Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* whikloj leaves00:31
* whikloj joins01:41
* ksclarke leaves01:45
* whikloj leaves01:46
* dhlamb_ joins05:31
* dhlamb leaves05:33
* dhlamb joins07:12
* dhlamb_ leaves07:15
* dhlamb_ joins07:49
* dhlamb leaves07:52
* dhlamb_ leaves07:56
* ajs6f joins08:14
escowles: ping08:17
<escowles>ajs6f: gotta take the kids to the bus stop -- but i'll be back in ~10 mins08:29
<ajs6f>escowles: Cool— we just sat down to breakfast. Let's catch up later this morning.08:31
* bseeger joins08:37
* acoburn joins08:48
* dhlamb joins08:50
<acoburn>awoods: ping09:02
* whikloj joins09:05
* dwilcox joins09:12
* dhlamb leaves09:21
* dhlamb joins09:24
<acoburn>awoods: I'm working on fcrepo-1791 (webac on binary resources)
awoods: the issue is I can't seem to find the magic incantation to get the right triples09:29
awoods: e.g. if I have a container: /rest/foo
awoods: and a file: /rest/foo/bar
awoods: the file is of type NonRdfSourceDescription09:30
awoods: (/rest/foo/bar)
awoods: but where do the triples live?
awoods: If I call getTriples() with PropertiesRdfContext, I don't get the right ones09:31
awoods: do I need to call getDescribedResource and fetch the triples on that?
awoods: that, at least, doesn't seem to work09:32
awoods: through the HTTP api, /rest/foo/bar/fcr:metadata provides the triples I want09:33
* ksclarke joins
<acoburn>awoods: I've also tried permutations with ContentRdfContext on the original/described resource, but no luck
<awoods>acoburn: the triples are stored on the NonRdfSourceDescription... see this ticket:09:34
acoburn: https://jira.duraspace.org/browse/FCREPO-164009:36
<acoburn>awoods: thanks09:37
<awoods>acoburn: https://github.com/fcrepo4/fcrepo4/pull/864/files#diff-d372035f89c8d0846920e20bd13e0145R195
* peichman joins
<acoburn>awoods: that's very helpful09:38
<awoods>acoburn: and I stand corrected: the properties are on the actual RdfSource
* esm_ joins09:39
<acoburn>awoods: isn't it the case that resource.getDescribedResource() returns the FedoraBinary?09:40
* mikeAtUVa joins
<awoods>acoburn: yes09:41
<acoburn>awoods: and given your comment above about the RdfSource, are you saying that the triples are *not* stored on that FedoraBinary object?
<awoods>acoburn: I meant to say: NonRdfSource09:42
i.e. the real thing, not the description resource
<acoburn>awoods: got it09:43
<awoods>acoburn: this code-shuffling between binaries and their descriptions is also the source of these two (duplicative) tickets:09:44
<ajs6f>awoods: Is someone working those tix? They seem like #1 priority?09:45
<awoods>ajs6f: We discussed them on yesterday's tech call, but no one claimed them.09:46
ajs6f: acoburn is on the webac sprint, but maybe escowles will have time for these next week.09:47
<ajs6f>awoods: K. I might have time for them next week, too.09:48
* ksclarke leaves
<awoods>ajs6f: that would be wonderful and helpful.
<escowles>ajs6f: awoods: yes, 1803/1753 seems like a big problem (and probably easy to fix)
<awoods>ajs6f: It sounded like you and escowles were going to be collaborating on resource triples (i.e. API spec) next week as well.09:49
* ksclarke joins
<ajs6f>awoods: Yes, but I honestly don't think that's going to take much time. There are only a few endpoints and I think the work will mostly be spotting JCR stuff and marking it for death.09:50
<awoods>ajs6f: if it does not take much of your time, that is a plus... I value your play time.09:51
<ajs6f>awoods: Working on Fedora is never play. Play is pleasant. Play is something to which to look forward. Play makes one happy.09:53
<awoods>ajs6f: hence my joy in hearing that the work of killing JCR stuff will be brief, allowing more time for you to run around and play.09:55
<ajs6f>awoods: Thank you for your caring and considerate attitude.
* umgrosscol joins09:58
* escowles leaves10:00
* escowles joins
* petertr7 joins10:39
* mohamedar joins10:46
<awoods>ruebot: is this ticket on your radar: https://jira.duraspace.org/browse/FCREPO-1802 ? good morning.
* Code joins10:50
<ruebot>awoods: it totally is. today is release day for islandora... if we get this blocker issue resolved.10:52
awoods: i had planned on giving it a lot of time next week.
<awoods>ruebot: thanks... just checking, as I had not heard from you on that ticket. Enjoy the release.10:53
<ruebot>awoods: diegopino and acoburn had some good conversations about it in #islandora the other day as well.
* github-ff joins
[fcrepo-module-auth-webac] whikloj opened pull request #42: Fix 404 for non-admin when requesting versioned resources. (master...fcrepo-1760) http://git.io/vlLgN
* github-ff leaves
<whikloj>awoods/acoburn: ^^ I feel very stupid today10:54
<acoburn>whikloj: I'm sure it's nothing like I feel w/r/t accessing triples stored on binary resources. I clearly don't understand something.10:55
<whikloj>acoburn: perhaps, but i spent 2 days fighting and then "Hmm why does the debug output look different. Oh getString != toString"10:56
acoburn: Regarding adding a Link header, if it is currently inaccessibleResource for non-admins. Will we need to change whatever you and awoods did for that?10:57
<acoburn>whikloj: and I'm like: why can't I even get any debug output? it's just crickets chirping10:58
<whikloj>acoburn: hahahaha, I totally understand
acoburn: are we meeting today or just IRC standup?
<awoods>whikloj: hangout
<acoburn>whikloj: meeting, and we should talk about link headers
* chadmills joins11:06
* Code leaves11:15
* travis-ci joins11:26
jrgriffiniii/fcrepo4-client#2 (fcrepo-1443-jrgriffiniii - 0030789 : griffinj@lafayette.edu): The build passed.
Change view : https://github.com/jrgriffiniii/fcrepo4-client/compare/31aa509dbd0f...003078997d9d
Build details : https://travis-ci.org/jrgriffiniii/fcrepo4-client/builds/88360814
* travis-ci leaves
* github-ff joins11:29
[fcrepo4-client] jrgriffiniii opened pull request #34: FCREPO-1443 fcrepo-client should use Link rel="describedby" header for description (master...fcrepo-1443-jrgriffiniii) http://git.io/vltvA
* github-ff leaves
* travis-ci joins11:31
jrgriffiniii/fcrepo4-client#3 (fcrepo-1443-jrgriffiniii - 940a10e : griffinj@lafayette.edu): The build passed.
Change view : https://github.com/jrgriffiniii/fcrepo4-client/compare/003078997d9d...940a10e4c0e7
Build details : https://travis-ci.org/jrgriffiniii/fcrepo4-client/builds/88362915
* travis-ci leaves
* chadmills leaves11:39
* bseeger leaves11:54
* umgrosscol leaves12:03
* github-ff joins12:04
[fcrepo4] peichman-umd opened pull request #931: Allow an admin user to specify a delegated user. (master...fcrepo-1790) http://git.io/vltKI
* github-ff leaves
<peichman>@awoods/@acoburn/@whikloj: PR opened for 1790, I will be back at my desk around 2pm12:05
<acoburn>peichman: thanks
* osmandin joins12:39
* peichman1 joins13:00
* peichman leaves
* umgrosscol joins13:04
* pmurray joins
* bseeger joins13:20
* umgrosscol leaves13:23
* apb18 joins13:56
* peichman1 leaves13:58
* peichman joins14:00
* esm_ leaves14:02
* esm_ joins
* rdfloyd joins14:10
<dhlamb>apb18: not sure how you could do that without a reverse proxy?14:14
<acoburn>apb18: fcrepo-transform is a good example of that (module built into fedora)14:15
<awoods>acoburn: once we can drop in fcrepo-transform at runtime... that will be an exciting pattern.
* Dan_Davis joins14:18
<esm_>ping apb
<awoods>apb18: you there?
<dhlamb>id's or json!14:24
<peichman>FYI for WebAC sprint folks, I have started working on a "Quick Start with WebAC" README style guide14:30
* osmandin leaves14:31
<dhlamb>awoods: exactly, design begins with feature extraction from use cases14:41
* umgrosscol joins14:48
<peichman>I am running into some behavior I wasn't expecting w/r/t trying to overwrite a resource14:51
<peichman>running fcrepo-webapp-plus with webac, and if I try to do a PUT to replace a resource (even as admin), I am getting a 409 Conflict response
from the Jetty log:
INFO 14:50:19.714 (FedoraLdp) PUT resource 'foo'
ERROR 14:50:19.716 (PersistingRdfStreamConsumer) Found repository-managed mixin http://www.jcp.org/jcr/nt/1.0folder in triple [http://localhost:8080/rest/foo, http://www.w3.org/1999/02/22-rdf-syntax-ns#type, http://www.jcp.org/jcr/nt/1.0folder] on which we will not operate.
<rdfloyd>We were on the line but we aren't being heard. Trying to reconnect now...14:53
<awoods>peichman: you are maybe trying to change repository-managed triples?14:54
not allowed14:55
<peichman>@awoods: it is giving me that error even when the replacement resource just consists of sincgle triple, <> dc:title "New title".14:56
<awoods>peichman: because you are implicitly trying to remove repo-managed properties.
<peichman>@awoods: and I also get it if I PUT foo.ttl twice, the first time succeeds, the second time it fails14:57
<awoods>not surprising14:58
you need to PUT, GET, then PUT
that makes sense
okay, now I am getting a 400, which is better, but I was expecting a 40315:02
(I am testing a PUT to a resource where the user has only read permissions)
<awoods>peichman: 400 usually means user error15:03
<peichman>@awoods, yeah, I am just being pedantic about my response codes :-)15:04
<awoods>peichman: you may have better success using PATCH vs. PUT
<peichman>@awoods thats what I am thinking15:05
<awoods>peichman: and a sparql-update body
<peichman>especially for demonstration purposes
<peichman>I want to have simple example of "this user can read but not write"
"then admin changes the ACL and now they can write"
@awoods: Content-Type: text/sparql-update?15:06
<peichman>@awoods: got it
* Dan_Davis leaves15:08
<peichman>@awoods: hey look, a 403 just like I was expecting! thanks :-)
* rdfloyd leaves15:23
<whikloj>acoburn/awoods: I am stumped with this header thing.15:24
MessageBodyWriter does not work as only one is selected per response.
I tried to use a UriAwareResourceModelFactory, which might work except I'm not sure how to access the response to addHeaders15:25
<acoburn>whikloj: I'll have to defer to awoods on this
<whikloj>Last path was wondering why we change the value to node here: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-kernel-modeshape/src/main/java/org/fcrepo/kernel/modeshape/rdf/impl/SkolemNodeRdfContext.java#L8015:26
<acoburn>whikloj: however, if you push your code to a public branch, I can take a look later this afternoon
<whikloj>acoburn: Fair enough, I might have to as well :)
<awoods>whikloj: I am not sure how much I can help at the moment...15:27
* dwilcox leaves
<whikloj>awoods: it's fine I will keep searching the inter-web for more JAX-RS secrets
<ajs6f>whikloj: ValueConverter15:28
whikloj: Sorry, https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-kernel-modeshape/src/main/java/org/fcrepo/kernel/modeshape/rdf/converters/ValueConverter.java#L176
<awoods>whikloj: have you looked through the JAX-RS spec?
<ajs6f>whikloj: That conversion is tracing through links.
<awoods>whikloj: http://download.oracle.com/otndocs/jcp/jaxrs-2_0-fr-eval-spec/index.html
<whikloj>awoods: not the spec no15:29
ajs6f: what does the ValueConverter get me?
<ajs6f>whikloj: Look at that line refernce. That is the actual conversion. It's converting a value that is a reference to a node to that node.
whikloj: Dereferencing. Following a link.15:30
<whikloj>ajs6f: sorry, my question is really why in SkolemNodeRdfContext do we have to deference all the references from a Node, which causes problems because if you don't have permission to read the node you can't even know it exists.15:31
^^ Talking in a authentication sense, ie. having a restriction and being aware of it but not being able to read the restriction15:32
<ajs6f>whikloj: We're checking to see which one of those nodes might turn out to be a Skolem/bnode and therefore should be included in the RDF that SkolemNodeRdfContext is creating/returning.15:33
<whikloj>ajs6f: and if you don't have permission to read the node you can't even know it's URI.
ajs6f: which seems limiting, but I don't fully understand it.15:34
<ajs6f>whikloj: Fair enough— I think you'll have to put a filter in there to do the check.
whikloj: The auth check, that is.
whiklloj: Just one more filter, as long as you have some way to get at the rules of who is and ain't allowed to see the node.
<peichman>draft of a "Quick Start with WebAC" document: https://gist.github.com/peichman-umd/6895e84ebbee0f8579d915:35
@awoods/@acoburn/@whikloj: thoughts on a permanent home for it?
<awoods>peichman: wiki15:36
<whikloj>ajs6f: the auth check is handled in the valueToNode function, problem is if you don't have permission to read the node how can you get at it to display it?
<ajs6f>whikloj: You have in hand an ID for the node, right?15:37
<peichman>@awoods: if wiki, then I would deifnitely link to it from the fcrepo-webapp-plus README
<whikloj>peichman: I agree with awoods, the wiki is best
<whikloj>ajs6f: it appears to be a property, but yes.
<peichman>3 votes for wiki, wiki it is :-)
<ajs6f>whikloj: Sorry, yes, the value of that prop is a Node ID, right?15:38
* dwilcox joins
<whikloj>ajs6f: I believe so15:39
<ajs6f>whikloj: Okay, so we need something that lets us give an ID to JCR and get back an auth result, right?
<whikloj>ajs6f: not exactly. My current use case is the <> acl:accessControl <some other resource in F4>15:41
<ajs6f>whikloj: Not sure I follow you. How is SkolemNodeRdfContext getting involved?15:42
<whikloj>ajs6f: honestly I'm not sure it is. I may have left the path and wandered into the woods again.
<ajs6f>whikloj: I'm going to start walking quietly backwards now.15:48
<whikloj>ajs6f: sorry here is the correct story
ajs6f: When using PropertiesRdfContext to display the properties on a resource, we use the ValueConverter to translate a PropertyToTriple15:49
<ajs6f>whikloj: Sounds plausible. I'll buy that.15:50
<whikloj>ajs6f: But if you don't have access to that property (which is also a resource in Fedora), can we return the URI. Instead of getting an Access Denied
ajs6f: It seems like traversing all the links is expensive for presenting a RdfStream, no?15:51
<ajs6f>whikloj: How is that property a resource in Fedora? Isn't it a predicate in a triple?
<whikloj>ajs6f: no it is the object in a triple
ajs6f: so when I search as a user with access to the resource I see15:52
acl:accessControl fedora:inaccessibleResource
<ajs6f>whikloj: I think we have to be careful here. The node on the other end of that link is the object, right? The property is a predicate, the value of which is an ID for a node, right?
<whikloj>ajs6f: I think so... but I think lots of things15:53
* github-ff joins
[fcrepo-module-auth-webac] awoods pushed 1 new commit to master: http://git.io/vlYci
fcrepo-module-auth-webac/master 4871919 Aaron Coburn: Implement support for acl:agentClass triples...
* github-ff leaves
<ajs6f>whikloj: I think that Rokusaburo Michiba was the greatest Iron Chef of all.
* github-ff joins15:54
[fcrepo-module-auth-webac] awoods closed pull request #40: Implement support for acl:agentClass triples (master...fcrepo-1793) http://git.io/vWbMZ
* github-ff leaves
<ajs6f>whikloj: But more importantly, I think that ValueConverter has some gear for this already.
whikloj: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-kernel-modeshape/src/main/java/org/fcrepo/kernel/modeshape/rdf/converters/ValueConverter.java#L159
<whikloj>ajs6f: yes, but that is why I get fedora:inaccessibleResource in the RDF. Why can't I know the ID of the node, even if I can't "GET" it15:55
ajs6f: it seems those are two separate actions
one is getting the properties on a resource I do have access to, the other is accessing any referenced resources15:56
<ajs6f>whikloj: Yeah, I get what your're saying, but I think that may be the way that JCR permissions work. Take a look here:
whikloj: Notice that all of the "hasPermission" methods require a path.
whikloj: With an ID, you can only get that from http://www.day.com/specs/javax.jcr/javadocs/jcr-2.0/javax/jcr/Session.html#getNodeByIdentifier%28java.lang.String%2915:57
whikloj: Which will already throw the exception.
<whikloj>ajs6f: ok...that makes sense. Thanks for explaining to me15:58
<ajs6f>whikloj: Are you saying you believe that? OMG I can't believe I got away with at!
* travis-ci joins
fcrepo4/fcrepo-module-auth-webac#51 (master - 4871919 : Aaron Coburn): The build passed.
Change view : https://github.com/fcrepo4/fcrepo-module-auth-webac/compare/06dc9c615d05...487191972e03
Build details : https://travis-ci.org/fcrepo4/fcrepo-module-auth-webac/builds/88419988
* travis-ci leaves
<whikloj>ajs6f: wha... but... who the.
* ajs6f scampers away giggling gleefully16:01
* github-ff joins16:04
[fcrepo-webapp-plus] peichman-umd opened pull request #28: Link to Quick Start with WebAC guide on the wiki. (master...webac-quickstart-guide) http://git.io/vlY4B
* github-ff leaves
<ajs6f>whikloj: If you wanted to be probably much too clever, you could use http://www.day.com/specs/javax.jcr/javadocs/jcr-2.0/javax/jcr/Session.html#impersonate%28javax.jcr.Credentials%29 to get the path (and thence the URI). But that seems really too clever by half.
* dhlamb leaves16:06
<whikloj>ajs6f: I'll hang on to that, though I wonder if you would already have to have admin access or throw a LoginException
* github-ff joins16:07
[fcrepo-webapp-plus] ajs6f pushed 2 new commits to master: http://git.io/vlYBy
fcrepo-webapp-plus/master 1ed67d7 Peter Eichman: Link to Quick Start with WebAC guide on the wiki.
fcrepo-webapp-plus/master 8c514a4 A. Soroka: Merge pull request #28 from peichman-umd/webac-quickstart-guide...
* github-ff leaves
<ajs6f>whikloj: No, the point of that method is that you offer new Credentials. I'm not seriously suggesting you use it.
<whikloj>ajs6f: wha...but... who the.16:08
ajs6f: currently proposed solution is to generate an internal session and use that to get the URI.
<ajs6f>whikloj: Yeah, it's not that different. Keep in mind that you must _NOT_ leave long-lived open sessions. That is massive performance hit.16:09
<whikloj>ajs6f: noted
<ajs6f>whikloj: the idea with impersonate is that you aren't shifting "personas", so you can still have some amount of auth policy executed by the repo.16:10
whikloj: The JCR repo, that is.
* codeFerret joins
<whikloj>ajs6f: But you can get different permissions?16:11
<ajs6f>whikloj: Yeah, if the JCR auth allows it.
* mikeAtUVa leaves
<whikloj>ajs6f: cool
<ajs6f>whikloj: The idea is that the right to impersonate someone is just another permission you may have.
whikloj: It's like this: https://www.youtube.com/watch?v=_P1bu4HUAMs16:12
* travis-ci joins16:13
fcrepo4-exts/fcrepo-webapp-plus#87 (master - 8c514a4 : A. Soroka): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-webapp-plus/compare/77ade85a1658...8c514a42d986
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-webapp-plus/builds/88422190
* travis-ci leaves
* codeFerret leaves
<whikloj>ajs6f: now that is something I can get in to16:14
<ajs6f>whikloj: It's all just a bunch of razzle-dazzle, anyway.
<whikloj>ajs6f: foo-shiz16:16
* github-ff joins16:22
[fcrepo-module-auth-webac] awoods pushed 2 new commits to master: http://git.io/vlY2F
fcrepo-module-auth-webac/master c3f3916 Jared Whiklo: Fix 404 for non-admin when requesting versioned resources....
fcrepo-module-auth-webac/master fa718f2 Andrew Woods: Merge pull request #42 from whikloj/fcrepo-1760...
* github-ff leaves
* travis-ci joins16:27
fcrepo4/fcrepo-module-auth-webac#52 (master - fa718f2 : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo-module-auth-webac/compare/487191972e03...fa718f2c3568
Build details : https://travis-ci.org/fcrepo4/fcrepo-module-auth-webac/builds/88424688
* travis-ci leaves
<ajs6f> Out for the week. Time to go take the boy to the downtown merchants' Halloween dogs-in-costumes parade, which should thoroughly blow his 18-month-old mind.16:30
* ajs6f leaves
* umgrosscol leaves16:31
* mohamedar leaves16:38
* escowles leaves16:48
* dwilcox leaves16:50
* bseeger leaves17:00
* esm_ leaves17:12
* acoburn leaves17:22
* chadmills joins17:28
* peichman leaves17:33
<whikloj>awoods: ping17:42
* avmich joins17:59
my claster doesn't seem to fom (i.e., two nodes don't join into a cluster). How can I figure out where the process of forming the cluster breaks? I'm using Fedora 4.218:00
to form
* chadmills leaves18:03
* whikloj leaves
* github-ff joins19:28
[fcrepo4-vagrant] jrgriffiniii opened pull request #29: FCREPO-1777 Provide Quality Assurance for fcrepo4-vagrant (master...fcrepo-1777-jrgriffiniii) http://git.io/vlOdd
* github-ff leaves
* avmich leaves19:47
* the_mgt_ joins20:21
* the_mgt leaves20:25
* apb18 leaves21:00
* dhlamb joins21:19
* github-ff joins21:59
[fcrepo-module-auth-webac] acoburn opened pull request #43: implement support for access controls directly on binary resources (master...fcrepo-1791) http://git.io/vl3uT
* github-ff leaves
* apb18 joins22:01
* apb18 leaves22:02
* ksclarke leaves22:44
* github-ff joins22:58
[fcrepo-webapp-plus] acoburn opened pull request #29: Remove MaxPermSize flag from maven configuration (master...remove-permgen) http://git.io/vl360
* github-ff leaves
* dhlamb leaves23:39

Generated by Sualtam