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

Using timezone: Eastern Standard Time
* nbanks joins00:08
* nbanks leaves00:13
* nbanks joins02:00
* nbanks leaves02:04
<pivotal-bot>Chris Beer started "HATEOAS RDF response for object/datastream describe endpoints" https://www.pivotaltracker.com/story/show/4891927102:14
* nbanks joins02:30
* nbanks leaves02:34
* jotudin joins03:17
* nbanks joins04:31
* nbanks leaves04:36
* nbanks joins06:02
* nbanks leaves06:06
* eddies leaves06:08
* eddies joins06:31
* eddies leaves
* eddies joins
* eddies leaves07:08
* eddies joins07:41
* eddies leaves
* eddies joins
* nbanks joins08:02
* nbanks leaves08:07
* ajs6f joins09:26
cbeer: Thanks for doing that merge ysterday.09:27
<pivotal-bot>A. "Bupholutac" Soroka added comment: "I just get a 500 at that address..." https://www.pivotaltracker.com/story/show/4988262709:29
* nbanks joins09:36
* nbanks leaves09:42
<pivotal-bot>Edwin Shin added comment: "Some more reading: http://wiki.nikhef.nl/grid/CLARIN/OAuth2_use_case" https://www.pivotaltracker.com/story/show/4545824909:56
Edwin Shin added comment: "And I suppose a filter might be the most flexible, out-of-box approach? e.g., http://cxf.apache.org/docs/jax-..." https://www.pivotaltracker.com/story/show/4545824909:57
* ajs6f leaves10:08
* ajs6f joins10:27
Hm. Is http://fcrepo4.fcrepo.org/ down?10:35
I'm getting 500s for everything.10:36
* escowles joins
<pivotal-bot>A. "Bupholutac" Soroka added comment: "For the fcrepo3 predicates without obvious analogs in fcrepo4, I'm happy to just put them in some..." https://www.pivotaltracker.com/story/show/4988262710:41
A. "Bupholutac" Soroka started "Declare some fcrepo3 predicate names for migrating object and datastream properties" https://www.pivotaltracker.com/story/show/49882627
Edwin Shin accepted "Add RDF output to field search api" https://www.pivotaltracker.com/story/show/4915602310:43
Edwin Shin accepted "RDF response for describe repository" https://www.pivotaltracker.com/story/show/4986404710:45
<eddies>ajs6f: have a min?10:46
<ajs6f>I'm on my way back to my office for standup— a min for IRC, or for voice?
<eddies>voice
<ajs6f>Not until I get back to office: I'll try to hurry.10:47
<eddies>k. just find on hangouts
*find me
<ajs6f>Our usual hangout?
<eddies>yah
<ajs6f>cool.
<bljenkins>Project fcrepo-kitchen-sink build #300: SUCCESS in 5 min 37 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/300/10:55
<cbeer>ajs6f: SEVERE: The provider class, class org.fcrepo.responses.HtmlProvider, could not be instantiated. Processing will continue but the class will not be utilized10:56
java.lang.NullPointerException
<ajs6f>ARG!
<cbeer>i wonder why
<ajs6f>Me, too.
Is there a stacktrace?
<cbeer> at org.fcrepo.responses.HtmlProvider.init(HtmlProvider.java:100)
* gregjansen joins
<ajs6f>Hm. I wonder if I missed something between fcrepo-webapp and kitch-sink?10:57
<cbeer>ajs6f: component scanning for org.fcrepo.responses is probably important?10:58
* ajs6f leaves
* github-ff joins11:00
[fcrepo-kitchen-sink] cbeer pushed 1 new commit to master: http://git.io/yLwBxA
fcrepo-kitchen-sink/master d78b148 Chris Beer: add org.fcrepo.responses to component-scan
* github-ff leaves
* ajs6f joins11:03
<bljenkins>Project fcrepo-kitchen-sink build #301: SUCCESS in 6 min 14 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/301/11:06
<cbeer>http://fcrepo4.fcrepo.org/fcrepo/rest/
ajs6f: ^11:07
oh, you missed it on your way out:
i said earlier:
"ajs6f: component scanning for org.fcrepo.responses is probably important?"
<ajs6f>cbeer++
Way to remember what I forgot.
That Which I Should Have Done I Did Not Do, but cbeer does it anyway.11:08
<cbeer>ajs6f: eddies and i also wonder if, we're indeed using spring for the fcrepo alpha, we should probably reorganize our spring configs to be sensible.
<ajs6f>Yes, we should have done that a while ago. :)
We can kill the generator-rdf module to remove some tricky Spring XML.
The DC generator can use @Autowire to simplify the Spring.11:09
Or maybe we don't need DC gen if the RDF is good enough for harvesting?
<pivotal-bot>Chris Beer edited "HATEOAS RDF response for object/datastream describe endpoints" https://www.pivotaltracker.com/story/show/4891927111:14
* github-ff joins11:16
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/5y5ghw
fcrepo4/master 59e443d Chris Beer: add some more structured data to the HTML output
* github-ff leaves
<cbeer>ajs6f: what happened to barmintor's info:fedora => http:// converter?11:17
<ajs6f>Eh?
We have one?11:18
<cbeer>https://github.com/futures/fcrepo4/commit/32c9622d69615f43c9c5b2680e1db8b74bb8bd2a
ah, i see11:19
<ajs6f>I didn't even know that what there! Shucks, my fault. We should be able to add that back into the two providers I added.11:21
Or rather, back into the JAX-RS methods.
That seems to be how it gets integrated, right?11:22
We're getting a lot of duplication in the RDF triples. I wonder if we can cut that out by using owl:sameAs to link the various identifiers that are appearing for a given node.11:24
Hm.
But I don't want to maintain a co-reference store.
<cbeer>escowles:
https://github.com/futures/fcrepo4/tree/master/fcrepo-http-commons/src/main/resources/views
https://github.com/futures/fcrepo4/blob/master/fcrepo-http-commons/src/main/java/org/fcrepo/responses/HtmlProvider.java
https://github.com/futures/fcrepo4/blob/master/fcrepo-http-api/src/main/java/org/fcrepo/api/FedoraNodes.java#L77
https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/utils/JcrRdfTools.java11:25
https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/FedoraResource.java#L190
there. i think that's the important bits
<escowles>cbeer: thanks!
<cbeer>and i'm going to fix up ajs6f's change so you'll get HTTP URIs for resources11:26
<ajs6f>cbeer++11:27
<cbeer>let's just serialize it as RDF and use inferencing!11:37
* travis-ci joins11:38
[travis-ci] futures/fcrepo4#572 (master - 59e443d : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/1e0008c9a9d4...59e443d19575
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7325711
* travis-ci leaves
* github-ff joins11:39
[fcrepo4] cbeer pushed 2 new commits to master: http://git.io/nkUObA
fcrepo4/master 278a4e4 Chris Beer: make sure describe endpoint uses info => http translation
fcrepo4/master 08342f9 Chris Beer: re-add GraphSubjects translator to http endpoints
* github-ff leaves
<cbeer>escowles: that latest commit should have the right URIs again.. whenever github gets around to telling us about it
fcrepo4: we can screw things up, but we'll try not to!11:44
<ajs6f>We'll _try_ to try.
* gregjansen leaves11:48
* gregjansen joins11:49
<bljenkins>Project fcrepo4 build #675: UNSTABLE in 15 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/675/11:55
* Chris Beer: make sure describe endpoint uses info => http translation
* Chris Beer: re-add GraphSubjects translator to http endpoints
<eddies><dependency>11:57
<groupId>org.apache.amber</groupId>
<artifactId>amber-oauth2-resourceserver</artifactId>
<version>0.22-incubating</version>
</dependency>
https://cwiki.apache.org/confluence/display/OLTU/OAuth+2.0+Resource+Server11:58
<ajs6f>http://shiro.apache.org/11:59
* travis-ci joins12:00
[travis-ci] futures/fcrepo4#573 (master - 08342f9 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/59e443d19575...08342f99b89a
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7326368
* travis-ci leaves
<ajs6f>https://github.com/sdorra/shiro-oauth12:01
<pivotal-bot>Edwin Shin edited "Add an OAuth service to the REST API's JAX-RS container" https://www.pivotaltracker.com/story/show/4545824912:08
Edwin Shin edited "Add an OAuth service to the REST API's JAX-RS container" https://www.pivotaltracker.com/story/show/45458249
<ajs6f>http://tools.ietf.org/html/rfc674912:17
Fun.12:18
77 pages of fun.
Well, it's still simpler than Shib.
<pivotal-bot>A. "Ternuculious" Soroka started "Add an OAuth service to the REST API's JAX-RS container" https://www.pivotaltracker.com/story/show/4545824912:19
Edwin Shin edited "Add an OAuth service to the REST API's JAX-RS container" https://www.pivotaltracker.com/story/show/4545824912:21
Edwin Shin added comment: "This was a good, general read: http://aaronparecki.com/articles/2012/07/29/1/oauth2-simplified" https://www.pivotaltracker.com/story/show/45458249
Edwin Shin added comment: "Some code examples for Spring: https://github.com/SpringSource/spring-security-oauth/tree/master/samples/oaut..." https://www.pivotaltracker.com/story/show/4545824912:23
Edwin Shin edited "Add an OAuth2 service to the REST API's JAX-RS container" https://www.pivotaltracker.com/story/show/4545824912:24
* eddies leaves12:34
* eddies joins12:35
* eddies leaves
* eddies joins
<cbeer>ajs6f: what's the story on https://www.pivotaltracker.com/story/show/49205649 again?12:40
<pivotal-bot>feature: Inject authenticated sessions into the JAX-RS resources (finished) / owner: A. "Ternuculious" Soroka
Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/63" https://www.pivotaltracker.com/story/show/4986661112:41
<ajs6f>Waiting on Nigel to tell Eddie how long to wait.12:46
<cbeer>wait for what?12:47
<ajs6f>For him to merge his branch, which subsumes mine.
* ajs6f leaves13:38
* ajs6f joins13:42
* ajs6f leaves
* github-ff joins13:43
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/U6Nj1Q
fcrepo4/master d40943b Chris Beer: update GraphStoreStreamingOutput to accept several flavors of Jena data.
* github-ff leaves
* ajs6f joins
<bljenkins>Yippie, build fixed!13:57
Project fcrepo4 build #676: FIXED in 13 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/676/
Chris Beer: update GraphStoreStreamingOutput to accept several flavors of Jena data.
Project fcrepo-kitchen-sink build #302: SUCCESS in 3 min 52 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/302/14:02
* travis-ci joins14:04
[travis-ci] futures/fcrepo4#574 (master - d40943b : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/08342f99b89a...d40943bb2e71
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/733018614:05
* travis-ci leaves
* ajs6f leaves14:06
* ajs6f joins14:12
<cbeer>ajs6f: does "Unable to find artifact. Could not find artifact org.jboss.as:jboss-eap:zip:6.1.0.Beta in jboss-eap-techpreview (http://maven.repository.redhat.com/techpreview/eap6/6.1.0.Beta1/maven-repository)" mean anything to you?14:17
before i ask in #modeshape..
<ajs6f>It does not, in the context of fcrepo. Are you building MODE?14:18
<cbeer>yeah, trying to run their IT14:19
<ajs6f>Erm. Could be a network hiccup, because I've never seen that. But then, I'm not sure I ever actually ran all their ITs… Do we do that as part of our MODE CI build?
<cbeer>i don't think so14:20
<ajs6f>Anyway, probably best to ask them. It doesn't ring any tubular bells for me. Not even a glockenspiel.
<cbeer>oh. hm. looks like i can download that file from redhat.. behind some kind of EULA14:22
this is beta software, don't come crying to us, etc, etc.
<ajs6f>Well, that's not unreasonable, although kind of annoying.14:23
After all, they're in the business of selling support.
rhauch brought that up as an issue as to why MODE hasn't been releasing snapshots— they don't want to spend time answering questions from goons who put beta gear into production.
(I don't know anyone who does that. :))14:26
<cbeer>yeah.14:28
it'd just be nice if they documented some of this
<ajs6f>Yes, maybe we can push them to open the dev process a little more. They seem like pretty open-minded types.14:29
Unlike those weird, secretive, paranoid fcrepo folks.
* github-ff joins14:31
[fcrepo4] ajs6f created SpringOAuth from master (+0 new commits): http://git.io/Lzm4kQ
* github-ff leaves
<pivotal-bot>Chris Beer started "Fix up legacy API and re-add it to the kitchen sink" https://www.pivotaltracker.com/story/show/5013814514:38
Chris Beer estimated "Move JAX-B models from -commons to legacy API as appropriate" as 1 point https://www.pivotaltracker.com/story/show/49930425
Chris Beer started "Move JAX-B models from -commons to legacy API as appropriate" https://www.pivotaltracker.com/story/show/49930425
<ajs6f>cbeer: I think _all_ of the JAXB should go to legacy, unless people hold a gun to our heads and demand XML output forever!14:39
I put them in commons, and I should never have done that
.
<cbeer>agreed. i suspect we still have some dependencies on them in the http api though14:40
e.g. the namespaces output
which we need to decide what to do with
<ajs6f>Oh, right. Fair enough. We ought to kill 'em, is what we ought to do. Take namespaces— we can deliver that in RDF, no prob.
Maybe "namespaces" should be a subresource of "repository"?14:41
Like you get the global description from repository, and a piece of it from namespaces, and interact with it there?14:42
Or the description in repository doesn't directly include the namespaces, just a ink to the namespaces endpoint? That's HATEOSy.
ish.
afk bbi514:43
* ajs6f leaves
* ajs6f joins14:47
<bljenkins>Project fcrepo-kitchen-sink build #303: SUCCESS in 4 min 12 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/303/
<ajs6f>back
<cbeer>ajs6f: well, right now, every RDF response includes every registered namespace14:50
because i'm too lazy to do anything else
<ajs6f>You mean every response to every endpoint?
<cbeer>but that doesn't help you pre-register namespaces with specific prefixes14:51
any RDF serialization includes the whole JCR namespace registry mapping.
<ajs6f>Oh, yeah. We should tweak Jena or something about that.
I was thinking that namespaces endpoint would supply an RDF serialization of the namespace registry, which you could adjust with the same SPARQL Update type deal we use for other endpoints.14:52
That's "verbose" but uniform across endpoints.
<cbeer>hm. that could work
<whatever namespace uri> <#hasPrefix> "whatever"14:53
<ajs6f>Yeah, something like that.
I don't know if there is any more structure in the registry than prefix->uri, but if so (and if it's useful) we could expose it with additional triples.
<cbeer>nope. just that afaik.14:54
<ajs6f>Easy enough, then.
Actually, do all the RDF serializations support prefixes? If so, you could just emit prefix declarations. No triples! But maybe that would just be confusing.14:55
And it's not clear how to remove a prefix mapping, then.
afk bbi1015:06
* ajs6f leaves
* ajs6f joins15:10
<cbeer>no, they don't.15:13
* gregjansen leaves15:14
<ajs6f>Who don't what?15:17
<cbeer>"do all the RDF serializations support prefixes?"
<ajs6f>Oh, well, that was kind of a joke-y thought anyway.
I think triples are fine.15:18
I think the benefit to namespaces via RDF is uniformity.
<pivotal-bot>Chris Beer started "fix addJcrContentLocationInformationToModel so it takes the jcr:content node instead of the parent node" https://www.pivotaltracker.com/story/show/5013804515:38
Chris Beer finished "fix addJcrContentLocationInformationToModel so it takes the jcr:content node instead of the parent node" https://www.pivotaltracker.com/story/show/50138045
Chris Beer added comment: "https://github.com/futures/fcrepo4/commit/7b4e883228468e3daf1506eaf46eee469acf8d8f" https://www.pivotaltracker.com/story/show/5013804515:39
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/SxxVVQ
fcrepo4/master 7b4e883 Chris Beer: update getLowLevelCacheEntries to take the jcr:content node or jcr:data property, instead of the parent node.
* github-ff leaves
<pivotal-bot>Chris Beer accepted "fix addJcrContentLocationInformationToModel so it takes the jcr:content node instead of the parent node" https://www.pivotaltracker.com/story/show/50138045
<cbeer>that could use more cleanup.. like actually separating the high-level fcrepo stuff from the low-level utils.15:40
<ajs6f>We have a happenstance attitude towards utils right now. they just pop up wherever they pop up.
<cbeer>yeah. they're easy enough to clean up later15:41
<ajs6f>I'm not too worried. It would be nice to have some kind of policy about it, tho'.15:42
<cbeer>factory bean factories for everyone!
<ajs6f>If we wrap bean factories in container factories, do we get taco factories?15:43
* travis-ci joins16:02
[travis-ci] futures/fcrepo4#575 (master - 7b4e883 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/d40943bb2e71...7b4e88322846
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7333392
* travis-ci leaves
<escowles>i'm seeing something weird in the HTML generation: it's consistently picking the wrong velocity template to use16:04
i think this is because the rdf graph also includes the child nodes, with multiple primaryType triples, so getFirstValueForPredicate is finding the values from child nodes16:06
<ajs6f>Urg: Thanks Spring:16:12
java.lang.IncompatibleClassChangeError: class org.springframework.core.type.classreading.ClassMetadataReadingVisitor has interface org.springframework.asm.ClassVisitor as super class
That's really helpful.
escowles: You can change in… hang on, let me find it.16:13
org.fcrepo.responses.HtmlProvider.writeTo() at lne...16:14
156.
I'm sorry I didn't catch that.
On the other hand, it's not clear to me how you'll know which is the "primary" node.
It's just a bag of triples.16:15
* jonathangee leaves
<escowles>i don't see how to figure it out -- we don't have the "main" subject to use that
<ajs6f>We may have to get clever and walk the graph.
Like look for the URI that isn't a child of anything else,
and takes _its_ primaryType.
<cbeer>not going to work.
<ajs6f>Why?
<cbeer>i include the parent in the graph dump16:16
<escowles>yes, we could look at the relationships to exclude some things
<cbeer>do we have the current URI in that context?
<escowles>cbeer: i don't think so
<ajs6f>We have only and exactly what the JAX-RS method puts in the RDF.
But we _could_ put it in there.
<cbeer>back when i was doing this, i wondered if we needed some kind of "this document is about this subject!" pointer
<ajs6f>Exactly!
<cbeer>but that seems kinda hacky
<escowles>yes, that would be helpful -- i could add that in FedoraNodes16:17
<cbeer>but maybe that's httprange14 all over again?
<ajs6f>I dunno. I think it's pretty standard tech.
Yeah, it gets into that. But we can take comfort in
the fact that we're doing RDF, not Linked Data.
We're passing RDF arround inside our system.
What we expose to the world via HTTP is something else.
(It's exactly what comes out of the RDFProvider.)16:18
You know, there's another way we could do it.16:20
With a named graph.
We're using the default graph now.
We could name the graph with the URI of the thing it describes.
I vote for that.16:21
<escowles>we do something similar at UCSD -- each object's metadata is in its own named graph -- we do that for retrieval efficiency, but it also gives you the name in all contexts16:22
<ajs6f>I guess that wold mean altering the method in FedoraResource that produces that Jena construct?
cbeer: +/-1?
<cbeer>-0. it's +1 to use named graphs internally.. i'm afraid of how to parse them externally though.16:23
but maybe it's unavoidable
this is at least the 2nd time we've said wouldn't named graphs be nice.
<ajs6f>That's true. As for the ext/int thing, I'm not at all suggesting we expose it.
<cbeer>we could always wrap the jena dataset in an fcrepo construct too16:24
<ajs6f>I think we pass one named graph around internally, and RDFGenerator serializes it as a default graph.
<cbeer>not saying that's any good.
<ajs6f>We could. Would it be like a "graph with metadata"?
<cbeer>yeah, something.
which is also a little silly
or..
<ajs6f>Clojure has metadata on data structures. I always thought that was cool, but dangerous.16:25
<cbeer>serialize properties to the default graph, but add another graph for request context?
i'm kinda doing that on field search anyway
<ajs6f>And produce both at the endpoint?
Or just pass them arond internally?
<cbeer>i'd be happy exposing it externally, why not?
it solves the same problem downstream.
<ajs6f>Because RDF/XML can't do that. No named graphs.16:26
<cbeer>"how do i know what i have"
<escowles>cbeer: would you like a side of metadata with your metadata?
<ajs6f>And I'm not sure how you do that in RDFa, either.
<cbeer>yes, but instead of metadata, can i substitute some fries?
<ajs6f>We heard you liked metadata, so we… aw, you know the rest.
I'm leaning towards two graphs passed around internally, deafult with the real metadata and a named graph with context, and one priduced externally, with just the real metadata.16:27
You know what resource its about because you asked for that resource's metadata.
(As a client.)16:28
(BTW, sorry Esme. I should have written more integration tests and I would have caught this.)
<escowles>ajs6f: not a problem -- though i was really confused when /rest/ worked but /rest/objects/ didn't (becuase it couldn't find the template for mode:root)16:29
<ajs6f>Oh, I don't think I even made a tempalte for mode:root. I forgot that that special node type exists.16:30
<cbeer>i think mode:root showed up after you did the work
because of that path-parsing bug16:31
<escowles>so, should we have a special named graph (with the same named graph across all resources) that holds the current resource's URI?16:33
<cbeer>as long as that isn't totally crazy.16:34
<escowles>i think it's only marginally crazy
<cbeer>there's probably other information that'd make sense to reflect in there too
<ajs6f>I like it. So the name of the named graph tells you what resource with which you are conerned,16:35
<cbeer>like i said, certainly on field search
<ajs6f>andthe contetnts tell you other context.
<cbeer>but could also include the fcrepo base uri, maybe?
<ajs6f>Sure. Whatever we think will be helpful.
Anything from the HTTP world that might be useful, for example.
(Like the base uri)
I could even see allowing people to pass in a base uri (like to get a "view" of the data as housed other than at the repo machine) for use with webapps through which they are accessing Fedora. But maybe that's a little crazy.16:37
* github-ff joins
[fcrepo-kitchen-sink] cbeer pushed 1 new commit to master: http://git.io/6i0b3g
fcrepo-kitchen-sink/master 7711b81 Chris Beer: Revert "remove legacy api"
* github-ff leaves
<ajs6f>Anyway, this should get us everything we need at temaplting to pick the right template, right?
<escowles>yes, i think so -- once we have the current URI, then it will be easy to pick the correct primaryType statement to use16:38
<ajs6f>Cool.
cbeer: am I right in think this is mostly about FedoraResource and the method that produces the roginal graph(s)?16:39
and then just passing it along through JAX-RS (annotating it with base uri on the way)
so that it can be used by the Providers appropriately?
<escowles>i'm going to be out picking up my kids & having dinner, so i'll check back in to see where this is in an hour or so16:40
<ajs6f>I'm going home for the night, so I'm not taking much further.
(Today, that is)
<cbeer>i think this change can happen in the http-api
it gets a Jena GraphStore
and just needs a way to serialize a Request to some triples
<ajs6f>Okay. That sounds easy. Shall I reopen the ticket for which I did the providers and continue work from there? Or do we want a new tickets?16:41
<cbeer>i guess we'd have to mess with FedoraResource if we're serializing the JCR stuff into a named graph
nah, that's a new ticket.
<ajs6f>Okay.
<cbeer>i'm happy to take a stab at it after this JAX-B thing
<ajs6f>Whoa— what's this about the JCR stuff?
You mean segregating JCR and Fedora triples?16:42
<cbeer>no. i'm saying -kernel shouldn't care about this named graph of http properties
<ajs6f>(Meaning triples with predicates from those different namespaces)?
Oh, okay.
<cbeer>it's totally between the JAX-RS resource and the provider
<ajs6f>Oh, right. You're 100% right.
It's an API concern.
(HTTP API.)
Do we want to stick the whole Request in there?16:43
Or just base uri, maybe a few other selected things.
<cbeer>i don't think we need it all
<ajs6f>Me niether.16:44
So far we know we want base uri.. anythign else?
MIMEtype is already getting passed via JAX-RS machinery.
We had talked about a "recursion" param.
But we don't have one yet.16:45
I don't think we'd want it in here anyway.
<cbeer>mimetype probably isn't important.. but maybe available mime types?
it'd be nice to have a way to find out what flavors are available
<ajs6f>Isn't that decided by Jena?
Jena and the fact that we can produce HTML.16:46
Does it vary on a resource-to-resource basis?
(or a request-to-request basis, for that matter.)
<cbeer>just what we've declared @Produces on.16:51
<ajs6f>Right.16:53
So do we need to include tthat in the context?
<cbeer>our decisions were somewhat arbitrary
<ajs6f>True.
<cbeer>and mime types for these formats are somewhat arbitrary.
<ajs6f>True.
<cbeer>whether we need to provide that on every request is debatable, sure.
but it's something that we could expose.. somehow.
<ajs6f>Doesn't it come back from an HTTP OPTIONS request, or is that just the allowable methods?16:54
<cbeer>ooh. maybe it does
<ajs6f>I dunno. I could be getting that wrong.
Lemm check.
http://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html16:55
Doesn't say. Urg.
Well, we could certainly return that as headers in the response to OPTIONS16:56
I wonder what practice is out theere in the Web for this?
* github-ff joins16:57
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/a7xINg
fcrepo4/master 9b2d029 Chris Beer: yank a bunch of unused JAX-B models out of http-commons
* github-ff leaves
<ajs6f>Certainly seems like something you'd want to be able to find out from an endpoint (HATEOS and all).
<bljenkins>Project fcrepo-kitchen-sink build #304: FAILURE in 1 hr 1 min: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/304/16:58
<cbeer>ajs6f: ew. you get wadl back using OPTIONS17:00
<ajs6f>BLORP!?
From Jersey itself, I guess?
Cause we didn't do that.
<bljenkins>Project fcrepo-kitchen-sink build #305: NOW UNSTABLE in 2 min 52 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/305/17:01
<ajs6f>Actually, looking around the web, it looks like a lot of people do return WADL for OPTIONS, I guess since HTTP doesn't define a message body for OPTIONS.17:02
http://books.google.com/books?id=XUaErakHsoAC&lpg=PA98&ots=5jdlHplGqC&dq=RESTful%20web%20services%20OPTIONS%20method%20response&pg=PA98#v=onepage&q=RESTful%20web%20services%20OPTIONS%20method%20response&f=false17:04
says that there just is no standard.
I'm outta here. See y'all tomorrow.17:07
* ajs6f leaves
<pivotal-bot>Chris Beer finished "Move JAX-B models from -commons to legacy API as appropriate" https://www.pivotaltracker.com/story/show/4993042517:13
Chris Beer added comment: "https://github.com/futures/fcrepo4/commit/9b2d02946c6128271031f3f4cecb611b870e2902" https://www.pivotaltracker.com/story/show/49930425
Chris Beer added "Add HTTP context information to RDF responses as a named graph" https://www.pivotaltracker.com/story/show/5022033917:15
Chris Beer estimated "Add HTTP context information to RDF responses as a named graph" as 1 point https://www.pivotaltracker.com/story/show/50220339
Chris Beer started "Add HTTP context information to RDF responses as a named graph" https://www.pivotaltracker.com/story/show/50220339
<bljenkins>Yippie, build fixed!17:16
Project fcrepo-kitchen-sink build #306: FIXED in 3 min 58 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/306/
* travis-ci joins17:19
[travis-ci] futures/fcrepo4#576 (master - 9b2d029 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/7b4e88322846...9b2d02946c61
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7336495
* travis-ci leaves
<cbeer>ok, i'm backing away from adding a named graph about context, i think.17:40
because it's then a matter of what to call things..
mainly the graph and the subject within the graph
<escowles>cbeer: do we need the triples for the children? if we only had the triples for the object we were describing, then we wouldn't have this problem18:00
<cbeer>escowles: yes, i think so.18:16
i think it's common to want the datastreams for an object
or, if you have a datastream, the fixity information is actually on the child fcr:content node18:17
<pivotal-bot>Chris Beer added "Replace fixity JAX-B serialization with RDF" https://www.pivotaltracker.com/story/show/5022821919:02
Chris Beer started "Replace fixity JAX-B serialization with RDF" https://www.pivotaltracker.com/story/show/50228219
Chris Beer unstarted "Add HTTP context information to RDF responses as a named graph" https://www.pivotaltracker.com/story/show/5022033919:33
* github-ff joins21:40
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/qGGxcA
fcrepo4/master 99f67f7 Chris Beer: replace fixity JAX-B response with RDF; move fixity work up to DatastreamService, and add a RDF serialization for fixity results.
* github-ff leaves
<pivotal-bot>Chris Beer finished "Replace fixity JAX-B serialization with RDF" https://www.pivotaltracker.com/story/show/50228219
Chris Beer estimated "Replace fixity JAX-B serialization with RDF" as 2 points https://www.pivotaltracker.com/story/show/50228219
Chris Beer added comment: "https://github.com/futures/fcrepo4/commit/99f67f7b85e1612ae766d0507e4f14b84c44fd17" https://www.pivotaltracker.com/story/show/50228219
Chris Beer delivered "Replace fixity JAX-B serialization with RDF" https://www.pivotaltracker.com/story/show/50228219
Chris Beer added "Update FedoraIdentifiers to remove JAX-B dependency" https://www.pivotaltracker.com/story/show/5023433121:49
Chris Beer started "Update FedoraIdentifiers to remove JAX-B dependency" https://www.pivotaltracker.com/story/show/50234331
* travis-ci joins22:01
[travis-ci] futures/fcrepo4#577 (master - 99f67f7 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/9b2d02946c61...99f67f7b85e1
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/7343029
* travis-ci leaves
<bljenkins>Project fcrepo-kitchen-sink build #307: SUCCESS in 11 min: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/307/22:29