepo/AbstractResourceTest.java#L1118:41
is causing me trouble.
<ajs6f>How?
<cbeer>so, i'm settting the port i want to use via mvn -Dtest.port=8888 clean test
<ajs6f>I mean, do you want to parameterize that?
Oh yeah.
So we need to parameterize it.18:42
<cbeer>I tried to use System.getProperty
but it didn't work as i expected.
<ajs6f>Right.
NO?
Mm.
<cbeer>maybe i'm just doing something wrong
<ajs6f>Are the tests being run in a forked JVM?
<cbeer>i'm just running
$ mvn -Dtest.port=9988 clean test
anyway, we're having a process check-in.
<ajs6f>Shouldn't be, then.
Well, I'm cooking, so let me know if I can help with it.18:43
The port thing.
<cbeer>k
<ajs6f>Signing off for the evening. Stay warm in Chi-town, y;all.18:54
* ajs6f leaves
* github-ff joins18:59
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/D5vTQw
fcrepo4/master feef2ae Chris Beer: use the test.port in the AbstractResourceTest too
* github-ff leaves
* eddies leaves19:10
* eddies joins19:12
* eddies leaves
* eddies joins
* JasonDGI leaves19:32
* eddies leaves22:22
* eddies joins22:23
* eddies leaves
* eddies joins
* github-ff joins22:36
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/f0SarQ
fcrepo4/master f6c0548 Chris Beer: and in the DublinCore test (which should be DRY-ed out..s
* github-ff leaves
<cbeer>root@futures6:/opt/fedora# rm -rf derby data; server tomcat6 restart22:43